SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 1
Created by Christy Owen of Liberty University’s Online Writing Center
onlinewriting@liberty.edu; last date modified: November 7, 2021
Sample APA Paper: Professional Format for Graduate/Doctoral Students
Claudia S. Sample
School of Behavioral Sciences, Liberty University
Author Note
Claudia S. Sample (usually only included if author has an ORCID number)
I have no known conflict of interest to disclose.
Correspondence concerning this article should be addressed to Claudia S. Sample.
Email: cssample123456789@liberty.edu
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 2
Table of Contents
(Only Included for Easy Navigation; Hyperlinked for Quick Access)
Sample APA Paper: Professional Format for Graduate/Doctoral Students ……………………………… 6
Basic Rules of Scholarly Writing ……………………………………………………………………………………… 7
Brief Summary of Changes in APA-7 ………………………………………………………………………………… 8
Running Head, Author Note, and Abstract …………………………………………………………………………. 9
Basic Formatting Elements …………………………………………………………………………………………….. 10
Font ………………………………………………………………………………………………………………….. 10
Line Spacing ……………………………………………………………………………………………………… 10
Spaces After Punctuation …………………………………………………………………………………….. 11
Footnotes …………………………………………………………………………………………………………… 11
Heading Levels—Level 1 ………………………………………………………………………………………………. 11
Level 2 Heading …………………………………………………………………………………………………. 12
Level 3 Heading ………………………………………………………………………………………. 13
Level 4 Heading. Must be bolded and indented ½”. Add a period, one
space, and begin your content on the same line as shown here. ………………………………… 13
Level 5 Heading …………………………………………………………………. 13
Specific Elements of Academic Papers ……………………………………………………………………………. 13
Tables of Contents and Outlines …………………………………………………………………………… 13
Annotated Bibliographies ……………………………………………………………………………………. 14
Appendices ………………………………………………………………………………………………………… 14
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 3
Crediting Your Sources………………………………………………………………………………………………….. 15
Paraphrasing and Direct Quotes ……………………………………………………………………………. 15
Paraphrasing ……………………………………………………………………………………………. 16
Block Quotes …………………………………………………………………………………………… 16
How Often to Cite Your Source in Each Paragraph ………………………………………………… 17
Rule for Omitting the Year of Publication ……………………………………………………………… 17
Arranging the Order of Resources in Your Citations ………………………………………………. 17
Two Works by the Same Author in the Same Year …………………………………………………. 18
Two Works by Two Different Authors with the Same Last Name ……………………………. 18
Three or More Authors Cited In-Text ……………………………………………………………………. 18
Number of Authors in the Reference List ………………………………………………………………. 19
Numbers ………………………………………………………………………………………………………………………. 19
Displaying Titles of Works in-Text …………………………………………………………………………………. 19
Primary Sources versus Secondary Sources ……………………………………………………………………… 20
Personal Communications ……………………………………………………………………………………………… 20
Resources Canonically Numbered Sections (i.e., the Bible and Plays) …………………………………. 21
Bible and other Classical Works …………………………………………………………………………… 21
Plays …………………………………………………………………………………………………………………. 22
Lectures and PowerPoints ………………………………………………………………………………………………. 22
Dictionary Entries …………………………………………………………………………………………………………. 23
Changes in Reference Entries …………………………………………………………………………………………. 23
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 4
Electronic Sources ………………………………………………………………………………………………………… 24
Adding Color ……………………………………………………………………………………………………………….. 24
Self-Plagiarism ……………………………………………………………………………………………………………… 25
Final Formatting Tweaks ……………………………………………………………………………………………….. 26
Exhaustive Reference List Examples & Additional Helpful Resources ………………………………… 26
Conclusion …………………………………………………………………………………………………………………… 29
References ……………………………………………………………………………………………………………………. 30
Appendix ……………………………………………………………………………………………………………………… 40
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 5
Abstract
Begin your abstract at the left margin. This is the only paragraph that should not be indented.
Unless otherwise instructed, APA recommends an abstract be no more than 250 words. It should
generally not contain any citations or direct quotes. This should be a tight, concise summary of
the main points in your paper, not a step-by-step of what you plan to accomplish in your paper.
Avoid phrases such as “this paper will,” and just structure your sentences to say what you want
to say. The following three sentences exemplify a good abstract style: There are many
similarities and differences between the codes of ethics for the ACA and the AACC. Both include
similar mandates in the areas of —-, —, and —. However, each differs significantly in the areas
of —, —, and —. For more detailed information, see “Writing an Abstract” at
https://www.liberty.edu/casas/academic-success-center/wp-content/uploads/sites/28/2019/04/
Writing_an_Abstract_Revised_2012.pdf (note that you would not include any links in your
abstract). This is just now at 168 words, so eyeball how brief your abstract must be. Think of
your paper as a movie you want to sound enticing, and the abstract as the summary of the plot
you would share to draw people’s interest into wanting to come and see your movie. You want to
really hook and intrigue them. What you have to say is important! Remember to stay under 250,
words. Keywords highlight the search terms someone would use to find your paper in a database.
Keywords: main words, primary, necessary, search terms
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 6
Sample APA Paper: Professional Format for Graduate/Doctoral Students
The title of your paper goes on the top line of the first page of the body (American
Psychological Association [APA], 2019, section 2.11). It should be centered, bolded, and in title
case (all major words—usually those with four+ letters—should begin with a capital letter)—see
p. 51 of your Publication Manual of the American Psychological Association: Seventh Edition
(APA, 2019; hereinafter APA-7). It must match the title that is on your title page (see last line on
p. 32). As shown in the previous sentence, use brackets to denote an abbreviation within
parentheses (bottom of p. 159). Write out the full name of an entity or term the first time
mentioned before using its acronym (see citation in first sentence in this paragraph), and then use
the acronym throughout the body of the paper (section 6.25).
There are many changes in APA-7. One to mention here is that APA-7 allows writers to
include subheadings within the introductory section (APA, 2019, p. 47). Since APA-7 now
regards the title, abstract, and term “References” to all be Level-1 headings, a writer who opts to
include headings in his or her introduction must begin with Level-2 headings as shown above
(see section 2.27) for any divisions within the introductory section.
If you do choose to include headings in your introduction section (which is optional), be
sure to include two or more subheadings, since APA (2019) forbids stand-alone heading levels.
A second notable change in APA-7 is that writers are no longer required to cite their source every
single sentence that content from it is mentioned (section 8.1). As demonstrated in this paper,
since all of the content (other than the examples included for illustration and reference-entry
variation purposes) comes directly from the APA-7 itself, citations to the APA-7 are only
included for the first instance in each paragraph. Section and/or page numbers are included
parenthetically throughout for the sake of students who desire to know exactly where the stated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 7
rule appears in the APA-7 itself. In your academic papers, however, it is critical to include the
required author(s) and year, as applicable, for all citations that are included; this may include
more than one citation for each resource per paragraph, as required to avoid any confusion about
the source of that content.
Basic Rules of Scholarly Writing
Most beginning students have difficulty learning how to write papers and also format
papers correctly using the seventh edition of the APA manual. However, the Liberty University
Online Writing Center’s (OWC) mission includes helping students learn how to be autonomous,
proficient writers. The OWC also provides students with templates to help them with basic
formatting elements, but this sample paper is designed to help graduate and doctoral students
learn to master APA rules and formatting on their own, which will prove helpful as they progress
in their studies and work toward future publication in scholarly journals.
For the purpose of instruction, this paper will use second person (you, your), but third
person (this author) must be used in most student papers. First person (I, me, we, us, our) is not
generally permitted in academic papers. Students should refrain from using first or second person
in college courses (even though the APA manual encourages this in other writing venues) unless
the assignment instructions clearly permit such (as in the case of personal reflection sections or
life histories). If in doubt, students should clarify with their professors.
APA-7 delineates separate rules and guidelines between “student” and “professional”
writers (APA, 2019). Because a primary purpose of graduate and doctoral studies is to prepare
those students to publish professionally, Liberty University has decided to have undergraduate
students follow APA-7’s guidelines for “student papers,” and graduate/doctoral students follow
APA-7’s guidelines for “professional papers.” Separate templates are available for each level.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 8
This sample paper illustrates and discusses the rules and formatting of professional papers, as
required for all Liberty University graduate and doctoral courses using APA-7 style.
Brief Summary of Changes in APA-7
Most of these changes will be discussed in more detail below; this is just a very brief
overview here. APA-7 reverts back to only one space after closing punctuation in the body of the
paper (APA-6 required two spaces; APA, 2019, section 6.1). Student (undergraduate) papers no
longer include a running head or abstract (sections 2.2 and 2.8); professional (graduate/doctoral)
papers require an abstract but the running head is now the same on all pages (the added phrase
“Running head:” from APA-6 has been eliminated; see section 2.8). Title pages are different for
both student and professional formats. The title of a paper is no longer limited to 12 words
(section 2.4).
Citations of all resources with three or more authors now use the first author’s last name
and the term et al. (APA, 2019, section 8.17). Reference entries must name up to the first 19
authors before adding an ampersand and ellipsis (up from APA-6’s six authors; section 9.8).
APA-7 omits the phrase DOI and instead standardizes DOIs to be presented in hyperlink format
(i.e., https://doi.org/10.1177%2F1524838017742386; section 9.35). Formatting guidelines for
annotated bibliographies are included in APA-7 (section 9.51), as well as expanded and
standardized reference entry examples. As discussed above, it is no longer necessary to cite a
source every single time you refer to content gleaned from it as long as it is clear the content
comes from that source (section 8.1); APA-7 also expanded the specific location noted in the
citation to include page, paragraph, section (as used throughout this sample paper, to direct the
student to the exact relevant content), chapter, timestamp, etc. (section 8.13).
APA-7 allows for “self-plagiarism” (clarified and defined below). It also invites writers to
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 9
highlight the most relevant work first, rather than just present all works in alphabetical order
(APA, 2019, section 8.12).
Heading-level formatting has changed, and APA-7 provides more flexibility in font and
line spacing (APA, 2019). The Bible must now be included in the reference list and its citations
must include the editor’s details and year (section 8.28); there are also new rules for dictionary
entries. Publisher city and state details are omitted from all reference entries except those
involving presentations or conferences, as is the phrase “retrieved from.” Hyperlinks should be
live, but they may be either presented as blue underlining or plain black text.
Running Head, Author Note, and Abstract
APA (2019) delineates separate formatting requirements for what it terms “student” and
“professional” papers. Its descriptions for those labels, however, suggests that it regards
undergraduate-level writing to fall within the student purview, and graduate/doctoral-level
writing (including dissertations and theses) to fall within the professional purview. Since a
significant goal in graduate and post-graduate studies is preparing those students to publish in
scholarly journals at and beyond graduation, it makes sense to train those students in the
formatting that is required for professionals. As such, Liberty University has opted to require the
APA-7’s “student” version format for all undergraduate assignments using APA, and its
“professional” version for all graduate and doctoral assignments. To that end, this being the
sample paper for professional formatting, it includes the additional elements required for such: a
running head (same on all pages), an author’s note, and an abstract. Graduate and doctoral
students will use this format. Note that the first “paragraph” under the author’s note is generally
only included if the author has an ORCID number, which most students will not have. However,
it is included in this sample paper and the corresponding template because the purpose of these
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 10
resources is to prepare students to publish manuscripts post-graduation. The student’s full
address, however, is intentionally omitted from the Liberty University template and this sample
paper for privacy and safety reasons, since student papers are often unfortunately published
online and disclosing their home addresses could pose safety risks.
Basic Formatting Elements
Font
APA-7 does not prescribe a specific font or size (APA, 2019, section 2.19) but rather
allows for some choice (e.g., 12-point Times New Romans, 11-point Calibri, 11-point Arial, 11-
point Georgia, or 10-point Lucinda Sans Unicode). Most journals and academic institutions will
have a preference, however, as even APA-7 acknowledges on p. 44. For this reason—and
because font size can easily be changed if an editor interested in publishing a student’s work
prefers a different font—Liberty University recommends that students use 12-point Times New
Romans or 11-point Calibri font for the body text in all academic papers. Data in charts, figures,
and tables should be presented in 8- to 14-point size in either Calibri, Arial, or Lucinda Sans
Unicode font. Students are not permitted to use any fonts such as script, calligraphy, poster,
decorative, or others not found in published scholarly journals. Since APA-7 itself authorizes a
variety of fonts and sizes, assignments will be gauged by word count rather than page count.
Word count constitutes the number of words within the body of the paper, and excludes the title
page, abstract, reference list, appendices, and other supplemental resources.
Line Spacing
APA-7 adds extra/blank lines on the title page (APA, 2019, sections 2.5, 2.7, 2.21). It also
specifies that spacing in tables and figures may be single-, 1-1/2-, or double-spaced; equations
can be triple- or quadruple-spaced. Footnotes, when used at the bottom of a page, should be
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 11
single-spaced (section 2.21).
Spaces After Punctuation
APA-7 reverts back to just one space after closing punctuation in the body of the paper, as
well as in reference entries (APA, 2019, section 6.1). Ordinarily, it would be improper to have a
paragraph with only one sentence, though APA itself asserts that for its purposes “sentences and
paragraphs of any length are technically allowed.”1
Footnotes
This leads to another new rule in APA-7, one allowing the inclusion of footnotes (APA,
2019, section 2.13). Footnotes should be use very sparingly and are appropriate to include
information such as that in the prior section to alert the reader to supplemental material that is
available online for that thought. Though APA-7 authorizes placement of footnote content either
at the bottom of the page (as in this sample paper) or on a separate page after the reference list
(section 2.21), Liberty University recommends that student place them, when used, at the bottom
of the page, as shown here.
Heading Levels—Level 1
This sample paper uses primarily two levels of headings (Levels 1 and 2). APA style,
however, has five heading levels, which will be demonstrated briefly for visual purposes. See
section 2.27 of your APA-7 (APA, 2019) for more details on heading levels and formatting. In
APA-7, all heading levels are now bolded and in title case (capitalize each major word—usually
those with four or more letters, including hyphenated compound words). Do not capitalize
articles (a, an, the) in headings unless they begin a title or follow a colon. Level 1 headings are
centered, with the content falling on the line beneath each, in standard paragraph format.
1 See https://blog.apastyle.org/apastyle/2016/05/index.html
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 12
Many students misunderstand that you progress from Level 1 to Level 2 to Level 3 to
Level 4 to Level 5, but that is not correct. In fact, your paper may have only Level 1 headings, or
just Levels 1 and 2. The rule of thumb is that you must have at least two of each heading level
that you use, otherwise omit that heading level.
Headings are basically styling ways of organizing your paper, without using an outline
format. APA specifies five levels of headings; you would likely never use Level 5 and only very
rarely use Level 4 as a student. Think of each level as the different levels in an outline. Roman
numerals, for example, would be Level 1 headings. Capital letters would be Level 2 headings.
Numerals would be Level 3 headings. Lowercase letters would be Level 4. And lowercase
Roman numerals would be Level 5. You must always have two or more of each subheading, but
you do not need every level. You start with Level 1 and work down from that (but not
consecutive 1-2-3-4-5). Under a Level 1, you would either have two+ Level 2 headings or none
at all (just one big section in paragraphs before the next Level 1 section).
Special note about conclusion sections: Please note that some of the sample papers
published by APA to demonstrate proper APA-7 format (including the “professional” sample on
pp. 50-60 of the APA-7 manual) depict the “Conclusion” section with a Level-2 heading. This is
limited to empirical papers that are being submitted for publication in scholarly journals, as those
conclusions pertain to the “Discussion” sections in such papers and are not conclusions of the
overall papers themselves. Conclusions in academic papers at Liberty University will be Level 1
headings (including dissertations and theses, which are divided by chapters, unlike journal article
manuscripts).
Level 2 Heading
Level 2 headings are left-justified (APA, 2019, p. 48). The supporting information is
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 13
posed in standard paragraph form beneath it. Never use only one of any level of heading. You
must use two or more of any level you use, though not every paper will require more than one
level. The heading levels are simply demonstrated here for visual purposes, but you would
always have two or more of each under a larger heading, as shown throughout all the other
sections of this sample paper.
Level 3 Heading
Level 3 headings are bolded, left-justified, and italicized; the content falls on the line
underneath, as with Levels 1 and 2.
Level 4 Heading. Must be bolded and indented ½”. Add a period, one space, and begin
your content on the same line as shown here.
Level 5 Heading. Same as Level 4, but also italicized. Despite heavy writing experience,
this author has never used Level 5 headings.
Specific Elements of Academic Papers
Tables of Contents and Outlines
APA (2019) does not regulate every type of paper and some elements in various
assignments are not addressed in the APA-7 manual, including outlines and tables of content. In
those cases, follow your professor’s instructions and the grading rubric for the content and
format of the outline or annotations, and use standard APA formatting for all other elements
(such as running head, title page, body, reference list, 1″ margins, double-spacing, permitted
font, etc.). Note that most academic papers will not require a table of contents, nor would one be
appropriate. One was included in this paper simply for ease-of-access so students could go
directly to the content they want to see. Generally speaking, no table of contents would be
necessary for papers less than 20 pages of content, unless otherwise required by your professor.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 14
That being said, when organizing outlines in APA format, set your headings up in the
proper levels (making sure there are at least two subheadings under each level), and then use
those to make the entries in the outline. As discussed above, Level 1 headings become uppercase
Roman numerals (I, II, III), Level 2 headings become capital letters (A, B, C), Level 3 headings
become numbers (1, 2, 3), Level 4 headings become lowercase letters (a, b, c), and Level 5
headings become lowercase Roman numerals (i, ii, iii). Many courses now require “working
outlines,” which are designed to have the bones and foundational framework of the paper in
place (such as title page, abstract, body with title, outline/heading divisions, supporting content
with citations, and references), without the full “meat” that fills out and forms a completed paper.
Annotated Bibliographies
Many Liberty University courses also now require students to prepare and submit an
annotated bibliography as a foundational step to building a research paper. There is significant
merit in these assignments, as they teach students to critique the resources they have found and
rationalize why each is relevant for their paper’s focus. APA (2019) includes a section on
annotated bibliographies (9.51; see the example provided on p. 308). The appendix attached to
this sample paper also includes a sample annotated bibliography.
Appendices
Appendices, if any, are attached after the reference list (APA, 2019, section 2.14). You
must refer to them (i.e., “callout”) in the body of your paper so that your reader knows to look
there (see the yellow-highlighted callouts to Table 1 on p. 54 and to Footnote 1 on p. 55 of your
APA-7 for visuals on how this should appear in your paper). The word “Appendix” is singular;
use it to refer to individual appendices. APA-7 regards it as a Level 1 heading so it should be
bolded. I attached a sample Annotated Bibliography as a visual aid (see Appendix). You will see
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 15
that I included the title “Appendix” at the top of the page and formatted it in standard APA
format beneath that. Because I only included one appendix, it is simply titled as such. If there are
more appendices, assign a letter to each and denote each by that: “Appendix A” and “Appendix
B.”
Crediting Your Sources
Paraphrasing and Direct Quotes
Paraphrasing is rephrasing another’s idea in one’s own words by changing the wording
sufficiently without altering the meaning (remember not to just change a word here or there or
rearrange the order of the original source’s wording). Quoting is using another’s exact words.
Both need to be cited; failure to do so constitutes plagiarism. Include the author(s) and year for
paraphrases, and the author(s), year, and page or paragraph number for direct quotes. APA-7 also
expands this to include figure number, time stamp, etc.—whatever detail is necessary to get the
reader directly to that content. Page numbers should be used for any printed material (books,
articles, etc.), and paragraph numbers should be used in the absence of page numbers (online
articles, webpages, etc.; see APA, 2019, section 8.13). Use p. for one page and pp. (not italicized
in your paper) for more than one (section 8.25). Use para. for one paragraph and paras. (also not
italicized in your paper) for two or more (section 8.28). For example: (Perigogn & Brazel, 2012,
pp. 12–13) or (Liberty University, 2019, para. 8). Section 8.23 of the APA (2019) manual
specifies that it is not necessary to include a page or paragraph number for paraphrases (just for
direct quotes), but writers may choose to do so to help their readers find that content in the cited
resource.
When naming authors in the text of the sentence itself (called a narrative citation), use the
word “and” to connect them. For example, Perigogn and Brazel (2012) contemplated that . . .
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 16
Use an ampersand (&) in place of the word “and” in parenthetical citations and reference lists:
(Perigogn & Brazel, 2012).
Paraphrasing
Only use quotes when the original text cannot be said as well in your own words or
changing the original wording would change the author’s meaning. You cannot simply change
one word and omit a second; if you paraphrase, the wording must be substantially different, but
with the same meaning. Regardless, you would need to cite the resource you took that
information from. For example, Benoit et al. (2010) wrote that “although, a link between
attachment and posttraumatic stress disorder (PTSD) symptoms has been established, the
mechanisms involved in this link have not yet been identified” (p. 101). A paraphrase for that
quote might be: A link between dysfunctional attachment and the development of PTSD has
been made, though there is insufficient data to determine exactly how this mechanism works
(Benoit et al., 2010).
Block Quotes
Quotes that are 40 or more words must be blocked, with the left margin of the entire
quote indented ½ inch. Maintain double-spacing of block quotes. APA prefers that you introduce
quotes but note that the punctuation falls at the end of the direct quote, with the page number
outside of that (which is contrary to punctuation for non-blocked quotes). For example, Alone
(2008) claims:2
Half of a peanut butter sandwich contains as much bacteria as the wisp of the planet
Mars. Thus, practicality requires that Mrs. Spotiker nibble one bit at a time until she is
assured that she will not perish from ingesting it too quickly. (p. 13)
2 Note that there are no quotation marks for block quotes, as shown in the example.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 17
Usually quotes within quotes use single quotation marks; however, use double quotation marks
for quotes within blocked quotes, since there are no other quotation marks involved. Also
understand that direct quotes should be used sparingly in scholarly writing; paraphrasing is much
preferred in APA format (APA, 2019, section 8.23), as it demonstrates that you read, understood,
and assimilated other writers’ content into one cohesive whole.
How Often to Cite Your Source in Each Paragraph
As already mentioned above, APA’s (2019) new official rule is that you no longer must
cite your source every single time you refer to material you gleaned from it (section 8.1). It is
now acceptable to cite your source the first time you refer to content from it in your paragraph,
and then not again in that same paragraph unless your phrasing does not make the source of your
content clear. This is demonstrated throughout this sample paper.
Rule for Omitting the Year of Publication
That being said, APA (2019) has clarified its special rule that excludes the year of
publication in subsequent narrative in-text citations (when you name the authors in the text of the
sentence itself), after the first narrative citation in each paragraph. It should continue to appear in
all parenthetical citations (see section 8.16). For example, Alone (2008) portrays imagery of Mrs.
Spotiker. This includes her devouring a peanut butter sandwich (Alone, 2008). Alone conveys
this through the lens of astronomy. Note that the year of publication was omitted from the second
narrative citation (underlined for visual purposes).
Arranging the Order of Resources in Your Citations
If the material you cited was referred to in multiple resources, separate different sets of
authors with semicolons, arranged in the order they appear (alphabetically by the first author’s
last name) in the reference list (i.e., Carlisle, n.d.-a; Prayer, 2015) (APA, 2019, section 8.12).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 18
APA-7 now invites writers to prioritize or highlight one or more sources as most prominent or
relevant for that content by placing “those citations first within parentheses in alphabetical order
and then insert[ing] a semicolon and a phrase, such as ‘see also,’ before the first of the remaining
citations” (APA., 2019, p. 263)—i.e., (Cable, 2013; see also Avramova, 2019; De Vries et al.,
2013; Fried & Polyakova, 2018). Periods are placed after the closing parenthesis, except with
indented (blocked) quotes.
Two Works by the Same Author in the Same Year
Authors with more than one work published in the same year are distinguished by lower-
case letters after the years, beginning with a (APA, 2019, section 8.19). For example, Double
(2008a) and Double (2008b) would refer to resources by the same author published in 2008.
When a resource has no date, use the term n.d. followed by a dash and the lowercase letter (i.e.,
Carlisle, n.d.-a and Carlisle, n.d.-b; see APA, 2019, section 8.19).
Two Works by Two Different Authors with the Same Last Name
Citations in the body of the paper should include only the last names, unless you have
two or more resources authored by individuals with the same last name in the same year (or are
citing a personal communication). When there are two different authors with the same last name
but different first names who published in the same year, include the first initials: Brown, J.
(2009) and Brown, M. (2009) (APA, 2019, section 8.20).
Three or More Authors Cited In-Text
When referring to material that comes from three or more authors, APA-7 now requires
that all citations name just the first author’s last name followed by the words et al. (without
italics) (APA, 2019, section 8.17). Et al. is a Latin abbreviation for et alii, meaning “and others,”
which is why the word “al.” has a period, whereas “et” does not. Alone et al. (2011) stipulated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 19
that peacocks strut. Every single time I refer to their material, I would apply APA-7’s rule: Alone
et al. (2011) or (Alone et al., 2011). Since et al. denotes plural authors, the verb must be plural to
match, too: Alone et al. (2011) are… This applies to all citations within the body of the paper
with three or more authors.
Number of Authors in the Reference List
For resources with 20 or fewer authors in the reference list, write out all of the authors’
last names with first and middle initials, up to and including the 20th author (APA, 2019, section
9.8). APA-7 has a special rule for resources with 21 or more authors: Write out the first 19
authors’ last names with initials, insert an ellipsis (…) in place of the ampersand (&), and finish
it with the last name and initials of the last author. See example #4 provided on page 317 of your
APA-7, as well as this paper’s reference list for visuals of these variances (Acborne et al. 2011;
Kalnay et al., 1996).
Numbers
Numbers one through nine must be written out in word format (APA, 2019, section 6.33),
with some exceptions (such as ages—see section 6.32). Numbers 10 and up must be written out
in numerical format (section 6.32). Always write out in word format any number that begins a
sentence (section 6.33).
Displaying Titles of Works in-Text
The names of journals, books, plays, and other long works, if mentioned in the body of
the paper, are italicized in title case (APA, 2019, section 6.17). Titles of articles, lectures, poems,
chapters, website articles, and songs should be in title case, encapsulated by quotation marks
(section 6.7). The year of publication should follow the author’s name, whether in narrative or
parenthetical format: Perigogn and Brazel (2012) anticipated…, or (Perigogn & Brazel, 2012).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 20
The page or paragraph number must follow after the direct quote. Second (2015) asserted that
“paper planes can fly to the moon” (p. 13). You can restate that with a parenthetical citation as:
“Paper planes can fly to the moon” (Second, 2015, p. 13). Second (2011) is another resource by
the same author in a different year.
Primary Sources versus Secondary Sources
APA (2019) strongly advocates against using secondary sources; rather, it favors you
finding and citing the original (primary) resource whenever possible (section 8.6). On the rare
occasion that you do find it necessary to cite from a secondary source, both the primary (who
said it) and secondary (where the quote or idea was mentioned) sources should be included in the
in-text citation information. If the year of publication is known for both resources, include both
years in the citation (section 8.6). Only the secondary source should be listed in the reference
section, however. Use “as cited in” (without the quotation marks) to indicate the secondary
source. For example, James Morgan hinted that “goat milk makes the best ice cream” (as cited in
Alone, 2008, p. 117). Morgan is the primary source (he said it) and Alone is the secondary
source (he quoted what Morgan said). Only the secondary source is listed in the reference section
(Alone, and not Morgan) because if readers want to confirm the quote, they know to go to page
117 of Alone’s book.
Personal Communications
APA (2019) rationalizes the exclusion of references for information obtained through
personal communication (such as an interview, email, telephone call, postcard, text message, or
letter) in the reference list because your readers will not be able to go directly to those sources
and verify the legitimacy of the material. Instead, these items are cited only in the body of the
paper. You must include the individual’s first initial, his or her last name, the phrase “personal
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 21
communication” (without the quotation marks), and the full date of such communication (section
8.9). As with other citations, such citations may be either narrative or parenthetical. For example,
L. Applebaum advised him to dip pretzel rolls in cheese fondue (personal communication, July
13, 2015). The alternative is that he was advised to dip pretzel rolls in cheese fondue (L.
Applebaum, personal communication, July 13, 2015). Note that there is no entry for Applebaum
in the reference list below.
Resources Canonically Numbered Sections (i.e., the Bible and Plays)
These resources should be cited in book format (APA, 2019, Section 9.42). The Bible and
other religious works are generally regarded as having no author; an annotated version would be
treated as having an editor. Include republished dates as necessary. The OWC will publish a list
of reference entries for various Bible versions on its APA Quick Guide webpage.
Bible and other Classical Works
Works such as the Bible, ancient Greek or Roman works, and other classical works like
Shakespeare must be cited in the body of the paper (APA, 2019, section 8.28). APA-7 now also
requires that they be included in the reference list, too (section 9.42), which is a significant
change from APA-6. Republished dates are included as well (see section 9.41). As such, you
would add a parenthetical phrase at the end of your reference entry with the original publication
details; note that there should be no punctuation following such parenthetical content at the end
of a reference entry (the reference entries depicting this in the reference list below are correctly
punctuated).
Citations for the Bible will include the Bible version’s name in the author’s position (as
an anonymous work), original and republished years, and then the book chapter/verse (spelled
out) in place of the page number (i.e., King James Bible, 1769/2017, Genesis 3:8)—see sections
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 22
8.28 and 9.42. Note that APA (2019) requires book titles to be italicized in every venue,
including citations and reference entries. Because Liberty University is a distinctly-Christian
institution and many of its courses require biblical integration, most if not all of its students will
cite the Bible in virtually every course. The examples provided on pp. 274 and 325 of APA-7 are:
(note the italics in each)
Narrative citation: King James Bible (1769/2017)
Parenthetical citation: (King James Bible, 1769/2017, Song of Solomon 8:6)
Reference entry: King James Bible. (2017). King James Bible Online.
https://www.kingjamesbibleonline.org/ (Original work published 1769)
Plays
When citing plays, “cite the act, scene, and line(s), in a single string, separated by
periods. For example, ‘1.3.36-37’ refers to Act 1 Scene 3, Lines 36-37” (APA, 2019, section
8.28; see also example #37 on p. 325).
Lectures and PowerPoints
APA (2019) has expanded and standardized its rules for citations and reference entries in
an effort to best credit the original sources. It now includes rules for crediting content in course
or seminar handouts, lecture notes, and PowerPoint presentations (see #102 on p. 347). When
citing a PowerPoint presentation, include the slide number rather than the page number. For
purposes of Liberty University course presentations and lectures (which are not readily available
to the public), reference each as a video lecture with the URL (if available) for the presentation,
naming the presenter(s) in the author’s position. Include the course number, lecture title, and
enough details for others to identify it within that course, in a sort of book format, naming
Liberty University as publisher. Peters (2012) is an example of this in the reference list of this
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 23
paper. If the presenter for a Liberty University class lecture is not named, credit Liberty
University as the author; see Liberty University (2020) in the reference list below as an example.
Dictionary Entries
In keeping with its efforts to standardize reference entries, APA (2019) now requires
citation and referencing of word definitions from dictionaries to follow the same rules for
chapters in an edited book (see #47 and #48 on p. 328; section 8.13). As such, you will now
name either the individual, group, or corporate author of the dictionary in the author’s place (e.g.,
Merriam-Webster, n.d.). If you searched online, include the retrieval date and the URL to the
exact webpage. If you used a hard copy book, include the publisher details. The in-text citation
in the body of the paper would follow standard author/year format (e.g., Merriam-Webster, n.d.).
Changes in Reference Entries
There are a number of notable changes in APA-7 from past versions. For the most part,
these simplify and unify the formats to be more consistent across the different resource venues.
Some of these have already been discussed above (i.e., naming up to 19 authors’ names before
adding an ellipsis, and crediting authors and editors of classical works and dictionaries). Other
changes include italicizing names of webpages and website resources in the reference list (APA,
2019, section 6.22), as well as book titles even when named in the author’s position (such as
King James Bible). The city and state locations of publishers are no longer required; only include
those details “for works that are associated with a specific location, such as conference
presentations” (p. 297, section 9.31). Issue numbers are required for all journal articles that have
such, regardless of what page number each issue begins with (section 9.25). If two or more
publishers are listed on the copyright page, include all of them in the order listed, separated by
semicolons (section 9.29). Omit the word Author in the publisher’s place when it is the same as
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 24
the author (section 9.24).
Electronic Sources
Note that since the APA 6th edition was published in 2010, great strides have been made
in online and electronic resource accessibility, and APA’s position on electronic resources has
shifted to embrace this. More and more resources are available electronically through the
Internet. The advent of this increased availability has resulted in APA-7’s effort to standardize
the formatting of resources, which in turn simplifies them to some extent. All reference entries
follow the same basic details: Author(s), year of publication, name of resource, and location
details (i.e., either journal name/volume/issue/page numbers, or book publisher, or webpage).
APA (2019) requires inclusion of a Digital Object Identifier (DOI) in the references
whenever available (section 9.34); if not, then a webpage, if available. In keeping with its
unification of resources, APA-7 now standardizes all DOIs and URLs to be presented in
https://doi.org/10.1177%2F1524838017742386 format. The phrase “Retrieved from” is now
excluded except when the content may have changed (such as dictionary entries, Twitter profiles,
Facebook pages; see section 9.16). APA-7 requires all hyperlinks to be active (so your reader can
click on one to go directly to that webpage), but they may appear as either blue-underlined text
or simple black text (section 9.35). There should be no period after any URL. APA-7 no longer
requires authors to break long URLs with soft returns (hold down the Shift key and press the
Enter key) at forward slashes, periods, or underscores to avoid unsightly spacing gaps, but it may
be best to do so in academic papers.
Adding Color
Though APA (2019) authorizes writers to include the use of color in photographs and
figures (section 7.26), Liberty University discourages this in academic papers. It risks becoming
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 25
distracting for both students in their quest to be creative, and professors in their quest to focus on
academic content.
Self-Plagiarism
APA (2019) also invites writers to repurpose some of their work in future papers.
Specifically, APA-7 states that:
In specific circumstances, authors may wish to duplicate their previously used works
without quotation marks or citation …, feeling that extensive self-referencing is
undesirable or awkward and that rewording may lead to inaccuracies. When the
duplicated material is limited in scope, this approach is permissible. (p. 8.3)
APA-7 adds “Do not use quotation marks or block quotation formatting around your own
duplicated material” (p. 256).
Liberty University, however, has stringent rules against self-plagiarism, as do many
scholarly journals. Liberty University students receive grades for their class papers; those who
have received feedback and a grade from a prior professor on a prior paper have an advantage
over their classmates, both in having the benefit of that feedback/grade and in not having to write
a whole paper from scratch during the subsequent class. Student papers are also submitted to
SafeAssign to deter plagiarism. For these reasons, Liberty University expressly forbids students
using significant portions of a prior paper in a subsequent course (either a retake of the same
course or a new class altogether). It is conceivable that students who are building their
knowledge base in a subject matter—particularly at the graduate and post-graduate levels—
would reasonably justify incorporating brief excerpts from past papers into current ones. In such
case, Liberty University authorizes students to utilize APA-7’s disclosure (i.e., “I have previously
discussed”), along with a citation to the prior class paper and a reference entry (i.e., Owen, 2012;
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 26
Yoo et al., 2016). Such self-references and re-use of content from prior papers should be used
sparingly and disclosed fully in the current paper; that content should not constitute a significant
portion of any academic assignment, however.
Final Formatting Tweaks
The templates provided by Liberty University are already formatted with proper spacing,
margins, heading level structure, and hanging indents, as necessary. With the exceptions of the
title page, figures, and equations, papers in APA format should be double-spaced throughout,
with no extra spacing between lines. Academic papers at Liberty University should also be in
one of the accepted fonts throughout (recommended: Times New Romans, 12-point font).
Sometimes when you format your paper or cut-and-paste material into it, things get skewed. One
quick way to ensure that your paper appears correct in these regards is to do a final formatting
tweak after you have completed your paper. Hold down the “Ctrl” button and press the “A” key,
which selects and highlights all of the text in your paper. Then go to the Home tab in Microsoft
Word and make sure that whichever acceptable font/size you choose to use is selected in the Font
box. Next, click on the arrow at the bottom of the Paragraph tab. Set your spacing before and
after paragraphs to “0 pt” and click the “double” line spacing. The extra spacing required on the
title page is already programmed into the template and should not change even when you
complete these actions.
Exhaustive Reference List Examples & Additional Helpful Resources
The reference list at the end of this paper includes an example of a myriad of different
sources and how each is formatted in proper APA-7 format. One example of each of the primary
types of resources will be included in the reference list, as cited in the body of this paper.
Remember that, for purposes of this paper only, many of the sources cited in the body of the
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 27
paper were provided for illustrative purposes only and thus are fictional, so you will not be able
to locate them if you searched online. Nevertheless, in keeping with APA-7 style, all resources
cited in the body of the paper are included in the reference list and vice versa (except for personal
communications, per APA-7’s published exceptions). Be absolutely sure that every resource cited
in the body of your paper is also included in your reference list (and vice versa), excepting only
those resources with special rules, such as personal communications and primary sources you
could not access directly.
The reference list in this paper is fairly comprehensive and will include a book by one
author who also appears as one of many authors in another resource (Alone, 2008; Alone et al.,
2011); chapters in edited books (Balsam et al., 2019; Haybron, 2008; Perigogn & Brazel, 2012;
Weinstock et al., 2003); electronic version of book (Strong & Uhrbrock, 1923); electronic only
book (O’Keefe, n.d.); edited books with and without DOIs, with multiple publishers (Hacker
Hughes, 2017; Schmid, 2017); work in an anthology (Lewin, 1999); journal articles (Andrews,
2016; Carlisle, n.d.-a, n.d.-b; De Vries R. et al., 2013; McCauley & Christiansen, 2019);
newspaper article (Goldman, 2018; Guarino, 2017); online webpages (Liberty University, 2019;
Prayer, 2015); resource with corporate author as publisher (American Psychological Association,
2019); resources by two authors with the same last name but different first names in the same
year of publication (Brown, J., 2009; Brown, M., 2009); two resources by same author in the
same year (Double, 2008a, 2008b; Carlisle, n.d.-a, n.d.-b); two resources by the same author in
different years (Second, 2011, 2015); resource with 20 authors (maximum allowed by APA-7
before special rule applies) (Acborne et al., 2011); resource with 21 or more authors (Kalnay et
al., 1996); dictionary entries (American Psychological Association, n.d.; Graham, 2019;
Merriam-Webster, n.d.); Liberty University class lecture using course details (Peters, 2012);
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 28
PowerPoint slides or lecture notes, not including course details (Canan & Vasilev, 2019); citing a
student’s paper submitted in a prior class, in order to avoid self-plagiarism (Owen, 2012);
unpublished manuscript with a university cited (Yoo et al., 2016); code of ethics (American
Counseling Association, 2014); diagnostic manual (American Psychiatric Association, 2013);
religious and classical works, including the Bible (Aristotle, 350 BC/1994; King James Bible,
1769/2017; Shakespeare, 1623/1995); dissertation or thesis (Hollander, 2017; Hutcheson, 2012);
review of a book (Schatz, 2000); video (Forman, 1975); podcast (Vedentam, 2015); recorded
webinar (Goldberg, 2018); YouTube or other streaming video (University of Oxford, 2018); clip
art or stock image (GDJ, 2018); map (Cable, 2013); photograph (McCurry, 1985); data set (Pew
Research Center, 2018); measurement instrument (Friedlander et al., 2002); manual for a test,
scale, or inventory (Tellegen & Ben-Porah, 2011); test, scale, or inventory itself (Project
Implicit, n.d.); report by a government agency or other organization (National Cancer Institute,
2018); report by individual authors at a government agency or other organization (Fried &
Polyakova, 2018); annual report (U.S. Securities and Exchange Commission, 2017); conference
session (Fistek et al., 2017); and webpages (Avramova, 2019; Centers for Disease Control and
Prevention, 2018; National Nurses United, n.d.; U.S. Census Bureau, n.d.).
Lastly, below are a few webpages that address critical topics, such as how to avoid
plagiarism and how to write a research paper. Be sure to check out Liberty University’s Online
Writing Center (https://www.liberty.edu/online/casas/writing-center/) for more tips and tools, as
well as its Facebook page (https://www.facebook.com/groups/LUOWritingCenter). Remember
that these links are only provided for your easy access and reference throughout this sample
paper, but web links and URLs should never be included in the body of scholarly papers; just in
the reference list. Writing a research paper (https://www.youtube.com/watch?v=zaa-PTexW2E
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 29
or https://www.youtube.com/watch?v=KNT6w8t3zDY and avoiding plagiarism
(https://www.youtube.com/watch?v=VeCrUINa6nU).
Conclusion
The conclusion to your paper should provide your readers with a concise summary of the
main points of your paper (though not via cut-and-pasted sentences used above). It is a very
important element, as it frames your whole ideology and gives your readers their last impression
of your thoughts. Be careful not to introduce new content in your conclusion.
After your conclusion, if you are not using the template provided by the Online Writing
Center, insert a page break at the end of the paper so that the reference list begins at the top of a
new page. Do this by holding down the “Ctrl” key and then clicking the “Enter” key. You will go
to an entirely new page in order to start the reference list. The word “References” (not in
quotation marks) should be centered and bolded. Items in the reference list are presented
alphabetically by the first author’s last name and are formatted with hanging indents (the
second+ lines of each entry are indented 1/2” from the left margin). APA authorizes the use of
singular “Reference” if you only have one resource.3 Students would, of course, NOT include
any color-coding or footnotes in their reference entries. However, for the sake of clarity and
ease in identifying what each entry represents, each one included in the reference list of this
sample paper is color-coordinated to its corresponding footnote, with a brief description of what
each depicts.
3 https://apastyle.apa.org/instructional-aids/creating-reference-list.pdf
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 30
References
Acborne, A., Finley, I., Eigen, K., Ballou, P., Gould, M. C., Blight, D., Callum, M., Feist, M.,
Carroll, J. E., Drought, J., Kinney, P., Owen, C., Owen, K., Price, K., Harlow, K.,
Edwards, K., Fallow, P., Pinkley, O., Finkel, F., & Gould, P. P. (2011). The emphasis of
the day after tomorrow. Strouthworks. 4
Alone, A. (2008). This author wrote a book by himself. Herald Publishers. 5
Alone, A., Other, B., & Other, C. (2011). He wrote a book with others, too: Arrange
alphabetically with the sole author first, then the others. Herald Publishers. 6
American Counseling Association. (2014). 2014 ACA code of ethics.
https://www.counseling.org/knowledge-center 7
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). https://www.doi.org/10.1176/appi.books.9780890425596 8
American Psychological Association. (n.d.). Positive transference. In APA dictionary of
psychology. Retrieved August 31, 2019, from https://dictionary.apa.org/positive-
transference 9
American Psychological Association. (2019). Publication manual of the American Psychological
Association (7th ed.). 10
4 Resource with 20 authors (maximum allowed by APA before special rule applies).
5 Entry by author who also appears as one of many authors in another resource (single author
appears first in list).
6 Multiple authors appear after same single-author resource.
7 Code of ethics.
8 Diagnostic manual.
9 Entry in a dictionary, thesaurus, or encyclopedia, with group author.
10 Resource with corporate author as publisher.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 31
Andrews, P. M. (2016). Congruence matters. Educational Leadership, 63(6), 12-15. 11
Aristotle. (1994). Poetics (S. H. Butcher, Trans.). The internet Classics Archive.
http://classics.mit.edu/Aristotle/poetics.html (Original work published ca. 350 B.C.E.) 12
Avramova, N. (2019, January 3). The secret to a long, happy, heathy life? Think age-positive.
CNN. https://www.cnn.com/2019/01/03/health/respect-toward-elderly-leads-to-long-life-
intl/index.html 13
Balsam, K. F., Martell, C. R., Jones, K. P., & Safren, S. A. (2019). Affirmative cognitive
behavior therapy with sexual and gender minority people. In G. Y. Iwamasa & P. A.
Hays (Eds.), Culturally responsive cognitive behavior therapy: Practice a supervision
(2nd ed., pp. 287-314). American Psychological Association.
https://doi.org/10.1037/0000119-012 14
Benoit, M., Bouthillier, D., Moss, E., Rousseau, C., & Brunet, A. (2010). Emotion regulation
strategies as mediators of the association between level of attachment security and PTSD
symptoms following trauma in adulthood. Anxiety, Stress, and Coping, 23(1), 101-118.
http://dx.doi.org/10.1080/10615800802638279
Brown, J. (2009). Ardent anteaters. Brockton.
Brown, M. (2009). Capricious as a verb. Journal of Grammatical Elements, 28(6), 11-12. 15
11 Journal article without DOI, from most academic research databases or print version.
12 Ancient Greek or Roman work.
13 Webpage on a news website.
14 Chapter in an edited book with DOI.
15 Resources by two authors with the same last name but different first names in the same year of
publication. Arrange alphabetically by the first initials.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 32
Cable, D. (2013). The racial dot map [Map]. University of Virginia, Weldon Cooper Center for
Public Service. https://demographics.coopercenter.org/Racial-Dot-Map 16
Canan, E., & Vasilev, J. (2019, May 22). [Lecture notes on resource allocation]. Department of
Management Control and Information Systems, University of Chile. https:// uchilefau.
academia.edu/ElseZCanan 17
Carlisle, M. A. (n.d.-a). Erin and the perfect pitch. Journal of Music, 21(3), 16-17. http:// make-
sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include 18
Carlisle, M. A. (n.d.-b). Perfect pitch makes sweet music. Journal of Music, 24(8), 3-6. http://
make-sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include
Centers for Disease Control and Prevention. (2018, January 23). People at high risk of
developing flu-related complications. https://www.cdc.gov/flu/highrisk/index.htm 19
De Vries R., Nieuwenhuijze, M., Buitendijk, S. E., & the members of Midwifery Science Work
Group. (2013). What does it take to have a strong and independent profession of
midwifery? Lessons from the Netherlands. Midwifery, 29(10), 1122-1128.
https://doi.org/10.1016/j.midw.2013.07.007 20
Double, C. (2008a). This is arranged alphabetically by the name of the title. Peters.
Double, C. (2008b). This is the second (“the” comes after “arranged”). Peters. 21
16 Map.
17 PowerPoint slides or lecture notes.
18 Online journal article with a URL and no DOI; also depicts one of two resources by the same
author with no known publication date.
19 Webpage on a website with a group author.
20 Journal article with a DOI, combination of individual and group authors.
21 Two resources by same author in the same year. Arrange alphabetically by the title and then
add lowercase letters (a and b, respectively here) to the year.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 33
Fistek, A., Jester, E., & Sonnenberg, K. (2017, July 12-15). Everybody’s got a little music in
them: Using music therapy to connect, engage, and motivate [Conference session].
Autism Society National Conference, Milwaukee, WI, United States.
https://asa.confex.com/asa/2017/webprogramarchives/Session9517.html 22
Forman, M. (Director). (1975). One flew over the cuckoo’s nest [Film]. United Artists. 23
Fried, D., & Polyakova, A. (2018). Democratic defense against disinformation. Atlantic Council.
https://www.atlanticcouncil.org/in-depth-research-reports/report/democratic-defense-
against-disinformation/ 24
Friedlander, M. L., Escudero, V., & Heatherton, L. (2002). E-SOFTA: System for observing
family therapy alliances [Software and training videos] [Unpublished instrument].
http://www.softa-soatif.com/ 25
GDJ. (2018). Neural network deep learning prismatic [Clip art]. Openclipart.
https://openclipart.org/detail/309343/neural-network-deep-learning-prismatic 26
Goldberg, J. F. (2018). Evaluating adverse drug effects [Webinar]. American Psychiatric
Association. https://education.psychiatry.org/Users/ProductDetails.aspx?
ActivityID=6172 27
Goldman, C. (2018, November 28). The complicate calibration of love, especially in adoption.
22 Conference session.
23 Video.
24 Report by individual authors at a government agency or other organization.
25 Measurement instrument.
26 Clip art or stock image.
27 Webinar, recorded.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 34
Chicago Tribune. 28
Graham, G. (2019). Behaviorism. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy
(Summer 2019 ed.). Stanford University.
https://plato.stanford.edu/archives/sum2019/entries/behaviorism 29
Guarino, B. (2017, December 4). How will humanity react to alien life? Psychologists have some
predictions. The Washington Post. https://www.washingtonpost.com/news/speaking-of-
science/wp/2017/12/04/how-will-humanity-react-to-alien-life-psychologists-have-some-
predictions/ 30
Hacker Hughes, J. (Eds.). (2017). Military veteran psychological health and social care:
Contemporary approaches. Routledge. 31
Haybron, D. M. (2008). Philosophy and the science of subjective well-being. In M. Eid & R. J.
Larsen (Eds.), The science of subjective well-being (pp. 17-43). Guilford Press. 32
Hollander, M. M. (2017). Resistance to authority: Methodological innovations and new lessons
from the Milgram experiment (Publication No. 10289373) [Doctoral dissertation,
University of Wisconsin-Madison]. ProQuest Dissertations and Theses Global. 33
Hutcheson, V. H. (2012). Dealing with dual differences: Social coping strategies of gifted and
lesbian, gay, bisexual, transgender, and queer adolescents [Master’s thesis, The College
28 Newspaper article without DOI, from most academic research databases or print version
29 Entry in a dictionary, thesaurus, or encyclopedia, with individual author.
30 Online newspaper article.
31 Edited book without a DOI, from most academic research databases or print version.
32 Book chapter, print version.
33 Doctoral dissertation, from an institutional database.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 35
of William & Mary]. William & Mary Digital Archive.
https://scholarworks.wm.edu/etd/1539272210/ 34
Kalnay, E., Kanimitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S.,
White, G., Whollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J.,
Mo, K. C., Ropelewski, C., Wang, J., Leetmaa, A., … Joseph, D. (1996). The
NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological
Society, 77(3), 437-471. http://doi.org/ fg6rf9 35
King James Bible. (2017). King James Bible Online. https://www.kingjamesbibleonline.org/
(Original work published 1769) 36
Lewin, K. (1999). Group decision and social change. In M. Gold (Ed.), The complex social
scientist: A Kurt Lewin reader (pp. 265-284). American Psychological Association.
https://doi.org/10.1037/10319-010 (Original work published 1948) 37
Liberty University. (2019). The online writing center. https://www.liberty.edu/online/casas/
writing-center/ 38
Liberty University. (2020). BIOL 102: Human biology. Week one, lecture two: Name of class
lecture. https://learn.liberty.edu 39
34 Thesis or dissertation, from the web (not in a database).
35 Resource with 21 or more authors. Note the ellipse (…) in place of the ampersand (&).
36 Religious work.
37 Work in an anthology.
38 Online webpage with URL.
39 Liberty University class lecture with no presenter named.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 36
McCauley, S. M., & Christiansen, M. H. (2019). Language learning as language use: A cross-
linguistic model of child language development. Psychological Review, 126(1), 1-51.
https://doi.org/10.1037/rev0000126 40
McCurry, S. (1985). Afghan girl [Photograph]. National Geographic.
https://www.nationalgeographic.com/magazine/national-geographic-magazine-50-years-
of-covers/#/ngm-1985-jun-714.jpg 41
Merriam-Webster. (n.d.). Heuristic. In Merriam-Webster.com dictionary. Retrieved 01/02/2020,
from http://www.m-w.com/dictionary/heuristic 42
National Cancer Institute. (2018). Facing forward: Life after cancer treatment (NIH Publication
No. 18-2424). U.S. Department of Health and Human Services, National Institutes of
Health. https://www.cancer.gov/publications/patient-education/life-after-treatment.pdf 43
National Nurses United. (n.d.). What employers should do to protect nurses from Zika.
https://www.nationalnursesunited.org/pages/what-employers-should-do-to-protect-rns-
from-zika 44
O’Keefe, E. (n.d.). Egoism & the crisis in Western values. http:// www. onlineoriginals.com/
showitem.asp?itemID-135 45
40 Typical journal article with doi.
41 Photograph.
42 Dictionary entry.
43 Report by a government agency or other organization.
44 Webpage on a website with no date.
45 Electronic only book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 37
Owen, C. (2012, Spring). Behavioral issues resulting from attachment have spiritual
implications [Unpublished manuscript]. COUN502, Liberty University. 46
Perigogn, A. U., & Brazel, P. L. (2012). Captain of the ship. In J. L. Auger (Ed.) Wake up in the
dark (pp. 108-121). Shawshank Publications. 47
Peters, C. (2012). COUN 506: Integration of spirituality and counseling. Week one, lecture two:
Defining integration: Key concepts. Liberty University.
https://podcasts.apple.com/us/podcast/defining-integration-key-
concepts/id427907777?i=1000092371727 48
Pew Research Center. (2018). American trend panel Wave 26 [Data set].
https://www.pewsocialtrends.org/dataset/american-trends-panel-wave-26 49
Prayer. (2015). http:// www exact-webpage 50
Project Implicit. (n.d.). Gender–Science IAT. https://implicit.harvard.edu/implici/taketest.html 51
Schatz, B. R. (2000, November 17). Learning by text or context? [Review of the book The social
life of information, by J. S. Brown & P. Duguid]. Science, 290, 1304.
https://doi.org/10.1126/science.290.5495.1304 52
Schmid, H.-J. (Ed.). (2017). Entrenchment and the psychology of language learning: How we
reorganize ad adapt linguistic knowledge. American Psychological Association; De
46 Citing a student’s paper submitted in a prior class, in order to avoid self-plagiarism.
47 Chapter from an edited book.
48 Liberty University class lecture using course details.
49 Data set.
50 Online resource with no named author. Title of webpage is in the author’s place.
51 Test, scale, or inventory itself.
52 Review of a book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 38
Gruyter Mouton. https://doi.org/10.1037/15969-000 53
Second, M. P. (2011). Same author arranged by date (earlier first). Journal Name, 8, 12-13.
Second, M. P. (2015). Remember that earlier date goes first. Journal Name, 11(1), 18. 54
Shakespeare, W. (1995). Much ado about nothing (B. A. Mowat & P. Werstine, Eds.).
Washington Square Press. (Original work published 1623) 55
Strong, E. K., Jr., & Uhrbrock, R. S. (1923). Bibliography on job analysis. In L. Outhwaite
(Series Ed.), Personnel research series: Vol. 1, Job analysis and the curriculum (pp. 140-
146). https://doi.org/10.1037/10762-000 56
Tellegen, A., & Ben-Porah, Y. S. (2011). Minnesota Multiphasic Personality Inventory–2
Restructured Form (MPI-2-RF): Technical manual. Pearson. 57
U.S. Census Bureau. (n.d.). U.S. and world population clock. U.S. Department of Commerce.
Retrieved July 3, 2019, from https://www.census.gov/popclock 58
U.S. Securities and Exchange Commission. (2017). Agency financial report: Fiscal year 2017.
https://www.sec.gov/files/sec-2017-agency-financial-report.pdf 59
University of Oxford. (2018, December 6). How do geckos walk on water? [Video]. YouTube.
https://www.youtube.com/watch?v=qm1xGfOZJc8 60
53 Edited book with a DOI, with multiple publishers.
54 Two resources by the same author, in different years. Arrange by the earlier year first.
55 Shakespeare.
56 Electronic version of book chapter in a volume in a series
57 Manual for a test, scale, or inventory.
58 Webpage on a website with a retrieval date.
59 Annual report.
60 YouTube or other streaming video.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 39
Vedentam, S. (Host). (2015-present). Hidden brain [Audio podcast]. NPR. https://www.npr.org/
series/423302056/hidden-brain 61
Weinstock, R., Leong, G. B., & Silva, J. A. (2003). Defining forensic psychiatry: Roles and
responsibilities. In R. Rosner (Ed.), Principles and practice of forensic psychiatry (2nd
ed., pp. 7-13). CRC Press. 62
Yoo, J., Miyamoto, Y., Rigotti, A., & Ryff, C. (2016). Linking positive affect to blood lipids: A
cultural perspective [Unpublished manuscript]. Department of Psychology, University of
Wisconsin-Madison. 63
61 Podcast.
62 Chapter in an edited book without a DOI, from most academic research databases or print
version.
63 Unpublished manuscript with a university cited.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 40
Appendix
Annotated Bibliography
Cross, D. & Purvis, K. (2008). Is maternal deprivation the root of all evil? Avances en
Psycologia Latinoamericana, 26(1), 66-81.
Weaving spiritual applications throughout the article, the authors incorporate a plethora
of references to substantiate that maltreatment has a direct connection to attachment
disorders. They provide articulate and heavily-supported reasoning, detailing the specific
causes of maternal deprivation individually and then incorporating them in a broader
sense to answer the article’s title in the affirmative.
Feldman, R. (2007), Mother-infant synchrony and the development of moral orientation in
childhood and adolescence: Direct and indirect mechanisms of developmental continuity.
American Journal of Orthopsychiatry, 77(4), 582-597.
This longitudinal study tracked 31 Israeli children from ages 3 months to 13 years
(infancy to adolescence). There were direct parallels noted between increased
attachment/coherence and the child’s moral cognition, empathy development, and verbal
IQ. Toddlers who were able to regulate their own behavior later proved to excel in lead-
lag structures and language skills.
The impact of efficacy on work attitudes across cultures
Fred Luthans *, Weichun Zhu, Bruce J. Avolio
Department of Management, Gallup Leadership Institute, University of Nebraska-Lincoln,
Lincoln, NE 68588-0491, USA
Abstract
To answer the call for more cross-cultural research, this study analyzed the efficacy and work attitudes of employee samples
from the U.S. and Southeast Asia (Indonesia, Malaysia, and Thailand). The results showed that across these two samples, general
efficacy had a significant positive relationship with organizational commitment and a significant negative relationship with
intention to turnover. Further analysis also indicated that job satisfaction mediated the relationship between general efficacy and
organizational commitment and intention to quit in the U.S. sample. The relationship between general efficacy and organizational
commitment was stronger in the U.S. than in the three combined countries sampled in Southeast Asia.
# 2005 Elsevier Inc. All rights reserved.
www.socscinet.com/bam/jwb
Journal of World Business 41 (2006) 121–132
Over the years, considerable research, summarized in
meta-analytic reviews, have clearly demonstrated that a
significant relationship exists between various psycho-
logical capacities, such as Big Five personality traits
(Barrick & Mount, 1991), self-evaluation traits (Judge &
Bono, 2001), specific self-efficacy (Stajkovic & Luthans,
1998a), and desirable employee work attitudes and
performance. But for a few exceptions (e.g., Born &
Iwawaki, 1997; Markus & Kitayama, 1991), this
relationship has not been tested to see if it generalizes
across cultures. In addition, neither the complexity nor
theoretical richness of the relationship has been tested for
possible mediators. Thus, the two-fold purpose of this
study was to begin to fill these gaps by first examining
whether U.S. and Southeast Asian employees’ job
satisfaction mediated the relationship between general
self-efficacy and work attitudes (organizational commit-
ment and turnover intention). Secondly, we examined
whether the relationship between general self-efficacy
and employee attitudes/outcomes, in terms of organiza-
* Corresponding author. Tel.: +1 402 472 2324;
fax: +1 402 472 5855.
E-mail address: fluthans1@unl.edu (F. Luthans).
1090-9516/$ – see front matter # 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jwb.2005.09.003
tional commitment and turnover intentions, differs
between U.S. and Southeast Asian samples.
1. Theoretical foundation and hypotheses
The theoretical foundation for self-efficacy is closely
associated with the extensive work of Albert Bandura
(1997). His conception of self-efficacy represents a
person’s belief and confidence in his or her own
capability to perform a specific task in a specific context.
This self-efficacy influences a person’s choice of
activities, avoiding tasks judged too difficult or
impossible, and embracing activities deemed manage-
able. Taken to the workplace, Stajkovic and Luthans
(1998b: 66) defined self-efficacy as ‘‘an individual’s
conviction (or confidence) about his or her abilities to
mobilize the motivation, cognitive resources, and courses
of action needed to successfully execute a specific task
within a given context’’. Bandura (1997) and others have
demonstrated a significant positive relationship between
this specific self-efficacy and various motivational,
affective, and behavioral outcomes in clinical (e.g.,
Bandura, Adams, Hardy, & Howells, 1980; Locke,
Frederick, Lee, & Bobko, 1984), educational (e.g., Lent,
F. Luthans et al. / Journal of World Business 41 (2006) 121–132122
Brown, & Larkin, 1984; Schunk, 1995), and organiza-
tional settings (e.g., Bandura, 1997, 2000; Wood &
Bandura, 1989). Stajkovic and Luthans (1998a) in a
meta-analysis of 114 studies found a strong (.38)
relationship between specific self-efficacy and work-
related performance outcomes.
Theory-building and research on self-efficacy has
been primarily contextually specific. Some organiza-
tional behavior researchers, however, have conceptua-
lized and have provided research for a generalized sense
of self-efficacy (Eden & Zuk, 1995; Judge & Bono,
2001; Judge, Erez, Bono, & Thoresen, 2003; Sherer
et al., 1982). An argument for general self-efficacy is
based on an individual’s past experience with success or
failure that results in a general set of expectations that
will affect his or her expectations of success across a
broad array of new and specific challenges/situations.
Eden and Zuk (1995: 629) defined general efficacy as
‘‘a generalized trait consisting of one’s overall estimate of
one’s ability to effect requisite performances in
achievement situations’’. For example, Sherer et al.
(1982) indicated that individuals with historical and
continuous successes invarious situations are more likely
to have a positive general self-efficacy in a greater variety
of situations than those people with less successful
experiences. General self-efficacy represents a global
sense of confidence in one’s coping ability across a wide
range of demanding or difficult situations and reflects a
broad and stable confidence in dealing effectively with
rather stressful situations (Judge & Bono, 2001).
There are only a few studies that have examined the
influence of general self-efficacy on employee work
attitudes (e.g., Judge et al., 2003). In particular, Judge
and Bono (2001) indicated that general efficacy, a
person’s belief he/she can conduct a task across
different situations and contexts, is a major contributor
to a positive core evaluation trait. General self-efficacy
influences a person’s choice of action, level of effort,
perseverance and resilience in the face of difficulties
and obstacles, adversity, failures and the characteristics
of individual thoughts, which could be self-hindering or
self-aiding (Judge et al., 2003; Wood & Bandura, 1989).
Sherer et al. (1982) indicated that general self-
efficacy is made up of three components—initiative,
effort and persistence. These components help deter-
mine how much effort people will expend on an activity,
how long they will persevere when confronting
obstacles, and how resilient they will be in the face
of adverse situations. Although Bandura (1997: 42)
argues the prevailing conceptual view that an ‘‘efficacy
belief is not a decontextualized trait’’, there is never-
theless increasing evidence of a strong positive
relationship between specific self-efficacy and the more
trait-like general efficacy (Judge et al., 2003; Leganger,
Kraft, & Roysamb, 2000; Wang, 2000).
1.1. Impact of general self-efficacy on employee
attitudes
General self-efficacy (Eden & Zuk, 1995; Judge &
Bono, 2001) has been shown to have a positive
relationship with work-related performance. For exam-
ple, in a meta-analysis, Judge and Bono (2001) found a
.23 relationship between general self-efficacy and job
performance. General efficacy is also expected to be
related to job satisfaction and turnover because of the
significant positive relationship reported between job
satisfaction and performance (Harter, Schmidt, & Hayes,
2002; Judge, Thoresen, Bono, & Patton, 2001). For
example, in a meta-analysiss of 312 samples, Judge et al.
(2001) found the mean true correlation to be .30.
Judge and Bono (2001) and Herold (2000) have
argued that general self-efficacy would affect job
satisfaction through its association with practical
success on the job. Individuals with high self-efficacy
deal more effectively with difficulties and persist in the
face of failure and obstacles (Gist & Mitchell, 1992).
Moreover, such individuals are more likely to attain
valued outcomes through persistence and thus derive
intrinsic satisfaction from their jobs. It then follows that
those with higher general self-efficacy are more likely
to be satisfied with their job. This was shown in the
Judge and Bono (2001) meta-analysis that reported a .45
relationship between generalized self-efficacy and job
satisfaction. Other studies have shown that efficacy is
positively related to organizational commitment (Wer-
bel, Landau, & DeCarlo, 1996) and overcoming burnout
(Cordes & Dougherty, 1993).
1.2. Mediating role of job satisfaction
A major purpose of this study is to examine whether
job satisfaction mediates the relationship between
employees’ general self-efficacy and their work attitudes
across cultures. This research question recognizes the
theoretical complexity and richness of the relationship
between general-efficacy and employee attitudes. We
argue that people with higher general self-efficacy are
more likely to be satisfied with their job because they
understand their job is making a significant and
meaningful contribution to their organization and to
their own success. Therefore, those with high efficacy
and job satisfaction are more likely to be committed to the
organization and have a lower intention to turnover.
F. Luthans et al. / Journal of World Business 41 (2006) 121–132 123
Considerable theory and research strongly supports
a negative relationship between job satisfaction and
turnover intentions (Cote & Morgan, 2002). In
general, employees with high job satisfaction are
more likely to enjoy their current position and stay in
the same organization. On the other hand, those with
low job satisfaction are more likely to search for
opportunities to leave their present organization. We
would expect to find a negative relationship between
an employee’s level of job satisfaction and his/her
intention to quit.
Based on the above background and shown in our
proposed conceptual model in Fig. 1, the following
hypotheses were developed for this study:
Hypothesis 1. The positive relationship between gen-
eral self-efficacy and organizational commitment is
mediated by job satisfaction.
Hypothesis 2. The negative relationship between gen-
eral self-efficacy and turnover intention is mediated by
job satisfaction.
1.3. Moderating role of the cultural dimension of
individualism/collectivism
The cultural dimension of individualism/collecti-
vism is recognized to be a major dimension to
Fig. 1. General self-efficacy and work attitudes: mediating role of job satisf
dimension.
differentiate between East and West (Hofstede, 1980,
2001, 2003). There is considerable support that
collectivism/individualism is an effective way to
analyze cultures. It has been found to be a coherent,
integrated, and empirically testable dimension of
cultures (Chiu & Kosinski, 1999). Hofstede (1980,
2001, 2003) argued that individualistic societies are
characterized by person-centered conscientiousness,
autonomy and initiative. He found the United States to
be ranked highest on individualism, whereas Southeast
Asian countries such as Indonesia, Malaysia, and
Thailand have been found to be low on individualism
and therefore high on group-centered collectivism
(Hofstede, 1980, 2001, 2003).
Chan, Gelfand, Triandis, and Tzeng (1996) pointed
out that, in collectivist societies, especially Southeast
Asia, individuals adhere to basic Confucianism and
believe common interests and social harmony are more
important than individual interests and enjoyment. At
the individual level, people in the collective societies
are more socio-centric and have an interdependent view
of the self (Markus & Kitayama, 1991). It is suggested
that collectivists are more inclined to respond to
situations and restraint of their own intentions for the
greater welfare of the group.
Individualistic societies, on the other hand, are
egocentric, autonomous, separate, self-contained and
action and moderating role of the individualism-collectivism cultural
F. Luthans et al. / Journal of World Business 41 (2006) 121–132124
independent (Markus & Kitayama, 1991). Bond (1991)
also posited that people in individualistic societies
emphasize the positive, exciting, and fulfilling aspects
of life. They hold on to notions that match their personal
beliefs and values at work. Trandis et al. (1986) pointed
out that individualistic cultures are more likely to
consider themselves as independent members of a
group. Therefore, it is expected that the individual
difference construct of self-efficacy would play a
more important role in determining individuals’ work
attitudes, such as organizational commitment and
turnover intention. More specifically, it would seem
to follow that the impact of general self-efficacy on
organizational commitment and turnover intention is
stronger in individualistic societies than in collectivistic
societies. This is because people in collectivistic
societies are more likely to be influenced by group
values and beliefs while people in individualistic
cultures tend to make decisions based more on their
own thinking and mindsets.
Based on the above background, the following
hypotheses were developed for the study.
Hypothesis 3. The positive relationship between gen-
eral self-efficacy and organizational commitment is
stronger among more individualistic U.S. employees
than among more collectivist Southeast Asian employ-
ees.
Hypothesis 4. The negative relationship between gen-
eral self-efficacy and turnover intention is stronger
among more individualistic U.S. employees than among
more collectivist Southeast Asian employees.
2. Methods
The sample for this study was drawn from a wide
variety of organizations in the U.S. and Southeast Asia.
Among the 888 respondents, 753 were from U.S., and
67 from Indonesia, 17 from Malaysia, and 51 from
Thailand. Although there are obviously many cultural
differences among these Southeast Asian countries,
using the Hofstede (1980, 2001), Trompenaars (1994),
and Ronen and Shenkar (1985) cultural guidelines on
the individualism-collectivism dimension for clustering
countries, we combined respondents from Indonesia,
Malaysia, and Thailand into one Southeast Asian
sample. For example, on Hofstede’s (2001, p. 215)
individualism index, for 50 countries Malaysia ranked
36th, Thailand tied for 39–41, and Indonesia tied for
47–48.
Besides this similarity on the individualism-collec-
tivism cultural dimension, we also conducted one-way
ANOVA analysis and found that there was no statistical
difference in the general self-efficacy among the
respondents from these three Southeast Asian countries.
Furthermore, we did not find any significant difference
in industry type, years of education, and gender among
these three Asiain countries. This analysis provides
support for our aggregating the data from these three
Southeastern Asian countries. In terms of handling
missing data, we used the mean substitution method.
This method was used because the total number of
missing cases was fewer than 5% and the listwise-
deletion method has been shown to generate biased
results (Noh, Kwak, & Han, 2004).
For those from the U.S., as to the type of ownership,
47.2% came from private firms and 52.8% were from
public firms. The U.S. demographic profile indicated
that 35.4% were managers, 42.5% were female, and
70.2% had a college degree. For those from Southeast
Asia, as to the type of ownership, 45.4% were from
private firms and 54.6% were from public organiza-
tions. The demographic profile for the Southeast Asian
respondents indicated that 32.4% were managers,
37.42% were female, and 75.2% had a college degree.
In other words, the demographics between the U.S.
sample and the Southeast Asian sample were similar.
2.1. Measures
To minimize the language and cultural differences on
the widely used standardized questionnaires as much as
possible, Earley’s (1994) guidelines for re-translation/
back translation were used. For each of the Southeast
Asians sampled, a bilingual native speaker from each
country (Indonesia, Malaysia, and Thailand) translated
the English questionnaires into their language. Then this
translated questionnaire was re-translated back to
English by another bilingual expert. The original and
the re-translated versions of the general self-efficacy,
job satisfaction, turnover intention, and organizational
commitment questionnaires were then carefully com-
pared and any discrepancies were reconciled.
2.1.1. General self-efficacy
The general self-efficacy scale was drawn from
Sherer et al. (1982). There are 17 items in this scale.
Respondents used a 14-point Likert scale (1 ‘‘strongly
disagree’’ to 14 ‘‘strongly agree’’) to indicate their level
of agreement with these items. Sample items include:
‘‘When I make plans, I am certain I can make them
work’’ and ‘‘If I can not do a job the first time, I keep
trying until I can’’. Item ratings were averaged to form
overall scores for individuals’ general self-efficacy. The
F. Luthans et al. / Journal of World Business 41 (2006) 121–132 125
Cronbach alpha was .83 for the U.S. sample and .78 for
the Southeast Asian sample.
2.1.2. Job satisfaction
The job satisfaction scale was drawn from Hackman
and Oldham (1980). Three items were used. Respon-
dents used a 7-point Likert scale (1‘‘strongly disagree’’
to 7 ‘‘strongly agree’’) to indicate their level of
agreement with these items. Items included: ‘‘Generally
speaking, I am very satisfied with my job’’, ‘‘I am
generally satisfied with the feeling of worthwhile
accomplishment I get from doing this job’’, and ‘‘I am
generally satisfied with the kind of work I do in this
job’’. We averaged item ratings to form overall scores
for the respondent’s job satisfaction. The Cronbach
alpha was .90 for the U.S. sample and .86 for the
Southeast Asian sample.
2.1.3. Turnover intention
The turnover intention scale was drawn from Hom,
Griffeth, and Sellaro (1984). Three items were used.
Respondents used a 7-point Likert scale (1 ‘‘strongly
disagree’’ to 7 ‘‘strongly agree’’) to indicate their level
of agreement. Items included: ‘‘I am planning to search
for a new job outside this organization during the next
12 months’’, ‘‘I often think about quitting this job’’, and
‘‘If I have my own way, I will be working for some other
organization one year from now’’. Item ratings were
averaged to form overall scores for the respondent’s
turnover intention. The Cronbach alpha was .79 for the
U.S. sample and .95 for the Southeast Asian sample.
2.1.4. Organizational commitment
Level of organizational commitment was measured
by Mowday, Porter, and Steers’ (1982) 15-item scale.
This scale measures three basic components of
organizational commitment: identification, involve-
ment, and loyalty. Sample items included: ‘‘I am quite
proud to be able to tell people that I am part of this
organization’’; ‘‘I find that my values and the
organization’s values are very similar’’; and ‘‘I really
care about the fate of this organization’’. Ratings were
completed on a 5-point Likert scale. The ratings of all
15 items were averaged to form a single index of
organizational commitment. The Cronbach alpha was
.89 for the U.S. sample and .81 for the Southeast Asian
sample.
2.2. Control variables
Six control variables were used to account for a
variety of demographical and contextual factors that
might affect the relationship between the independent
variable (general self-efficacy) and the dependent work
attitude variables (job satisfaction, organizational
commitment, and turnover intentions). They were: (1)
months of work experience, (2) months of experience in
the present organization, (3) months of experience in the
current position, (4) age, (5) gender, and (6) years of
education.
2.3. Common source/rater effects
To analyze the potential of a common source/rater
effect, we pursued two strategies. First, we used
Harman’s one factor test (Podsakoff, MacKenzie, Lee,
& Podsakoff, 2003) to examine the extent to which a
common or single method factor existed that would
account for the variance in our findings. The premise
of this strategy is that either (a) a single factor will
emerge from exploratory analysis or (b) one general
factor will explain most of the variance among
all scales (Podsakoff et al., 2003). We did an
exploratory factor analysis by entering all the scales
used in this study and nine factors emerged.
The general factor explained 22.56% of the total
variance, which does not provide support for a
common method/source bias explanation for the
results of this study.
For the second strategy, as suggested by Podsakoff
et al. (2003), we deliberately included in our survey a
scale that was not compatible with the theoretical
framework of our study and thus not expected to relate
to the outcomes. This 16-item outside scale was
tolerance for ambiguity (Budner, 1962) with sample
questions such as: ‘‘What we are used to is always
preferable to what is unfamiliar?’’ and ‘‘Many of our
most important decisions are based on insufficient
information’’. The Cronbach alpha was .71 for the U.S.
sample and .70 for the Southeast Asian sample.
We included the tolerance for ambiguity scale in
order to test for discriminant validity in a manner
similar to the method proposed by Lindell and Whitney
(2001). Our study focused on four core constructs:
general self-efficacy, job satisfaction, organizational
commitment, and turnover intention. A fifth construct,
outside our theoretical framework, was tolerance for
ambiguity. The data from this scale was regressed on the
various independent, mediating and dependent vari-
ables as shown in Fig. 1. If there was evidence of
common method and source/rater bias effects, the
coefficients involving tolerance for ambiguity as well as
those involving general self-efficacy should be similarly
affected (Podsakoff et al., 2003).
F. Luthans et al. / Journal of World Business 41 (2006) 121–132126
We did find the coefficients for general self-efficacy
and its relationship to the dependent outcome variables
were significant, but none of the corresponding
coefficients when using the outside scale of tolerance
for ambiguity were significantly related to the vari-
ous outcome measures. Specifically, the coefficients
between tolerance for ambiguity and organizational
commitment were: U.S.: b = .03, p > .10; Southeast
Asia: b = �.14, p > .05; turnover intention: U.S.:
b = .01, p > .10; Southeast Asia: b = �.12, p > .05;
and job satisfaction: U.S.: b = .06, p > .05; Southeast
Asia: b = �.14, p > .05. Therefore, the results of both
of these strategies indicate that common method and
source/rater effects do not seem to have an impact on the
pattern of results reported in this study.
3. Results
Table 1 provides descriptive statistics and correla-
tions for all study variables. General self-efficacy was
significantly and positively related to job satisfaction in
the U.S. sample (b = .32, p < .01), but was not
significantly related in the Southeast Asia sample
(b = .15, p < .10). General self-efficacy was positively
related to organizational commitment (U.S.: b = .24,
p < .01; Southeast Asia: b = .26, p < .01), and nega-
Table 1
Descriptive statistics and correlations for study variables (U.S. Sample, N
Variable Mean S.D. 1 2 3
1. Months of
total working
experience
131.61 (124.43) 92.35 (45.63) 1.00 .56**
2. Months of
experience in
this organization
57.99 (67.98) 56.10 (41.99) .55** 1.00
3. Months of
experience in
present position
32.65 (30.23) 36.25 (16.46) .41** .64** 1
4. Age 34.33 (34.83) 8.47 (4.59) .82** .50**
5. Gender .57 (.63) .49 (.49) .07** .04
6. Years of
education
16.38 (17.05) 1.80 (1.36) �.09** �.04 �
7. General
self-efficacy
10.54 (10.47) 1.49 (1.46) .04 .04
8. Job
satisfaction
5.06 (5.30) 1.23 (1.02) .09** .06
9. Turnover
intention
3.14 (2.41) 1.87 (1.58) �.14** �.07* �
10. Organizational
commitment
4.28 (4.73) .97 (.73) .07* .09*
Values in parentheses of means and S.D. are the Southeast Asian sample.* p < .05.
** p < .01.
tively related to turnover intentions (U.S.: b = �.17;
p < .01; Southeast Asia: b = �.26, p < .01). Of the
control variables, general self-efficacy had no relation-
ship with any of the demographic or contextual
variables.
We used hierarchical regression models to examine
the hypothesized direct effects and generated a series of
successive models to determine the added value of
each step. A separate series of regression models was
generated to examine the relationship of general self-
efficacy with job satisfaction and turnover intention.
Before interpreting the full-sample equation, we
examined the data for any possible violations of
assumptions for conducting ordinary least squares
(OLS) regression. The result indicated that all of the
assumptions, i.e., normality of distribution and homo-
geneity of variance, were not violated. Control
variables, including months of work experience,
months of experience in the present organization,
months of experience in the current position, age,
gender, and years of education, were controlled for in
the equation.
To test for mediation, we used the following four-
step procedure outlined by Baron and Kenny (1986): (1)
we tested whether there was a significant, positive
relationship between general self-efficacy and job
= 753; Southeast Asian, N = 135)
4 5 6 7 8 9 10
.11 .83** .16 �.05 �.02 .09 �.05 .04
.05 .47** .07 .17 .01 .01 �.01 .01
.00 .14 .19* �.12 .06 .24** .05 �.11
.37** 1.00 .18* .09 �.13 .10 �.04 �.04
.01 .10** 1.00 .02 .09 .41** .05 .09
.07 �.05 .20** 1.00 �.24** �.16 .13 �.04
.03 .02 .05 .02 1.00 .15 �.26** .26**
.06 .11* .03 .01 .32** 1.00 �.13 .33**
.08* �.15** �.03 .08* �.17** �.58** 1.00 �.41**
.06 .06 .02 .01 .24** .65** �.60** 1.00
Correlations above diagonal are the Southeast Asian sample.
F. Luthans et al. / Journal of World Business 41 (2006) 121–132 127
satisfaction, (2) general self-efficacy and organizational
commitment, and turnover intention, and (3) job
satisfaction and organizational commitment, and turn-
over intention. We then proceeded to step (4) by
controlling for job satisfaction to determine whether the
relationships between general self-efficacy and the
work-related attitudes of organizational commitment
and turnover intention were reduced to non-significance
or became significantly smaller. When the relationships
between general self-efficacy and organizational com-
mitment, and with turnover intention, become insig-
nificant or less significant, there is evidence for full or
partial mediation (Baron & Kenny, 1986). Results of
these analyses are presented in Table 2 for both the U.S.
sample and the Southeast Asian sample.
Controlling for work experience, tenure in one’s
present organization, experience in one’s current posi-
tion, age, gender, and years of education, we found that
job satisfaction mediates the relationship between
general self-efficacy and organizational commitment
in the U.S. sample. Step 1 in Table 2 shows that general
self-efficacy is significantly related to job satisfaction in
the U.S. (b = .32, p < .01), but not in Southeast Asia
Table 2
Results of hierarchical regression analyses (U.S. Sample, N = 753; Southea
Model
1 2
Job satisfaction Commitment Turnover
intention
Control variables
Age .11 (.05) .01 (�.15) �.11 (�.14)
Gender .01 (.38**) �.01 (.09) �.03 (.07)
Education
(years)
.01 (�.15) .01 (.04) .08 (.09)
Months of total
work experience
�.02 (�.01) .02 (.20) �.04 (.04)
Months of experience
in this organization
�.02 (�.03) .06 (�.05) .03 (.02)
Months of experience
in this position
.02 (.15) .01 (�.14) �.03 (.07)
R2 .01** (.17**) .01** (.01**) .02** (.01**)
General self-efficacy .32** (.09) .24** (.26**) �.15** (�.26**
DR2 .10** (.01**) .05** (.05**) .03** (.04**)
Job satisfaction
General self-efficacy
� cultural dimension
(1 = U.S., 0 = Asia)
.22** �.15**
DR2
Values in parentheses are the Southeast Asian sample.* p < .05.
** p < .01.*** p .05). Model 2 shows that general self-
efficacy is significantly related to organizational com-
mitment (U.S.: b = .24, p < .01; Southeast Asia: b = .26,
p < .01), and turnover intention (U.S.: b = �.15,
p < .01; Southeast Asia: b = �.26, p < .01). Model 3
shows that job satisfaction is significantly related to
organizational commitment (U.S.: b = .65, p < .01;
Southeast Asia: b = .40; p < .05), and negatively with
turnover intention in U.S. (b = �.57, p < .01), but was
not significantly related in the Southeast Asia sample
(b = �.18, p .05), and with
turnover intention (U.S.: b = .02, p > .05), are no longer
significant or decreased in significance. These results
indicate that, with the U.S. sample, job satisfaction
mediates the relationship between general self-efficacy
and all of the various work attitudes. Thus, both
hypotheses 1 and 2 were supported for the U.S. sample.
In order to examine the moderating role of the
culture over the relationships between general self-
efficacy and organizational commitment and with
turnover intention, we conducted separate analyses
by creating an interaction term between general self-
efficacy and cultural grouping based on individualism-
collectivism (i.e., U.S. and Southeast Asian). Results in
Table 2 show that this culture moderated the relation-
ship between general self-efficacy and organizational
commitment (b = .22, p < .01), providing support for
Hypothesis 3.
In summary, our results provided partial support for
Hypothesis 1 in that job satisfaction mediates the
positive relationship between general self-efficacy
and organizational commitment in the U.S. sample.
Hypothesis 2, that job satisfaction mediates the negative
relationship between general self-efficacy and turnover
intention was also supported in the U.S. sample.
Hypothesis 3, which states that the positive relationship
between general self-efficacy and organizational com-
mitment is stronger among U.S. employees than among
a sample of employees from three Southeast Asian
countries, was supported. Hypothesis 4, the negative
relationship between general self-efficacy and turnover
intention is stronger for U.S. employees than for the
Southeast Asian employees, was not supported. Inter-
estingly, contrary to our hypothesis, we found that
general self-efficacy has a more negative effect on
turnover intention in Southeast Asia than in the U.S.
4. Discussion
Results of this study were consistent with previous
U.S.-based research that has shown employees with
higher levels of general self-efficacy have more positive
work-related attitudes (Cote & Morgan, 2002; Judge &
Bono, 2001; Judge et al., 2003; Sherer et al., 1982).
Besides replicating earlier U.S.-based findings, the
current study makes an important extension to selected
Southeast Asian countries. With both U.S. and a sample
of employees from three Southeast Asian countries, we
found the higher their general self-efficacy, the greater
their organizational commitment, as well as the lower
their intention to turnover. We also found that general
self-efficacy was positively related to job satisfaction in
the Asian sample, however this relationship was not
significant.
In this study, we have also begun to explore the
theoretical complexity and richness of how efficacy
may affect work-related attitudes. We found U.S.
employees’ general self-efficacy may be mediated in its
relationship with work-related attitudes through job
satisfaction. Specifically, we found that general self-
efficacy leads to job satisfaction, and that job
satisfaction leads to commitment to the organization
for employees in the U.S. We also found that general
self-efficacy was significantly related to work commit-
ment and negatively related to turnover intention in the
selected Southeast Asian countries. These results
suggest that U.S. and Southeast Asian countries
employee differences in organizational commitment
and employees’ intentions to turnover may be explained
at least in part by the differences in how satisfied
employees are with respect to their job and work
environment.
A possible explanation for the Southeast Asian
employees’ job satisfaction not mediating the relation-
ship between efficacy and turnover intention may be
due to their recent history of economic uncertainty (i.e.,
the late 1990s financial crisis in Southeast Asia) and the
collectivist cultural value that promote having a more
substantial reason for quitting their jobs than simply
being dissatisfied with their work. Another possible
explanation is that Southeast Asian companies in
general tend to have a relatively more permanent
employment relationship (i.e., carry-over from life-time
employment days) with their employees, so Southeast
Asian employees may be less likely to consider quitting
as an option. Moreover, it may still be more difficult to
switch from company to company in Southeast Asian
cultures due to the importance of perceived loyalty in
the organization and employee relationship. However,
we did find the Southeast Asian employees (and also
those in the U.S.) who were more satisfied with their job
are more likely to have higher commitment to their
organization and still have a lower intention to leave
their organization. Although this latter finding is
consistent with relevant prior research (e.g., Cote &
Morgan, 2002), this current study also indicated
the importance job satisfaction seems to have across
countries with a different cultural dimension of
individualism (U.S.) and collectivism (Southeast Asia).
As to the moderating role of individualism-collecti-
vism, we found that general self-efficacy may have a
stronger effect on organizational commitment in the
U.S. than in Southeast Asian countries. This provides
some initial empirical support to cross-cultural theories
F. Luthans et al. / Journal of World Business 41 (2006) 121–132 129
(Markus & Kitayama, 1991; Trandis et al., 1986) using
the individualism-collectivism construct. These the-
ories suggest that people in individualistic societies are
egocentric, autonomous, self-contained, and indepen-
dent and that such individual level factors play a more
important role in influencing employees’ work attitudes
in Western cultures such as the U.S. than in Southeast
Asian cultures.
However, contrary to our hypothesis, we did not find
that general efficacy has a significantly different
relationship with turnover intention between the U.S.
and our Southeast Asian samples. One possible
explanation for this non-significant difference is that
general self-efficacy has such an important role in
deciding an individual’s intention to leave the organiza-
tion. In other words, it may be that these sampled
Southeast Asian employees’ general self-efficacy will
promote their staying in their organization since they
are more likely to think that they will be successful in
their current organization. Furthermore, it is also
possible that low organizational commitment does
not necessarily mean high turnover intention. Indivi-
duals with low commitment may need to consider
whether they have the confidence to find a new job
before they have the intention to leave and the
opportunity available to move to a different organiza-
tion.
4.1. Implications for theory and practice
The study has at least two important implications for
the theory-building of general self-efficacy. First, the
study empirically tested the potential mediation of job
satisfaction and second, the study contributed at least
some initial evidence of cross-cultural generalization.
The results indicated that a personality-like ‘‘self-
construct’’, U.S. and Southeast Asian employees’
general self-efficacy, may be a key psychological
strength for increasing levels of organizational commit-
ment and reducing intention to leave. Testing our model
in two distinct (U.S. and Southeast Asian) cultural
contexts (i.e., individualism versus collectivism) per-
mits beginning cross-validation, and tests of general-
izability of the theoretical links. For the future, these
preliminary results may help contribute to theoretical
understanding of the context specificity versus general-
izability of the impact of efficacy on work attitudes.
Also, the study provides additional evidence to support
and expand Judge et al. (1997, 2003) theory of core self-
evaluation in that it examines the relationships of one’s
general sense of self-efficacy to organizational commit-
ment and turnover intention across cultures.
Finally, the current study may help complement and
augment other lines of theory development besides
general self-efficacy and job satisfaction. For example,
since efficacy judgments constitute a self-regulatory
concept inherent in motivational processes and resulting
performance (Bandura, 1997; Judge & Bono, 2001;
Judge et al., 2003; Stajkovic & Luthans, 1998a), the
notion of general self-efficacy may also contribute to
understanding the psychological strengths that result
in higher levels of organizational commitment and
employee retention.
Besides theory-building, our study results also have
several implications for practice. The positive relation-
ship found between general self-efficacy and organiza-
tional commitment and the negative relationship with
intention to quit suggests that managers in general, and
human resource professionals in particular, may need
to give greater attention to the development of general
self-efficacy in their employees. For example, organi-
zations might be able to test for higher levels of general
self-efficacy prior to selecting employees, as well as
managers developing it in their associates over time
(Luthans & Youssef, 2004). Specific programs (e.g.,
successful experience, modeling, coaching and posi-
tive feedback) could be designed to help employees
enhance their general efficacy (Luthans & Youssef,
2004; Stajkovic & Luthans, 1998b). The main goal of
such efficacy development programs would not
necessarily be to train new skills, but to enhance
employees’ beliefs and confidence that they can do well
with their current skills, knowledge and strategies
(Bandura, 1997). A secondary benefit of working to
enhance general efficacy may involve increasing
employees’ overall organizational citizenship behavior
(Organ & Lingl, 1995), as they feel more confident in
broadly contributing to the overall mission of their
organization.
4.2. Limitations and suggestions for future research
A major limitation of the current study involved the
collection of survey data from a single source using a
common method. The similarity of methods used to
measure both independent and dependent variables may
have inflated the observed relationships between
general self-efficacy and the work attitude variables.
In our study, we took the perspective of method as being
a combination of respondent (or rater or observer),
instrument, time, and place (see Avolio, Yammarino, &
Bass, 1991 for a similar view of ‘‘method’’). If two
measures share respondent, instrument, time, and place,
we could expect these measures to be positively
F. Luthans et al. / Journal of World Business 41 (2006) 121–132130
correlated with each other in the absence of a
substantive relationship.
Our analysis concerning both the one factor and
conceptually unrelated construct tests for single source
effects and bias reported in our methods section
indicated that this potential limitation may not have
had a significant impact on our findings. Furthermore,
the fact that our independent variable (general self-
efficacy) was substantially different from our dependent
outcome variables (e.g., turnover intention) should also
mitigate some of the inflation effects associated with
single source bias/effects. Nevertheless, it would be
useful if future research could examine the relationships
between the general self-efficacy and work/psycholo-
gical outcomes with data collected over time, especially
using different methods.
A second limitation related to the use of a cross-
sectional design for data collection was whether general
self-efficacy indeed has a direct impact on job
satisfaction, organizational commitment, and intention
to turnover, as well as an indirect effect through job
satisfaction. Because this study design does not allow
causal conclusions, it is likely that individuals that
are satisfied with their jobs may also have higher levels
of general self-efficacy, potentially contributing to the
mediational effects reported in our results. Moreover,
more satisfied employees with correspondingly higher
job performance may also report higher levels of
general self-efficacy due to establishing mastery experi-
ence and receiving positive feedback and reinforce-
ment.
Still another limitation is that we did not have large
sample sizes from Southeast Asia countries and thus
had to group them into one sample by using the cultural
criterion of individualism-collectivism. There are many
other cultural dimensions, but we are limited in this
study to only individualism-collectivism and not a more
complete cultural comparison study. However, we were
able to confidently group these three Southeast Asian
countries according to their individualism-collectivism
cultural dimension and this dimension is conceptually
most relevant to general self-efficacy. This efficacy was
also found to be statistically similar in our three
Southeast Asian countries in the study.
To have a better understanding of the various
relationships among general self-efficacy, job satisfac-
tion, organizational commitment, and turnover inten-
tion, we recommend that future research focus on the
following: (1) examine these relationships over time
and/or by first intervening to enhance general self-
efficacy and then measuring its effects over time; (2)
incorporate additional cultural dimensions beyond
individual/collective variables, to get a better under-
standing of how broader cultures may moderate the
relationships observed in the current study; and (3) test
other potential mediators such as task complexity or
emerging positive organizational behavior states such
as hope or resiliency (Luthans & Youssef, 2004).
5. Conclusion
The primary goal of this study was to examine
whether general self-efficacy was related to employees’
work attitudes across countries with a different cultural
dimension of individualism-collectivism. The second
goal was to better understand the richness of the impact
of efficacy by testing mediators (job satisfaction) and
moderators (U.S. and selected Southeast Asian coun-
tries). Other antecedents now need to be included in
future research to determine how variables such as
leadership affect general self-efficacy and how efficacy
in turn impacts outcome variables such as commitment,
turnover and especially work performance. Other
cultural dimensions and other countries need to be
studied as well. Future research also might explore
whether the antecedents to general self-efficacy are the
same as those of specific self-efficacy.
In conclusion, this exploratory study has not only
provided some additional support for the positive role
that general (as opposed to specific) efficacy may have
for desirable workplace attitudes/outcomes, but also
that some of these effects may hold across different
cultures. The study has also contributed to the
theoretical development that general efficacy may be
mediated by constructs such as job satisfaction in its
relation to workplace attitudes, at least in the U.S.
Acknowledgments
We would like to acknowledge the constructive
suggestions made by John Slocum on previous drafts of
this manuscript. We would also like to thank Alex
Stajkovic for his help in gathering the U.S. sample data.
References
Avolio, B. J., Yammarino, F. J., & Bass, B. M. (1991). Identifying
common methods variance with data collected from a single
source: An unresolved sticky issue. Journal of Management,
17: 571–588.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
Freeman.
Bandura, A. (2000). Cultivate self-efficacy for personal and organiza-
tional effectiveness. In E. A. Locke (Ed.), Handbook of principles
of organizational behavior (pp. 120–136). Oxford, UK: Blackwell.
F. Luthans et al. / Journal of World Business 41 (2006) 121–132 131
Bandura, A., Adams, N. E., Hardy, A. B., & Howells, G. H. (1980).
Test of the generality of self-efficacy theory. Cognitive Therapy
and Research, 4: 39–66.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator
variable distinction in social psychological research: Conceptual,
strategic, and statistical considerations. Journal of Personality and
Social Psychology, 51: 1173–1182.
Barrick, M. R., & Mount, M. K. (1991). The ‘‘Big Five’’ personality
dimensions and job performance: A meta-analysis. Personnel
Psychology, 44: 1–26.
Bond, M. H. (1991). Beyond the Chinese face: Insights from psychol-
ogy. Hong Kong: Oxford University Press.
Born, A., & Iwawaki, S. (1997). The assessment of optimistic self-
beliefs: Comparison of the Chinese, Indonesian, Japanese,
and Korean versions of the general self-efficacy scale. Psycho-
logia: An International Journal of Psychology in the Orient, 40:
1–13.
Budner, S. (1962). Intolerance of ambiguity as a personality variable.
Journal of Personality, 30: 29–50.
Chan, D. K. S., Gelfand, M. J., Triandis, H. C., & Tzeng, O. (1996).
Tightness-looseness revisited: Some preliminary analyses in Japan
and the United States. International Journal of Psychology, 31: 1–
12.
Chiu, R. K., & Kosinski, F. A. (1999). The role of affective disposi-
tions in job satisfaction and work strain: Comparing collectivist
and individualistic sccieties. International Journal of Psychology,
34: 19–28.
Cordes, C., & Dougherty, T. W. (1993). A review and an integration of
research on job burnout. Academy of Management Review, 18:
621–650.
Cote, S., & Morgan, L. (2002). A longitudinal analysis of the
association between emotional regulation, job satisfaction, and
intention to quit. Journal of Organizational Behavior, 23: 947–
962.
Earley, P. C. (1994). Self or group? Cultural effects of training on self-
efficacy and performance. Administrative Science Quarterly, 39:
89–117.
Eden, D., & Zuk, Y. (1995). Seasickness as a self-fulfilling prophecy:
Raising self-efficacy to boost performance at sea. Journal of
Applied Psychology, 80: 628–635.
Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical
analysis of its determinants and malleability. Academy of Manage-
ment Review, 17: 183–211.
Hackman, J. R., & Oldham, G. R. (1980). Work redesign. MA:
Addison-Wesley.
Harter, J. K., Schmidt, F. L., & Hayes, T. L. (2002). Business unit level
relationship between employee engagement and business out-
comes: A meta-analysis. Journal of Applied Psychology, 87:
268–280.
Herold, J. P. (2000). Nurse’s talk in relationship to their perceptions of
work satisfaction and self-efficacy. Unpublished dissertation. Col-
orado: University of Denver.
Hofstede, G. (1980). Culture’s consequences: International differ-
ences in work-ralated values. London: Sage Publications.
Hofstede, G. H. (2001). Culture’s consequences: Comparing values,
behaviors, institutions and organizations across nations. Thou-
sand Oaks, CA: Sage Publications.
Hofstede, G. (2003). What is culture? A reply to Baskerville. Account-
ing, Organizations & Society, 28: 811–814.
Hom, P. W., Griffeth, R. W., & Sellaro, C. L. (1984). The validity of
Mobley’s 1977 model of employees turnover. Organizational
Behavior and Human Performance, 34: 144–174.
Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evalua-
tions traits-self esteem, generalized self-efficacy, locus of control,
and emotional stability-with job satisfaction and job performance:
A meta-analysis. Journal of Applied Psychology, 86: 80–92.
Judge, T. A., Erez, A., Bono, J., & Thoresen, C. J. (2003). The core
self-evaluations scale: Development of a measure. Personnel
Psychology, 56: 303–331.
Judge, T. A., Locke, E. A., & Durham, C. C. (1997). The dispostional
causes of job satisfaction: A core evaluations approach. Research
in Organizational Behavior, 19: 151–188.
Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The
job satisfaction-job performance relationship. Psychological Bul-
letin, 127: 376–407.
Leganger, A., Kraft, P., & Roysamb, E. (2000). Perceived self-efficacy
in health behavior research: Conceptualization, measurement and
correlates. Psychology and Health, 15: 51–69.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1984). Self-efficacy in the
predictions of academic performance and perceived career
options. Journal of Counseling Psychology, 33: 265–269.
Lindell, M. K., & Whitney, D. J. (2001). Accounting for common
method variance in cross-sectional research designs. Journal of
Applied Psychology, 86: 114–121.
Locke, E. A., Frederick, E., Lee, C., & Bobko, P. (1984). Effect of self-
efficacy, goals, and task strategies on task performance. Journal of
Applied Psychology, 69: 241–251.
Luthans, F. (2002). The need for and meaning of positive organiza-
tional behavior. Journal of Organizational Behavior, 23: 695–706.
Luthans, F., & Youssef, C. M. (2004). Human, social, and now positive
psychological capital: Investing in people for competitive advan-
tage. Organizational Dynamics, 33: 143–160.
Markus, H. R., & Kitayama, S. (1991). Culture and self: Implications
for cognition, emotion, and motivation. Psychological Review, 98:
224–253.
Mowday, R. T., Porter, L. W., & Steers, R. M. (1982). Employee-
organization linkages. New York: Academic Press.
Noh, H., Kwak, M., & Han, I. (2004). Improving the prediction
performance of customer behavior through multiple imputation.
Intelligent Data Analysis, 8: 563–577.
Organ, D. W., & Lingl, A. (1995). Personality, satisfaction, and
organizational citizenship behavior. Journal of Social Psychology,
135: 339–350.
Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003).
Common method biases in behavioral research: A critical review
of the literature and recommended remedies. Journal of Applied
Psychology, 88: 879–903.
Ronen, S., & Shenkar, O. (1985). Clustering countries on attitudinal
dimensions. Academy of Management Journal, 28: 435–454.
Schunk, D. H. (1995). Self-efficacy and education and instruction. In
J. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: The-
ory, research, and application (pp. 281–303). New York: Plenum
Press.
Sherer, M., Maddux, J. E., Mercadante, B., Prentice-Dunn, S., Jacobs,
B., & Rodgers, R. W. (1982). The self-efficacy scale: Construction
and validation. Psychological Reports, 51: 663–671.
Stajkovic, A. D., & Luthans, F. (1998a). Self-efficacy and work-
related performance: A meta-analysis. Psychological Bulletin,
124: 240–261.
Stajkovic, A. D., & Luthans, F. (1998b). Going beyond traditional
motivational and behavioral approaches. Organizational
Dynamics, 26(4): 62–74.
Trandis, H. C., Bontempo, R., Betancourt, H., Bond, M., Leung, K.,
Brenes, A., et al. (1986). The measurement of etic aspects of
F. Luthans et al. / Journal of World Business 41 (2006) 121–132132
individualism and collectivism across cultures. Australian Journal
of Psychology, 38: 257–267.
Trompenaars, F. (1994). Riding the waves of culture. New York: Irwin.
Wang, C. (2000). Correlation among general self-efficacy, trait anxi-
ety, state anxiety, and test anxiety. Chinese Journal of Clinical
Psychology, 8: 229–230.
Werbel, J., Landau, J., & DeCarlo, T. E. (1996). The relationship of pre-
entry variables to early employment organizational commitment.
Journal of Personal Selling and Sales Management, 16: 25–36.
Wood, R., & Bandura, A. (1989). Social cognitive theory of organiza-
tional management. Academy of Management Review, 14: 361–
384.
The impact of efficacy on work attitudes across cultures
Theoretical foundation and hypotheses
Impact of general self-efficacy on employee attitudes
Mediating role of job satisfaction
Moderating role of the cultural dimension of individualism/collectivism
Methods
Measures
General self-efficacy
Job satisfaction
Turnover intention
Organizational commitment
Control variables
Common source/rater effects
Results
Discussion
Implications for theory and practice
Limitations and suggestions for future research
Conclusion
Acknowledgments
References
Is themeaning of ethical leadershipconstant across cultures? A test of
cross-culturalmeasurement invariance
Saima AhmadGraduate School of Business and Law, RMIT University, Melbourne, Australia
Syed Fazal-e-hasanAustralian Catholic University, Sydney, Australia, and
Ahmad KaleemCentral Queensland University, Melbourne, Australia
Abstract
Purpose – This paper empirically addresses the question of whether the meaning of ethical leadership isconstant across cultures. Drawing on the implicit leadership theory (ILT), we examine whether people inAustralia and Pakistan respond to perceived ethical leadership in a similar or different manner. By comparingemployees’ interpretation of the key attributes associated with ethical leadership, we advance construct-specific knowledge in cross-national contexts.Design/methodology/approach – Since meaningful cross-country comparisons of a research constructrequire an equivalent measurement of it, we examine the issue of cross-cultural measurement invariance ofethical leadership. Specifically, this study explores the configural, metric and scalar invariance of ethicalleadership by obtaining data from matched international samples.Findings –The findings broadly support cross-cultural generalisability of the construct’s meaning and cross-cultural transferability of the ethical leadership scale (ELS). They suggest that measures of ethical leadershipconstructs should be used in different cultures with caution because significant differences may exist at theitem level.Originality/value – This study provides cross-cultural endorsement to the construal of ethical leadership bypresenting evidence that supports convergence in the construct’s meaning across Eastern and Westerncultures. The study has enhanced the construct validity of ethical leadership through the use of the refinedmultiple-sample analytical approach. Previous studies have assumed that measures of ethical leadership areinvariant across various contexts. However, this is the first study to employ a robust methodological technique(metric and path invariance) that demonstrates the significant difference between each item and path andgeneralises the validity of ethical leadership construct and its measures by using international samples.
Keywords Pakistan, Australia, Ethical leadership, Cross-cultural research, Measurement invariance
Paper type Research paper
1. IntroductionThe rise of ethical failures in the international business environment (e.g., corporate scandalsof Enron and Arthur Anderson) has heightened the importance of ethical behaviourexemplified by organisational leadership. George (2003, p. 1) suggests: “The depth of your[Enron and Arthur Anderson] misconduct shocked the world and awakened us to the realitythat the business world was on thewrong track, worshipping the wrong idols, and headed forself-destruction”. Consequently, contemporary organisations are re-evaluating their strategicdiscernment along the lines of ethical leadership to enhance business sustainability. Ethicalleadership is recognised as a notable style of organisational leadership for the advancementof the common good through the virtues of integrity, care, compassion and justice (Grace,1998; Wilson and McCalman, 2017).
A widely applied framing of ethical leadership is attributed to the work of the USresearchers, Brown et al. (2005), who conceptualised such leaders as trustworthy and
Cross-culturalmeasurement
invariance
1323
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/0143-7720.htm
Received 18 February 2019Revised 17 December 2019
11 March 2020Accepted 3 April 2020
International Journal of ManpowerVol. 41 No. 8, 2020
pp. 1323-1340© Emerald Publishing Limited
0143-7720DOI 10.1108/IJM-02-2019-0079
legitimate role models acting in morally sound ways and conveying the importance of ethics.They also operationalised the measurement of such a leadership style by advancing anethical leadership scale (ELS) and validating it in US work settings. The existence andefficacy of ethical leadership in organisations are thus primarily determined by seekingfollowers’ perceptions through surveys based on ELS or other similar measures (see, e.g.Kalshoven et al., 2011).
The majority of available social-scientific empirical-descriptive evidence comes fromWestern countries and cultures and uses this approach. However, the international businessenvironment is characterised by a steady growth in the number of organisations thatoperates across Eastern and Western countries with stark cultural differences. The GLOBEstudy on Culture, Leadership and Organisations conveys an extensive account of suchdifferences and highlights that the behaviour of, and the preferences for, leadership areprofoundly shaped by culture (see also House et al., 2004). Accordingly, the current ethicalleadership theory may not be considered universal because it is likely to reflect the dominantculture where it has been conceptualised and validated (Zhu et al., 2019). Furthermore,previous research has shown that leadership practices in developing countries of the East donot fit theWesternmodels of leadership because of people’s differing perceptions of authorityand interpersonal relations (Hofstede et al., 2010; Houghton et al., 2014). This implies thatcultural differences can be reflected in the operationalisation and measurement of the ethicalleadership construct. Hence this warrants a greater understanding of whether themeaning ofethical leadership converges or diverges across two cultures (Martin et al., 2009). According toBrown et al. (2005, p. 131),
. . . cross-cultural research on ethical leadership could advance further with the availability of abroadermeasure of ethical leadership such as the one offered here [i.e., ELS]. Givenmultiple efforts todevelop global business ethics standards, it would be helpful to understand whether perceptions andmeasures of ethical leadership are similar across cultures.
While it is widely recognised that ethical leadership is important within both Western andEastern societies for eliciting positive work-outcomes (Ahmad, 2018; Resick et al., 2006),extant research is mute on clarifying its defining attributes and actions in the context ofcultural differences (Martin et al., 2009; Resick et al., 2011; Zhu et al., 2019). Zhu et al. (2019)have recently acknowledged the limitation thus: “significant gaps persist in the ethicalleadership literature. One gap is particularly salient: the lack of formal conceptualization andmeasurement of ethical leadership in Eastern culture as well as cross-culturally”.
Eastern countries provide significant contrasts to Western countries in terms ofprominent cultural dimensions. For example, Hofstede’s (2001) well-recognised and widelyapplied framing of cultural dimensions suggests that Asian countries score considerablyhigher on power distance and collectivism in comparison to countries in Europe, NorthAmerica and Oceania. Given the growing tendency for organisations to operate acrossEastern andWestern borders, both researchers and organisational policymakers are keen tounderstand whether the meaning of ethical leadership converges or diverges across cultures(Dickson et al., 2012; Resick et al., 2011; Zhu et al., 2019). Through this study, we aim to expandthe frontiers of ethical leadership research and theory by answering this question.Specifically, this research explores cultural differences in the operationalisation andmeasurement of ethical leadership by drawing on the implicit leadership theory (ILT) (Lordand Maher, 1991) and testing the extent to which culture shapes followers’ perceived ethicalleadership in organisational settings.
This paper proceeds as follows. It begins with an overview of the ethical leadershipliterature before delving into cross-cultural studies in the area. A description of the researchmethod, analysis of the research data and interpretation of the results follow. The paper
IJM41,8
1324
concludes with a discussion of the implications of research findings, limitations and aroadmap for future enquiry.
2. Literature review2.1 Ethical leadership: construct conceptualisation and measurementThe seminal research on ethical leadership comes from Brown et al. (2005), who defined it as“the demonstration of normatively appropriate conduct through personal actions andinterpersonal relationships, and the promotion of such conduct to followers through two-waycommunication, reinforcement, and decision-making” (p. 120). They argued that employeestend to emulate the ethical behaviour of their managers and leaders because the latter providecredible and legitimate sources of role modelling in the workplace. The status, power andlegitimate authority of leaders to administer rewards or punishment to followers make themeffective role models (Bandura, 1977; Brown et al., 2005). The underlying theoretical premiseof their research was based on Bandura’s (1977) seminal work on social learning theory.
Apart from providing the social learning-based theoretical framing for ethical leadership,Brown et al. (2005) also developed the ELS for measuring followers’ perceptions of it inorganisational settings. Researchers have demonstrated ethical leadership’s relationshipwith, and distinctiveness from, associated constructs in organisational research (e.g.,consideration leadership style, followers’ attitudes) through a “nomological network” thatincorporates such elements. Subsequent research, deploying this scale, reinforced the notionthat ethical leadership improves followers’ well-being and work-outcomes (e.g. Ahmad et al.,2018; Chughtai et al., 2015) and lowers their intentions to leave the job (Palanski et al., 2014).
The development of ELS followed robust methodological procedures such as the use ofmultiple samples and the assessment of the content, coverage, discriminant and nomologicaltypes of validity (see also Brown et al., 2005; Hinkin, 1998). The measure appeared valid,unidimensional and reliable, with high predictive power to explain a range of associatedvariables in organisational research. Brown et al. (2005) argued that ELS is neither linkedwith the respondents’ propensity to give socially desirable ripostes nor with theirdemographic profiles or common belief systems. Most notably, the measure is briefbecause it comprises only ten items, and it is beneficial for survey research from thestandpoint of respondents’ participation time (Brown et al., 2005; Hinkin, 1998). Themeasure’s conciseness, despite the breadth of the overarching ethical leadership construct, isa major strength.
Despite the promising developments mentioned above, a recent review of research onethical leadership suggests that
Further theory development is needed—for example, aroundwhat is unique to ethical leadership, therole of intentions, and the role of the context. Also, methodological challenges around measurement,validity, and research designs need tackling. (Den Hartog, 2015, p. 429)
Taking up this challenge, the role of cultural context and associated methodologicalchallenges around measurement are discussed in the following sections.
2.2 Ethical leadership in cross-cultural contextsCulture, generally defined in terms of norms, values and practices widely shared by peopleresidingwithin a nation-state, provides a valuable lens to understand acceptable behaviour inorganisations (Hofstede, 2001; House et al., 2004). It also predicts the effectiveness of andpreferences for organisational leadership in a country. For example, Daniels and Greguras(2014) have explained that leadership in countries with high power distance cultureseemingly provokes more imitation from followers because they are keener to follow theirleader’s actions as part of their cultural norm. A leadership style considered acceptable in one
Cross-culturalmeasurement
invariance
1325
culture may conflict with the accepted norms of another culture. This is because culturaldifferences have been associated with people’s sensitivity to ethical issues. For example,abusive supervision (defined as followers’ perceptions of the extent of their supervisor’ssustained display of unacceptable behaviour) is more prevalent in high than in low powerdistance cultures, mainly because the high power distance context provides a means toreinforce the supervisor’s stronger power in a society that readily accepts and permits strongpower differentials (Tepper, 2007).
Resick et al. (2006) initiated an early enquiry into understanding ethical leadership in thecontext of national cultural differences. By conducting an extensive review of both ethics andthe leadership literature, they identified six key attributes of ethical leaders: (1) character andintegrity, (2) ethical awareness, (3) community- or people-orientation, (4) motivating, (5)encouraging and empowering and (6) managing ethical accountability (see Appendix for adescription of the identified attributes). This perspective of ethical leadership differed fromthe conceptualization of Brown et al. (2005), explained earlier.Where the former focuses on thesocial learning perspective of ethical leadership, the latter maintains a focus on cognition andbehavioural epitomising of ethical leadership.
According to these researchers,
At the core of ethical leadership exists a cognitive component consisting of leaders’ values andknowledge (integrity, ethical awareness, and community/people-orientations) which then influencethe way leaders behave and use their social power (motivating, encouraging, and empoweringfollowers and holding people accountable). (Resick et al., 2006, p. 348)
To examine this notion, Resick et al. (2006) utilised secondary data from the GLOBE project.While the GLOBE project offered an in-depth account of cultures and leadership throughresearch conducted in 62 countries worldwide (see also House et al., 2004), it did notspecifically measure ethical leadership. By drawing on the project’s original questionnaireitems, Resick et al. derived a 15-item scale to measure ethical leadership. However, theresulting measure did not capture the elements of “ethical awareness” and “accountability”associatedwith ethical leadership. The authors acknowledged this limitation and argued thata coherent conceptualisation of ethical leadership would be incomplete without consideringthese. Nevertheless, analysis of the research data by Resick et al. (2006) revealed that four ofthe six perceived attributes of ethical leadership are universally endorsed (see Appendix).These include character and integrity, altruism, motivation and encouragement (see Resicket al., 2006 for details).
Subsequent research in the area demonstrated variations in the meaning of ethicalleadership in cross-cultural contexts. Most notably, a successive qualitative enquiry byResick et al. (2011) showed that, despite the universal endorsement of the key attributes ofethical leadership, followers’ level or extent of emphasis on each attribute varied acrosscultures. Specifically, the latter research revealed that leaders’ character and accountabilityare highly desired attributes in the US and Ireland, whereas consideration and respect forothers are highly preferred attributes in Hong Kong and China. The authors attributed thisdiscrepancy to differing cultural values. For example, the dominant attribute of theirparticipating managers from the US was character, which reflects the notion that “leadershipresides within the individual, who is considered to be a heroic figure” (p. 447). The researchersestablished that this notion is consistent with the popular culture of the US, even though the“great man theories of leadership have long been abandoned” (p. 447).
In a similar vein, research also supports cross-cultural variation in the effects of ethicalleadership on followers’ attitudinal and behavioural outcomes. For example, through across-cultural examination of the effects of ethical leadership on workplace bullying, Ahmad(2018) demonstrated some degree of variation in the perceived effects of ethical leadership inthe study’s respondents from Australia and Pakistan. She called for further research on
IJM41,8
1326
ethical leadership to obtain a culturally nuanced insight into a cross-cultural endorsement ofthe construct. To this end, scholars such as Vandenberg and Lance (2000) have advocated thesignificance of evaluating the quality of the construct’s scale through measurementinvariance tests. Measurement invariance is defined as the “property of a measurementinstrument (in the case of survey research, a questionnaire), implying that the instrumentmeasures the same concept in the same way across various subgroups of respondents”(Davidov et al., 2014, p. 58). The measurement invariance tests move beyond the classicalapproaches of “psychometric theory” (see also Raykov andMarcoulides, 2011) bymaking useof relatively new analytic tools to investigate questions such as whether theoperationalisation and measurement of the construct is similar or dissimilar acrossdifferent cultural groups (Vandenberg and Lance, 2000). This approach is particularly usefulto ascertain cross-cultural endorsement of the construct’s meaning and measurement(Ahmad, 2018; Davidov et al., 2014).
This paper responds to such calls by examining potential variations in the perceivedattributes of ethical leadership through the application of measurement invariance tests.Building on the ILT (Lord and Maher, 1991), this paper acknowledges that the key attributesof ethical leadership may portray a diverging range of endorsement from different culturalreference points. This notion is discussed next.
2.3 Implicit leadership theoryILT has garnered considerable attention in the field of cross-cultural leadership over the pasttwo decades. Its value and explanatory power have been validated through the GLOBEproject. The key premise of ILT is based on understanding leadership as a sociallyconstructed phenomenon that is shaped by the followers’ interpretation of the behaviouralattributes of their leaders (Lord and Maher, 1991). The followers’ interpretation of theattributes of effective leadership is influenced by their unique values, beliefs andassumptions, often culturally contextualised, which guide the process through which theydifferentiate leaders from non-leaders (Javidan et al., 2006). The theory has germaneimplications for the present study, because, if ethical leadership is considered as a sociallyconstructed phenomenon, then there are foreseeable variations in its meaning across Easternand Western cultures, given that their members are exposed to different socialisationprocesses and may hold different ethical viewpoints (House et al., 2004; Javidan et al., 2006;Robertson, 2017). Grounded in this perspective, the GLOBE project has laid the foundationsof the culturally endorsed ILT (House et al., 2004). The present study extends the literature byexploring the relevance of the theory to understand convergence and divergence in themeaning of ethical leadership in a cross-cultural context.
2.4 Measurement invariance of ethical leadership in cross-cultural contextsTesting for cross-cultural measurement invariance of ethical leadership is the cornerstone toevaluating whether the construct represents a similar or a different pattern of a behaviouralor social phenomenon in varied cultural and national contexts. This has emerged as a highlyrecommended procedure to assess constancy in the construct’s meaning in the internationalmanagement literature (Cascio, 2012). Yet, despite the substantive growth in the ethicalleadership literature since 2005, cross-cultural measurement invariance properties of the ELSremain poorly understood. This is because the extant literature maintains a prime focus onthe followers’ job-related outcomes of ethical leadership in organisations of a single country(Ahmad et al., 2018; Chughtai et al., 2015; Stouten et al., 2013). Since the business importanceof, and academic interest in, ethical leadership is recognised internationally, the issue of itscross-cultural measurement needs careful consideration.
Cross-culturalmeasurement
invariance
1327
A growing body of cross-cultural research elaborates the significance of examiningmeasurement invariance (MI), especially when the scales were originally developed andvalidated in one cultural context and subsequently deployed in another. In particular,research by Billiet (2013) conveys a remarkable account of how discounting the issue of cross-cultural MI can twist the comparative findings of people’s religious involvement in differentcountries. His views were based on the interesting findings of the European Social Survey.The survey contained items to measure religious involvement with an example item “Howoften the respondent attends religious services”. Billiet (2013) observed stark differences inthe interpretation of this question in the Turkish context, where, in contrast to secularEuropean countries, the majority of respondents adhere to the Islamic value system. It is notmandatory for women to attend religious services in a congregation in Islam. Accordingly, hecautioned that ignoring the issue of cross-cultural invariance in questions such as this leadsto an underestimation of the extent of females’ religious involvement and may lead to aninvalid conclusion that Turkish women are less religious. This demonstrates thatrespondents from different societal cultures may not share a common frame of reference inproviding answers to the same question.
Moving on to the application of the ELS in cross-cultural research, Resick et al. (2011,pp. 450–451) commented as follows:
The findings of the current study provide evidence that the content of Brown et al.’s (2005) ELS isrepresentative of the characteristics and actions attributed by managers from a selection of Easternand Western societies to ethical leaders.
Although it may appear from this comment that ELS is prototypical and it provides a culturalinvariant operationalisation of ethical leadership, we noted that research method by Resicket al. (2011) was neither quantitative nor did it deploy ELS. In other words, an investigationinto cross-cultural MI of ethical leadership is warranted to support such claims.
It is our contention that employees from different cultures may respond differently to thespecific items contained in the ELS because sensitivity to ethical issues tends to vary acrosscultures. This can be explained through an example of a common ethical problem of bullyingat work. Research in that area has revealed that people from different cultures respond tobullying behaviours differently (Samnani, 2013). Key reasons for this discrepancy areexplained in the light of cultural dimensions of power distance and individualism. Researchshowed that high power distance in a culture permits such behaviour and cultural emphasison bidimensional individualism-collectivism in interpersonal relations shapes the way peoplewould respond to bullying (Samnani, 2013). Where individualism is a highly emphasisedvalue in the West, because Western culture accentuates individual freedom and uniqueness,collectivism is more valued in the East where the larger group receives more importance thanindividual members of that group (Hofstede, 2001). Accordingly, it is also very likely thatleadership in an Eastern culture may give less importance to individuals than the largergroup in a decision-making process. A low-cost, high-quality or a higher profit-marginbusiness proposal made by a person outside the group may not be acceptable to a leadertrained in a collectivist culture because, irrespective of the cost/quality/profit, a proposaloffered by someone closer or loyal to his or her larger group would be more acceptable thanone coming from outside (see also Pitta et al., 1999). Such a pattern of decision-making by aleader may be considered unethical in individualistic cultures. Thus, perceptions of (un)ethical behaviour and practice of leadership can diverge across cultures.
Empirical research on ethical leadership performed outside the US has increasinglyapplied the definition and ELS of Brown et al. (2005) (e.g., Ahmad, 2018; Chughtai et al.,2015; Stouten et al., 2013). Since ELS was developed in the US, scholars from othercountries have examined whether the scale exhibits authentic psychometric propertiesin their specific cultural contexts (see, e.g. Chughtai et al., 2015; Wang et al., 2016). The
IJM41,8
1328
deployment of the scale in international research is still debatable because rarely is MItesting based on samples from both East and West reported within the single study. Inregard to cross-cultural MI, a standard convention is to ensure linguistic equivalencethrough procedures of translation and back-translation. However, Scandura andDorfman (2004) deemed this as a necessary but not a sufficient condition. They arguedthat, although linguistic translation helps to preserve equivalence of meaning to someextent, failure to ascertain MI will nonetheless result in reporting bias. In their words,“it is possible that some of the differences, reported as ‘cultural similarities’ and‘cultural differences,’ are [mere] artifacts” (p. 279). Davidov et al. (2014) concurred withthis perspective, arguing that “lack of measurement equivalence can threatenmeaningful comparisons across populations as well as conclusions based on pooleddata of different populations” (p. 59). They also emphasised that MI should not be onlyinterpreted as an absence of differences across two or more groups, but also as anassessment of the group’s similar position regarding the phenomenon of interest. Thisensures an equivalent response.
More specifically, in statistical terms, MI estimates the extent of similarity orcomparability in factor structures and factor loadings of the scale items. A measure isdeemed cross-culturally invariant when respondents from different contexts sharesimilarities in the conceptualisation of the underlying latent construct and in theirinterpretation of the contained items (Hult et al., 2008; Vandenberg and Lance, 2000).According to Cascio (2012, p. 2,533), “respondents also must interpret response options in thesame way”: he noticed that response options such as “neither disagree nor agree” impliedneutrality in some cultures and subtle agreement in others. This suggests that validation ofMI of ethical leadership by using international or cross-cultural contexts is an importantapproach to advancing knowledge in international settings.
3. MethodologyExtending the sparse literature that examines ethical leadership across Eastern andWesterncontexts (see Resick et al., 2006 for exceptions; Resick et al., 2011), the present study comparedperceived ethical leadership betweenAustralia and Pakistan. The culture of the two countriesdiffers in numerous ways. Figure 1 provides a comparison of the societal profiles forAustralia and Pakistan using Hofstede et al.’s (2010) findings.
As can be seen from Figure 1, the two societies differ remarkably in terms of their scoreson the following cultural dimensions: individualism (Australia 5 90, Pakistan 5 14), powerdistance (Australia 5 38, Pakistan 5 55), uncertainty avoidance (Australia 5 51,Pakistan 5 70), long-term orientation (Australia 5 21, Pakistan 5 50), masculinity(Australia5 61, Pakistan5 50) and indulgence (Australia5 71, Pakistan5 0) (see Hofstedeet al., 2010 for details). Thus, Australia has a high individualist culture with a score of 90,which indicates a loosely-knit society where employees are expected to be self-contained. Bycontrast, Pakistan’s score of 14 on the same dimension is indicative of a strong collectivisticsociety. This tendency also explains Pakistanis’ long-term commitment to their group, whereemployer-employee relationships can be perceived similar to a family link. Likewise,Australia has a low power distance culture in comparison to Pakistan, which implies thatfollowers in Pakistan are less likely to question or challenge the behaviour of their leadersthan are followers in Australia. Of the multitude of cultural differences between East andWest, both power distance and individualism are considered important for understandingleadership (House et al., 2004; Taras et al., 2010; Wasti et al., 2007). As one example, researchby Smith et al. (2002) on 47 countries showed that collectivism and power distance stronglyinfluence the extent to which middle managers rely on hierarchical and formal sources of
Cross-culturalmeasurement
invariance
1329
guidance. Javidan et al. (2006) further confirmed variations in the value dimensions betweenthe Anglo and Southern Asian cultural clusters (to which Australia and Pakistan belong,respectively).
To measure ethical leadership between the two cultures with the help of ELS, we utilisedthe data from Ahmad’s (2018) cross-cultural project. The data were collected through asurvey which addressed some of the methodological issues with cross-cultural researchdesigns (see also Cascio, 2012; Schaffer and Riordan, 2003). For example, Schaffer andRiordan (2003, p. 181) recommended that “Researchers should address the equivalency oftheir samples across contexts as well as the uniformity of their survey-administrationprocedures”. Ahmad (2018) has aligned the within-country work settings and surveyedrespondents with comparable job duties (i.e. academic staff in universities). The researchapplied similar sampling strategies to collect international data with the same timeframe.Table 1 presents the samples’ profile and shows that the matching of samples was based ontheir demographic characteristics.
3.1 Measurement of ethical leadership and cross-cultural MI analysisConsistent with Brown et al. (2005), the research participants provided perceived levels ofethical leadership of the head of their department through a 10-item ELS. The responses wereon a 5-point Likert scale, with 1 as “strongly disagree” and 5 as “strongly agree”. An exampleitem is: “My head of department sets an example of how to do things the right way in terms ofethics”. This scale’s reliability scores (Cronbach’s α) were 0.95 for the Australian group and0.92 for the Pakistani group.
We followed Vandenberg and Lance’s (2000) and Byrne’s (2004) recommendations to testfor cross-cultural MI of ELS with structural equation modelling in AMOS (Version 25). Thisinvolved a simultaneous application of the confirmatory factor analysis (CFA) procedure toboth cultural groups and an analysis of the respective input matrices before and afterconstraining the measured attributes (also called multi-group confirmatory factor analysis orMGCFA). Using these strategies, we determined (1) the extent to which the single-factorstructure of the ELS, previously established in the US context (Brown et al., 2005), wasreplicable in this study’s contexts and (2) the extent of MI between Australian and Pakistanigroups. Figure 2 shows the measurement model of the ELS.
Source(s): Hofstede et al. (2010)
0102030405060708090
100
Individualism Powerdistance
Masculinity Uncertaintyavoidance
Indulgence Long termorienta�on
Scor
es
Cultural dimensions
Australia Pakistan
Figure 1.Comparison of thenational cultures ofAustralia and Pakistan
IJM41,8
1330
The fitness of the measurement model was evaluated on the basis of commonly used fitindices of the χ2/df ratio, comparative fit index (CFI), Tucker–Lewis index (TLI), root meansquare error of approximation (RMSEA) and standardised root mean square residual (SRMR)(Byrne, 2004; Kline, 2011).
4. ResultsThe mean scores for each ELS item ranged from 3.18 to 3.78 in Australia and 3.31 to 3.67 inPakistan. Further, no statistically significant differences in the mean scores across all ELSitems were evident from the independent samples’ t-test of the cross-cultural datasets. Afterthis preliminary analysis, a series of progressively restrictive tests was performed on thecross-country ethical leadership data in MGCFA. First, configural invariance was tested toconfirm whether the factor structure of the ELS is constant across the Australian andPakistani samples. The model fit indices were acceptable in this case (Model 1: χ25 302.542,df 5 70, TLI 5 0.938, CFI 5 0.952, RMSEA 5 0.072, SRMR 5 0.037), hence validating thecommon factor structure and comparable pattern of factor loadings across the study’scompared groups. Second, metric invariance of ELSwas evaluated by constraining the factor
CharacteristicsAustralian group
(N 5 306)<Pakistani group
(N 5 330) Difference
Gender Male 102 (36.3) 196 (62.4) χ2(1,N 5 595) 5 40.47
Female 179( 63.7) 118 (37.6) p < 0.001Age group under 20 years 1 (0.4) 0 (0.0) χ2(5,
N 5 592) 5 55.5820–29 years 35 (12.5) 79 (25.2) p < 0.00130–39 years 89 (31.9) 139 (44.4)40–49 years 61 (21.9) 61 (19.5)50–59 years 72 (22.8) 23 (7.3)above 60 years 21 (7.5) 11 (3.5)
Job title Lecturer 66 (23.5) 81 (25.7) χ2(6,N 5 596) 5 66.365
Senior Lecturer/Assistant Prof.
43 (15.3) 130 (41.3) p < 0.001
Associate Prof. 28 (10.0) 27 (8.8)Professor 25 (8.9) 20 (6.3)Teaching Associate/Tutor
24 (8.5) 13 (4.1)
Research Associate/Assistant
50 (17.8) 21 (6.7)
Other 45 (16.0) 23 (7.3)Disciplines Business, Economics
and Law106 (38.7) 140 (44.4) χ2(3,N5 589)5 4.70
Arts, Design andHumanities
67 (24.5) 55 (17.5) p 5 0.195
Engineering andTechnology
20 (7.3) 23 (7.3)
Sciences 81 (29.6) 97 (30.8)Employment type Full time 206 (74.4) 287 (93.8) χ2(1,
N 5 583) 5 42.01Part time 71 (25.6) 19 (6.2) p < 0.001
Terms ofemployment
Fixed 132 (47.5) 110 (35.1) χ2(1,N5 591)5 9.27On-going 146 (52.5) 203 (64.9) p 5 0.002
Table 1.A Comparison betweenthe Australian sample
and the Pakistanisample
Cross-culturalmeasurement
invariance
1331
loadings to be equivalent across Australian and Pakistani group of respondents. Theconstrainedmodel (Model 2) fitted the data appropriately (χ25 309.456, df5 78, TLI5 0.945,CFI 5 0.952, RMSEA 5 0.068, SRMR 5 0.032), with only subtle differences (difference ofχ25 309.456/78–302.542/705 6.914/8, p-value5 0.546, ns). The calculated CAIC value for theModel 2 was 453.34, which was lower than the Model 1 value of 470.75. However, in view ofthe findings of non-statistically significant χ2 differences (p-value5 0.552) with no significantchanges in the values of the CFI, TLI, GH, SRMR and RMSEA, metric invariance of the ELSwas construed. Third, scalar invariance was tested by constraining the items’ intercepts to besimilar across the two groups in addition to the previously imposed constraints.
Although the model fit indices for the resulting model (Model 3) were acceptable(χ2 5 368.86, df 5 88, TLI 5 0.94, CFI 5 0.94, RMSEA 5 0.059, GH 5 0.92, SRMR 5 0.04),changes in the values of the CFI (0.01), GH (0.013), and McDonald’s non-centrality index(0.032), were significantly different to Model 2 (acceptable at <0.01, <0.01, <0.02). Inaccordance withMilfont and Fischer (2010), the difference (χ25 58.50, df5 9, p-value < 0.001)was found significant and the value of CAIC (486.61) larger than for Model 2 (453.34).Considering significant differences between Model 2 and Model 3 (p< 0.001) and statisticallysignificant decrease in the values of CFI, GH andMcDonald’s non-centrality index (McDonaldand Marsh, 1990) and increase of CAIC, scalar invariance of ELS was not supported. Table 2provides the summary results of the multi-group invariance tests.
Table 2 shows that both CAIC and ECVI values decrease and change in the model fitindices are less than the corresponding cut-offs. Accordingly, metric invariance of the ELSgained support. Furthermore, changes in CFI (50.01), GH (0.013), NCI (0.032) and an increaseof CAIC and ECVI indicate that Model 3 (with both factor loadings and interceptsconstrained) has significantly worse fit to the cross-cultural ethical leadership data compared
Figure 2.The measurementmodel for ethicalleadership
IJM41,8
1332
χ2(df)
χ2/(df)
RMSEA[90%
CI]
SRMR
Δχ2(Δdf)
CFI(ΔCFI)
GH(ΔGH)
NCI
(ΔNCI)
ECVI[90%
CI]
CAIC
Com
parison
Decision
Model1
302.542
4.322
0.072[0.064
0.081]
0.037
0.952
0.932
0.833
0.666[0.587
0.757]
470.749
Accepted
Configural
invariance
(70)
Model2
310.363
3.929
0.068[0.060
0.076]
0.038
7.820
0.952
0.932
0.833
0.65
[0.571
0.742]
453.339
Model1vs
Model2
Accepted
Metric
invariance
(79)
(�0.004)
(0.001)
(9)
(<0.001)
(<0.001)
(<0.001)
Model3
368.863
4.192
0.071[0.064
0.079]
0.038
58.500*
0.942
0.919
0.802
0.714[0.627
0.814]
486.608
Model2vs
Model3
Rejected
Scalar
invariance
(88)
(0.003)
(<0.001)
(9)
(0.010)
(0.013)
(0.032)
Model4
450.153
4.593
0.075[0.068
0.082]
0.047
81.290*
0.927
0.901
0.758
0.811[0.713
0.921]
539.864
Model3vs
Model4
Rejected
Error
variance
invariance
(98)
(0.00)
(0.01)
(9)
(0.01)
(0.02)
(0.04)
Model5
450.528
4.551
0.075[0.068
0.072]
0.052
0.374
0.927
0.900
0.758
0.808[0.710
0.918]
537.435
Model4vs
Model5
Rejected
Factor
variance
invariance
(99)
(<0.0001)
(0.00)
(1)
(<0.001)
(0.00)
(<0.001)
Model6
451.604
4.516
0.074[0.068
0.082]
0.052
1.076
0.927
0.901
0.758
0.807[0.709
0.917]
535.708
Model5vs
Model6
Rejected
Factor
means
invariance
(100)
(�0.001)
(<0.001)
(1)
(<0.001)
(-0.001)
(<0.001)
Acceptance
criteriafor
indices
(differences)**
<3
<0.08
(<0.015)
<0.08
(0.90
(0.90
(<0.01)
(<0.02)
min
(decreasing)
min
(decreasing)
Note(s):*Significantat
thesignificance
level,p
x < �1.96)Invariant ornon-invariant
1. My head of department can be trusted a1-1- a1-2 1.029 Invariant2. My head of department listens to whatemployees have to say
a2-1- a2-2 0.459 Invariant
3. My head of department defines success notjust by results but also the way that they areobtained
a3-1- a3-2 0.471 Invariant
4. When making decisions, my head ofdepartment asks “what is the right thing to do?
a4-1- a4-2 0.079 Invariant
5. My head of department disciplines employeeswho violate ethical standards
a5-1- a5-2 1.145 Invariant
6. My head of department conducts his/herpersonal life in an ethical manner
a6-1- a6-2 1.200 Invariant
7. My head of department has the best interestsof employees in mind
a7-1- a7-2 1.161 Invariant
8. My head of department makes fair andbalanced decisions
a8-1- a8-2 0.520 Invariant
9. My head of department discusses businessethics or values with employees
a9-1- a9-2 1.218 Invariant
10. My head of department sets an example ofhow to do things the right way in terms ofethics
a10-1- a10-2 �1.223 Invariant
Residual error Path 1 v1-1- v1-2 3.336 Non-invariantResidual error Path 2 v2-1- v2-2 3.507 Non-invariantResidual error Path 3 v3-1- v3-2 1.57 InvariantResidual error Path 4 v4-1- v4-2 3.736 Non-invariantResidual error Path 5 v5-1- v5-2 2.518 Non-invariantResidual error Path 6 v6-1- v6-2 1.250 InvariantResidual error Path 7 v7-1- v7-2 2.674 Non-invariantResidual error Path 8 v8-1- v8-2 3.817 Non-invariantResidual error Path 9 v9-1- v9-2 0.785 InvariantResidual error Path 10 v10-1- v10-2 1.264 Invariant
Note(s): Where ai (i 5 1-10)-1 5 Paths (measurement) for Australian sample- ai (i 5 1-10)-2 5 Paths(measurement) for Pakistan Sample, vi, (i5 1-10)-15 Residual error (measurement) for Australian Sample, vi,(i 5 1-10)-2 5 Residual error (measurement) for Pakistan sample
Table 3.Pairwise Parametercomparison throughcritical ratios fordifferences
IJM41,8
1334
contributions to the literature: (1) based on the culturally endorsed ILT (Javidan et al., 2006), itsheds light on both similarities and differences in perceptions of ethical leadership across twosocietal cultures – Australia and Pakistan – and (2) it validates the commonly appliedmeasure (i.e. ELS, Brown andTrevi~no, 2006) in cross-cultural contexts throughMI tests. In sodoing, the present study addressed the question surrounding convergence and divergence inthe meaning of ethical leadership across East and West.
Previously Resick et al. (2006) showed that key attributes associated with ethicalleadership are globally endorsed. Yet they also noted some variations in the level ofendorsement across the GLOBE study’s cultural clusters. The findings of the present study(see Table 4) align with their findings of Anglo and Southeast Asian clusters, as they noticedthat the former cluster endorsed almost all attributes of ethical leadership to a greater extentthan the latter cluster.
Yet both Australians and Pakistanis followers strongly emphasised that fair and balanceddecision-making, trust and keeping the best interests of employees in mind were attributes ofethical leadership. In the Australian sample, trust in leadership appeared to be the mostemphasised attribute by followers, whereas leaders’ ability to exemplify ethical actions bydoingthings in the right manner was the most emphasised attribute in the Pakistani sample. Thislends support to somedegree of convergence and somedegree of divergence in perceived ethicalleadership in the cross-cultural context. Previous studies revealed that the establishment of MIis an important criterion for construct generalisation and measurement validation in order tomake meaningful cross-country comparisons (Milfont and Fischer, 2010; Vandenberg andLance, 2000). Building on these insights, we assessed the configural, metric and scalarinvariance in ELS by using samples fromAustralia and Pakistan. Since Australia and Pakistanare culturally different countries (Hofstede et al., 2010), we expected variations in the ethical
ELS items and corresponding pathsStandardised factorloadings for Australia
Z-Value
Standardised factorloadings for Pakistan
Z-Value
1. My head of department can betrusted
0.886 1 0.812 1
2. My head of department listens towhat employees have to say
0.857 21.377 0.745 15.236
3. My head of department definessuccess not just by results but alsothe way that they are obtained
0.734 16.077 0.684 13.623
4. When making decisions, my head ofdepartment asks “what is the rightthing to do?
0.863 21.661 0.784 16.326
5. My head of department disciplinesemployees who violate ethicalstandards
0.653 13.436 0.565 10.773
6. My head of department conducts his/her personal life in an ethical manner
0.695 14.732 0.677 13.425
7. My head of department has the bestinterests of employees in mind
0.887 23.060 0.847 18.241
8. My head of department makes fairand balanced decisions
0.909 24.432 0.854 18.471
9. My head of department discussesbusiness ethics or values withemployees
0.722 15.645 0.691 13.783
10. My head of department sets anexample of how to do things theright way in terms of ethics
0.881 22.691 0.862 18.718Table 4.
Comparative factorloadings and
associated Z-values
Cross-culturalmeasurement
invariance
1335
behaviour of the leaders in these countries and examined the MI scores of ELS across the twocultural groups to ascertain the nature of the potential divergence. The findings supportedconfigural invariance of the ELS (see also Milfont and Fischer, 2010).
This implied that the conceptualisation of ethical leadership was similar across the twocompared groups. This finding broadly aligns with the earlier observations by scholars such asEisenbeiss (2012) and Resick et al. (2006) indicating commonality of moral norms andbehavioural attributes associated with ethical leadership. Eisenbeiss (2012) stated that “All thecurrent approaches to ethical leadership proceed from a Western perspective on ethicalleadership and do not consider viewpoints, principles or values of other cultural clusters”(p. 793). Yet his cogent review of the seminal writings from the East and theWest that coveredboth ancient and modern moral philosophy through the writings associated with Plato,Aristotle, Confucius, Tagore andKant, and in the light of ethical teachings in dominant religionsin the world (Christianity, Islam, Buddhism and Hinduism), revealed that the underlyingnormative principles for ethical actions are similar across cultures.His interdisciplinaryanalysisof Western and Eastern philosophical and religious traditions further illuminates the centralnormative principles of ethical leadership. Eisenbiess’ (2012) research findings revealed thatethical leaders “treat others with dignity and respect”, “make making fair and consistentdecisions”, do not discriminate, show “concern for the welfare of society and the environment”and demonstrate “temperance and humility”. Importantly, the identified principles providenormative reference points for ethical leadership in cross-cultural contexts and implyconvergence in the broader meaning of ethical leadership. The present study’s findings ofconvergence in the meaning of ethical leadership across Anglo and Southern Asian cultures,evident from the configural invariance test, empirically substantiate this notion.
Moving on to the results of metric invariance of ELS to explore the extent to which theAustralian and Pakistani participants interpreted the scale items in a similar style, we foundsimilarity in the relationship of various items with the underlying ethical leadershipconstruct. This suggests that ratings of perceived ethical leadership are analogous across thetwo cultural groups (Milfont and Fischer, 2010). However, the present study could notsupport scalar invariance of ELS. Based on this finding, we caution researchers againstmaking comparisons of the latent mean scores of ELS when undertaking comparativeinternational research. Based on our study’s findings, we argue that ethical leadership can beconceptualised as a “variform universal” style of leadership (see also Dickson et al., 2012)because the comparative factor loadings of ELS items demonstrate uniformity in terms ofstrength and significance of the scale (see Table 4). All in all, ethical leadership is seeminglyconvergent across Australia and Pakistan in terms of its meaning; however, the enactment ofthis particular leadership style may differ across cultures.
5.1 Implications, limitations and future research directionsThe trends towards the internationalisation of business operations, together with a generalexpectation of stakeholders that businesses should engage in morally and socially responsiblepractices, have positioned ethical leadership at the forefront of public interest. This paper hasimportant implications for the theory and practice of ethical leadership. It provides a cross-cultural endorsement of the relevance of ILT to understanding ethical leadership ininternational settings. Since the study’s findings conveyed convergence in the meaning ofethical leadership, this implies that it can be considered a global leadership construct (Bird andMendenhall, 2016). Therefore, organisations operating in the East or West should considerinvesting in ethical behavioural training in their leadership development programs.
The study makes another key contribution to the literature by delineating behaviouralattributes of ethical leadership that are endorsed differently by Australians and Pakistanis.Finally, the present study addresses a substantive limitation in the methodological
IJM41,8
1336
development of ethical leadership theory by providing cross-cultural evidence of thepsychometric properties through MI testing. The study has enhanced the construct validityof ethical leadership through the use of the refined multiple-sample analytical approach.While previous studies have assumed the measures of ethical leadership are invariant acrossvarious contexts, this is the first study that employs a robust methodological technique(involving metric and path invariance) which demonstrates the significant differencebetween each item and path and generalises the validity of ethical leadership construct andits measures by using international samples.
While “scale development is the first step in ensuring practitioner adoption of usefulmeasures rather than quasi-measures that fail to support organizations in leadershipdevelopment” (Crawford and Kelder, 2019, p. 142), the findings of our study will assist in themore effective development of ethical leaders in international organisations. Multinationalcorporations considering enhancing ethical leadership practices across East and West mayfind it useful to measure it consistently. Once the behaviour is appropriately measured, it willbe more convenient to include it in performance appraisals. They can also use it as a tool toscreen out candidates when making promotion decisions.
Despite contributing these advances to the ethical leadership literature, the present studyalso has limitations. First, it has a limited scope because we did not examine the relationshipsbetween specific cultural contingencies and ethical leadership.We therefore urge researchersto directlymeasure the specific cultural contingencies in their studies by examiningwhat sortof ethical behaviours and practices are valued in specific contexts. Second, the study’s datawere confined to a specific work context of the two countries: that is, a sample of academicsworking in Australia and Pakistan. This means that the findings are neither generalisable tothe general working population of these countries nor to other national/cultural contexts.Australia and Pakistan are, respectively, part of the Anglo culture cluster and SouthernAsian culture under the GLOBE project (House et al., 2004). While the study’s MI tests havesupported that both cultural groups perceive the meaning of ethical leadership in a similarmanner, the single-factor ELS model may or may not be generalisable to other cultures.Further replications in other cultures such as Latin America, the Middle East andSub-Saharan Africa are required to support cross-cultural MI of ELS (see also House et al.,2004, for details). Finally, our statistical analysis did not control for the potential confoundinginfluence of respondents’ demographic characteristics, such as gender, which can provideadditional dimensions to advance the understanding of the measurement differences.
In closing, this paper has presented an in-depth analysis of ethical leadership in across-cultural context. The study’s findings have informed cross-cultural transferability ofthe ethical leadership construct and supported configural andmetric invariance of the ELS inAustralian and Pakistani samples.
References
Ahmad, S. (2018), “Can ethical leadership inhibit workplace bullying across east and west: exploringcross-cultural interactional justice as a mediating mechanism”, European Management Journal,Vol. 36 No. 2, pp. 223-234, doi: 10.1016/j.emj.2018.01.003.
Ahmad, S., Fazal-E-Hasan, S.M. and Kaleem, A. (2018), “How ethical leadership stimulates academics’retention in universities: the mediating role of job-related affective well-being”, InternationalJournal of Educational Management, Vol. 32 No. 7, pp. 348-1362.
Bandura, A. (1977), Social Learning Theory, Prentice-Hall, Englewood Cliffs, NJ.
Billiet, J. (2013), “Quantitative methods with survey data in comparative research”, in Kennett, P. (Ed.),A Handbook of Comparative Social Policy, 2nd ed., Edward Elgar, Cheltenham, pp. 264-300.
Cross-culturalmeasurement
invariance
1337
Bird, A. and Mendenhall, M.E. (2016), “From cross-cultural management to global leadership:evolution and adaptation”, Journal of World Business, Vol. 51 No. 1, pp. 115-126.
Brown, M.E. and Trevi~no, L.K. (2006), “Ethical leadership: a review and future directions”, TheLeadership Quarterly, Vol. 17 No. 6, pp. 595-616.
Brown, M.E., Trevi~no, L.K. and Harrison, D.A. (2005), “Ethical leadership: a social learning perspectivefor construct development and testing”, Organizational Behavior and Human DecisionProcesses, Vol. 97 No. 2, pp. 117-134.
Byrne, B.M. (2004), “Testing for multigroup invariance using AMOS graphics: a road less traveled”,Structural Equation Modeling, Vol. 11 No. 2, pp. 272-300.
Cascio, W.F. (2012), “Methodological issues in international HR management research”, InternationalJournal of Human Resource Management, Vol. 23 No. 12, pp. 2532-2545, doi:10.1080/09585192.2011.561242.
Chughtai, A., Byrne, M. and Flood, B. (2015), “Linking ethical leadership to employee well-being: therole of trust in supervisor”, Journal of Business Ethics, Vol. 128 No. 3, pp. 653-663.
Chungkham, H.S., Ingre, M., Karasek, R., Westerlund, H. and Theorell, T. (2013), “Factor structure andlongitudinal measurement invariance of the demand control support model: an evidence fromthe Swedish Longitudinal Occupational Survey of Health (SLOSH)”, PloS One, Vol. 8 No. 8,e70541.
Crawford, J.A. and Kelder, J.-A. (2019), “Do we measure leadership effectively? Articulating andevaluating scale development psychometrics for best practice”, The Leadership Quarterly,Vol. 30 No. 1, pp. 133-144.
Daniels, M.A. and Greguras, G.J. (2014), “Exploring the nature of power distance implications formicro-and macro-level theories, processes, and outcomes”, Journal of Management, Vol. 40No. 5, pp. 1202-1229.
Davidov, E., Meuleman, B., Cieciuch, J., Schmidt, P. and Billiet, J. (2014), “Measurement equivalence incross-national research”, Annual Review of Sociology, Vol. 40, pp. 55-75.
Den Hartog, D.N. (2015), “Ethical leadership”, Annual Review of Organizational Psychology andOrganizational Behavior, Vol. 2 No. 1, pp. 409-434.
Dickson, M.W., Casta~no, N., Magomaeva, A. and Den Hartog, D.N. (2012), “Conceptualizing leadershipacross cultures”, Journal of World Business, Vol. 47 No. 4, pp. 483-492.
Eisenbeiss, S.A. (2012), “Re-thinking ethical leadership: an interdisciplinary integrative approach”,The Leadership Quarterly, Vol. 23 No. 5, pp. 791-808.
George, W. (2003), Authentic Leadership: Rediscovering the Secrets to Creating Lasting Value, Jossey-Bass, San Francisco, CA.
Grace, W.J. (1998), Ethical Leadership: In Pursuit of the Common Good, Vol. 1, CEL publications,Seattle, WA.
Hinkin, T.R. (1998), “A brief tutorial on the development of measures for use in surveyquestionnaires”, Organizational Research Methods, Vol. 1 No. 1, pp. 104-121.
Hofstede, G. (2001), Culture’s Consequences: Comparing Values, Behaviors, Institutions andOrganizations across Nations, 2nd ed., Sage, Thousand Oaks, CA.
Hofstede, G., Hofstede, G.J. and Minkov, M. (2010), Cultures and Organizations: Software of the Mind,3rd ed., McGraw-Hill, New York.
Houghton, J.D., Carnes, A. and Ellison, C.N. (2014), “A cross-cultural examination of self-leadership:testing for measurement invariance across four cultures”, Journal of Leadership andOrganizational Studies, Vol. 21 No. 4, pp. 414-430.
House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W. and Gupta, V. (2004), Culture, Leadership, andOrganizations: The GLOBE Study of 62 Societies, Sage, Thousand Oaks, CA.
IJM41,8
1338
Hult, G.T.M., Ketchen, D.J., Griffith, D.A., Finnegan, C.A., Gonzalez-Padron, T., Harmancioglu, N., . . .and Cavusgil, S.T. (2008), “Data equivalence in cross-cultural international business research:assessment and guidelines”, Journal of International Business Studies, Vol. 39 No. 6,pp. 1027-1044.
Javidan, M., Dorfman, P.W., De Luque, M.S. and House, R.J. (2006), “In the eye of the beholder: crosscultural lessons in leadership from project GLOBE”, Academy of Management Perspectives,Vol. 20 No. 1, pp. 67-90.
Kalshoven, K., Den Hartog, D.N. and De Hoogh, A.H.B. (2011), “Ethical leadership at workquestionnaire (ELW): development and validation of a multidimensional measure”, TheLeadership Quarterly, Vol. 22 No. 1, pp. 51-69.
Kline, R.B. (2011), Principles and Practice of Structural Equation Modeling, Guilford Press, New York.
Lord, R.G. and Maher, K.J. (1991), Leadership and Information Processing: Linking Perceptions andPerformance, Vol. 1, Unwin Hyman, Cambridge, MA.
Martin, G.S., Resick, C.J., Keating, M.A. and Dickson, M.W. (2009), “Ethical leadership across cultures:a comparative analysis of German and US perspectives”, Business Ethics: A European Review,Vol. 18 No. 2, pp. 127-144, doi:10.1111/j.1467-8608.2009.01553.x.
McDonald, R.P. and Marsh, H.W. (1990), “Choosing a multivariate model: noncentrality and goodnessof fit”, Psychological Bulletin, Vol. 107 No. 2, p. 247.
Milfont, T.L. and Fischer, R. (2010), “Testing measurement invariance across groups: applications incross-cultural research’, International Journal of Psychological Research, Vol. 3 No. 1,pp. 111-130.
Palanski, M., Avey, J.B. and Jiraporn, N. (2014), “The effects of ethical leadership and abusivesupervision on job search behaviors in the turnover process”, Journal of Business Ethics,Vol. 121 No. 1, pp. 135-146, doi:10.1007/s10551-013-1690-6.
Pitta, D.A., Fung, H.-G. and Isberg, S. (1999), “Ethical issues across cultures: managing the differingperspectives of China and the USA”, Journal of Consumer Marketing, Vol. 16 No. 3, pp. 240-256.
Raykov, T. and Marcoulides, G.A. (2011), Introduction to Psychometric Theory, Routledge, New York.
Resick, C.J., Hanges, P.J., Dickson, M.W. and Mitchelson, J.K. (2006), “A cross-cultural examination ofthe endorsement of ethical leadership”, Journal of Business Ethics, Vol. 63 No. 4, pp. 345-359.
Resick, C.J., Martin, G.S., Keating, M.A., Dickson, M.W., Kwan, H.K. and Peng, C. (2011), “What ethicalleadership means to me: Asian, American, and European perspectives”, Journal of BusinessEthics, Vol. 101 No. 3, pp. 435-457.
Robertson, D.C. (2017), “Business ethics across cultures”, Gannon, M.J. and Newman, K.L. (Eds), TheBlackwell Handbook of Cross Cultural Management, Blackwell Publishers, Oxford, pp. 361-392.
Samnani, A.-K. (2013), “The early stages of workplace bullying and how it becomes prolonged: the roleof culture in predicting target responses”, Journal of Business Ethics, Vol. 113 No. 1, pp. 119-132,doi:10.1007/s10551-012-1286-6.
Scandura, T. and Dorfman, P. (2004), “Leadership research in an international and cross-culturalcontext”, The Leadership Quarterly, Vol. 15 No. 2, pp. 277-307.
Schaffer, B.S. and Riordan, C.M. (2003), “A review of cross-cultural methodologies for organizationalresearch: a best-practices approach”, Organizational Research Methods, Vol. 6 No. 2,pp. 169-215, doi:10.1177/1094428103251542.
Smith, P.B., Peterson, M.F. and Schwartz, S.H. (2002), “Cultural values, sources of guidance, and theirrelevance to managerial behavior a 47-nation study”, Journal of Cross-Cultural Psychology,Vol. 33 No. 2, pp. 188-208.
Stouten, J., van Dijke, M., Mayer, D.M., De Cremer, D. and Euwema, M.C. (2013), “Can a leader be seenas too ethical? The curvilinear effects of ethical leadership”, The Leadership Quarterly, Vol. 24No. 5, pp. 680-695.
Cross-culturalmeasurement
invariance
1339
Taras, V., Kirkman, B.L. and Steel, P. (2010), “Examining the impact of culture’s consequences: athree-decade, multi-level, meta-analytic review of hofstede’s cultural value dimensions”, Journalof Applied Psychology, Vol. 95 No. 3, pp. 405-439, doi:10.1037/a0018938.
Tepper, B.J. (2007), “Abusive supervision in work organizations: review, synthesis, and researchagenda”, Journal of Management, Vol. 33 No. 3, pp. 261-289.
Vandenberg, R.J. and Lance, C.E. (2000), “A review and synthesis of the measurement invarianceliterature: suggestions, practices, and recommendations for organizational research”,Organizational Research Methods, Vol. 3 No. 1, pp. 4-70.
Wang, H., Lu, G. and Liu, Y. (2016), “Ethical leadership and loyalty to supervisor in China: the roles ofinteractional justice and collectivistic orientation”, Journal of Business Ethics, Vol. 146 No. 3,pp. 539-543.
Wasti, S.A., Tan, H.H., Brower, H.H. and €Onder, Ç. (2007), “Cross-cultural measurement of supervisortrustworthiness: an assessment of measurement invariance across three cultures”, TheLeadership Quarterly, Vol. 18 No. 5, pp. 477-489.
Wilson, S. and McCalman, J. (2017), “Re-imagining ethical leadership as leadership for the greatergood”, European Management Journal, Vol. 35 No. 2, pp. 151-154.
Zhu, W., Zheng, X., He, H., Wang, G. and Zhang, X. (2019), “Ethical leadership with both ‘moralperson’ and ‘moral manager’ aspects: scale development and cross-cultural validation”, Journalof Business Ethics, Vol. 158 No. 2, pp. 547-565.
Appendix
Corresponding authorSaima Ahmad can be contacted at: saima.ahmad@rmit.edu.au
For instructions on how to order reprints of this article, please visit our website:www.emeraldgrouppublishing.com/licensing/reprints.htmOr contact us for further details: permissions@emeraldinsight.com
Key attributes Defining features
Character and integrity Character refers “to the pattern of intentions, inclinations, and virtues thatprovide the ethical or moral foundation for behaviour”. Whereas “integrityentails the ability to both determine, as well as engage in morally correctbehaviour regardless of external pressures”
Ethical awareness “Ethical awareness is the capacity to perceive and be sensitive to relevant moralissues that deserve consideration in making choices that will have a significantimpact on others”
Community/people-orientation
The use of social power by a leader by considering “the collective interests of thegroup [or followers] over self-serving interests”
Motivating The process of “inspiring followers to work towards the leader’s vision for thegroup and to be committed to the group.”
Encouraging andempowering
Providing followers with “a sense of personal competence that allows them to beself-confident and self-sufficient”
Managing ethicalaccountability
Setting “standards and expectations of ethical conduct for followers” andholding them “accountable using the rewards and punishment systems that areavailable”
Source(s): Resick et al. (2006)
Table A1.The key attributes ofethical leadershipacross cultures
IJM41,8
1340
Reproduced with permission of copyright owner. Furtherreproduction prohibited without permission.
Is the meaning of ethical leadership constant across cultures? A test of cross-cultural measurement invariance
Introduction
Literature review
Ethical leadership: construct conceptualisation and measurement
Ethical leadership in cross-cultural contexts
Implicit leadership theory
Measurement invariance of ethical leadership in cross-cultural contexts
Methodology
Measurement of ethical leadership and cross-cultural MI analysis
Results
Discussion
Implications, limitations and future research directions
References
Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=cres20
Regional Studies
ISSN: 0034-3404 (Print) 1360-0591 (Online) Journal homepage: https://www.tandfonline.com/loi/cres20
Regional determinants of technical efficiency:evidence from the Greek economy
Theodore Tsekeris & Sotiris Papaioannou
To cite this article: Theodore Tsekeris & Sotiris Papaioannou (2018) Regional determinants oftechnical efficiency: evidence from the Greek economy, Regional Studies, 52:10, 1398-1409, DOI:10.1080/00343404.2017.1390312
To link to this article: https://doi.org/10.1080/00343404.2017.1390312
View supplementary material
Published online: 15 Nov 2017.
Submit your article to this journal
Article views: 719
View related articles
View Crossmark data
Citing articles: 8 View citing articles
Regional determinants of technical efficiency: evidence from theGreek economyTheodore Tsekerisa and Sotiris Papaioannoub
ABSTRACTThis paper emphasizes the role of regional determinants in the technical efficiency of the Greek economy. We employ aregional production function which is simultaneously estimated with an inefficiency equation. The results underline theconsiderable and growing core–periphery disparities in technical efficiency. They show the significant efficiency-enhancing impact of intra-regional market access and its dispersion. They also demonstrate that interregional marketaccess, specialization, sectoral concentration and human capital exert a positive but varying influence on the efficiencyof Greek regions.
KEYWORDSregional efficiency; agglomeration economies; market access; sectoral concentration; specialization
JEL D24, R12, R15HISTORY Received 25 November 2016; in revised form 29 September 2017
INTRODUCTION
This paper measures the technical efficiency of Greekregions and analyzes which factors determine its evolutionacross space and time. Most studies in the existing litera-ture have focused on the country- or industry-wide and,to a lesser extent, the regional analysis of productivity.However, the most relevant ones demonstrate considerablespatial variations (Mikelbank & Jackson, 2000; Romp &De Haan, 2007), suggesting the need for a comprehensivemodelling specification to treat various intrinsic issuesrelated to sources of inefficiency. Given that productionfunctions do not accurately capture productivity changes,since infrastructure investments affect the location of econ-omic activities (Haughwout, 1998), spatial effects on effi-ciency related to agglomeration externalities and marketaccess should be properly taken into account.
The scant existing literature has so far focused on how alimited range of agglomeration forces influence technicalefficiency. We contribute in this direction by examiningthe efficiency contribution of a wide range of regionaldeterminants, including various agglomeration economies,human capital and market access from both the inter- andintra-regional perspectives.
To this end, we use stochastic frontier analysis (SFA) toestimate simultaneously the influence of physical capitaland labour on regional output and the inefficiency
contribution of regional variables. We use data across 13regions of the Greek economy (see Figure A1 of AppendixA in the supplemental data online) for the period 2000–12.This period is characterized by the realization of massivecapital investments and the advent of the economic crisisin 2008. In contrast to other studies (Petrakos & Psycharis,2016; Psycharis, Rovolis, Tselios, & Pantazis, 2014) thathave concentrated on the effects of the crisis on regionaloutput, income and employment, we concentrate on thedeterminants of regional inefficiency during a periodencompassing a whole business cycle.
The analysis of technical efficiency at the regional levelof the Greek economy is of particular interest due to aunique set of intrinsic characteristics pertaining to its com-plex geography, which includes several scattered island andmountainous areas. Additionally, the spatial economy ofGreece is persistently characterized by a dual patternbetween the core region of Attiki, which accounted foralmost half the national gross domestic product (GDP)both before and during the crisis, maintaining previousdivergent trends (Petrakos & Psycharis, 2016), and periph-eral regions, which follow different development paths withhighly heterogeneous conditions of physical accessibilityand access to information and knowledge. Both thenational and regional economies of the country historicallyface severe institutional failures, with the informal andsheltered (public-sector) types of activities playing a
© 2017 Regional Studies Association
CONTACTa(Corresponding author) tsek@kepe.grCentre of Planning and Economic Research (KEPE), Athens, Greece.b sopa@kepe.grCentre of Planning and Economic Research (KEPE), Athens, Greece.
REGIONAL STUDIES2018, VOL. 52, NO. 10, 1398–1409https://doi.org/10.1080/00343404.2017.1390312
significant role. Importantly, as our sample encompassesthe 2000–12 period, we explore whether our baselineresults hold when considering the influence of the crisis.The results may provide useful insight into interpretinghow regional differences and crisis adversities affect theefficiency performance of a Southern European countrywhose economic model is structurally different than thatof the majority of Northern European countries.
We show that the urbanization economies and theirdispersion, interregional market access, specialization, sec-toral concentration and human capital are all significantdrivers of regional efficiency. We also show that region-and time-specific effects are significant factors ofinefficiency.
The paper is structured as follows. The next section dis-cusses the findings of the related literature. The third sec-tion introduces the econometric methodology. The fourthsection presents results about the efficiency of Greekregions and identifies factors that contribute to inefficiency.Finally, the fifth section summarizes and concludes.
RELATED LITERATURE
There is a growing body of research that employs SFA toexamine the influence of several regional determinants ontechnical (in)efficiency, where the production functionmostly concerns specific economic activities, such as inno-vation systems (e.g., Fritsch & Slavtchev, 2011) and other(mainly manufacturing) industries, rather than the totalsystem of sectors within a national economy. Comparedwith earlier studies, having used production function mod-elling to estimate regional differences in technical efficiencyamong industries, Beeson and Husted (1989) used SFA toshow that more human capital and more urbanization areboth associated with higher manufacturing efficiency acrossthe US states. A similar methodology was followed byMcCoy and Moomaw (1995) to indicate the positiveurbanization impact on manufacturing efficiency of Cana-dian cities, and by Driffield and Munday (2001) to showthe positive role of regional agglomeration on efficiencyof UK industries.
Regarding studies of Asian countries, Mitra (1999,2000) used SFA to stress the significance of urbanizationand industrial spread (up to a certain threshold) on the pro-ductivity of Indian manufacturing sectors. Otsuka, Goto,and Sueyoshi (2010) and Otsuka and Goto (2015) demon-strated that agglomeration economies and improvement ofmarket access have both a positive influence on the effi-ciency of Japanese industries. Moreover, Ke and Yu(2014) carried out SFA to show that whereas industrialagglomeration and human capital enhance efficiency, thedensity of employment negatively affects productivitygrowth in large Chinese cities, which should specializeand diversify in advanced producer services, rather thanmanufacturing. Several evidence in favour of public policiesfocusing on increased specialization to boost productivitygrowth can be found in the relevant literature, such asthose of Aiginger and Davies (2004) for European Unioncountries and Lee, Jang, and Hong (2010) for Korea.
The analysis of the regional dimension of total factorproductivity (TFP) and technical efficiency has been over-looked in the majority of studies concerning the Greekeconomy. The early study of Vagionis and Spence (1994)examined productivity developments in the manufacturingindustry of Greek regions during 1980–88. They showedthat the largest TFP gains are found in the non-centralregions hosting industrial area projects, which, in turn,can provide employment opportunities and infrastruc-ture-induced agglomeration economies. They also illus-trated that such gains are located in medium-sized citiesdue to the deployment of new technology, while theyargued that the availability of grants and incentives offereda significant advantage for productivity gains. The crucialrole of urbanization economies was demonstrated byLouri (1988), who used a production function only forthe metropolitan area of Athens and the rest of Greece asa whole for 20 manufacturing sectors during 1974–77.She illustrated that the productivity of manufacturing sec-tors is driven by the strong tendency of firms to locate inlarge urban centres.
In contrast with previous studies concerning the Greekeconomy, the present study employs a sound theoreticalframework on the basis of a production function thatencompasses the entirety of Greek regions during therecent period of 2000–12. Data availability across all sectorsof economic activity enables one to construct physical capi-tal stock series and diverse agglomeration-related variableswhich are likely to affect the efficiency of Greek regions.Another advancement of the present study refers to theestimation of regional inefficiency determinants simul-taneously with the estimation of a regional productionfunction. It is noted that although the scant previous litera-ture has attempted to measure the regional efficiency per-formance of Greek firms, such as those of themanufacturing industry (Floros, Voulgaris, & Lemonakis,2014), the absence of an appropriately detailed and com-plete data set does not allow the measurement of firm-level efficiency over the period of interest.
ECONOMETRIC METHODOLOGY
In order to estimate regional efficiency and identify itsdeterminants, we use SFA and rely on a model specificationthat jointly estimates a stochastic production function and atechnical inefficiency equation (Battese & Coelli, 1995).This estimation procedure is regarded as more accurateand provides robust results compared with the typicaltwo-stage estimation procedure. We model for the exist-ence of unobserved regional inefficiency, within a stochasticfrontier model, as follows:
ln (Yit) = b0 + l t + b1 ln (Lit)+ b2 ln (Kit)+ Vit
− Uit , (1)
where Yit is the output, expressed in value-added terms, ofeach region i at time t; l is the rate of technical change; andvariable t is a time trend, which captures technical progressover time; Lit is the labour input expressed as the number of
Regional determinants of technical efficiency: evidence from the Greek economy 1399
REGIONAL STUDIES
total hours worked; Kit is total physical capital; and the par-ameters b1 and b2 are the value added elasticities of labourand physical capital respectively. The two components ofthe error structure, Vit and Uit , compose the main featureof the stochastic frontier. The stochastic component Vit
describes random shocks to production, which are regionspecific and are assumed to be independent and identicallydistributed (i.i.d.) following a normal distribution N(0, s2
v).The stochastic component Uit , which is associated withtechnical inefficiency, is independently distributed fromVit and has an asymmetrical distribution equal to theupper half of the N(0, s2
u). Technical efficiency is a non-negative random variable, denoted as TEit = exp {−Uit},which is output-oriented and reaches its maximum whenTEit = 1.
As a way to study the effects of regional determinantson technical inefficiency, we model the mean mit of thetruncated distribution of Uit as follows:
mit = d0 + dHCHCit + dIMPIMPit + dDIMPDIMPit
+ dEMPEMPit + dSSit + dHHHHit
+ dPPi + dTTt + wit ,
(2)
where d0 is a constant; HC is the human capital variable,measured as the percentage of hours worked by personswith a tertiary education. A higher level of human capitalis expected to exert a significantly positive impact on pro-ductivity (Becker, 1993; Benos & Karagiannis, 2016).
The measure we adopt to represent urbanization econ-omies refers to the internal (intra-regional) market access,which is expressed by the variable of internal market poten-tial (IMP) to account for the importance of economies ofscale and transport costs. Given that the spatial unit ofanalysis refers to the NUTS-II level of region,1 a variableof the dispersion of the above measure (DIMP) is alsoused to account for possible significant intra-regional vari-ations among the constituent (NUTS-III level) prefecturesof each region. We expect that urbanization economiesincrease (decrease) in a region as the linkages of economicactivities among its constituent prefectures become closer(more distant). The level and dispersion of urbanizationeconomies available to all firms of a region irrespective ofsector typically have a positive association with pro-ductivity. The variable of external (interregional) marketpotential (EMP) constitutes a function of the proximityand ease of reaching sizeable economic activities in otherregions. Improved external market access generates moreeconomies of scale for distribution channels and industrialor service clusters, which further raise the productivity oflabour and physical capital in the region.
The variable of specialization S is expected to boost theproductivity of a region. When firms specialize geographi-cally, they can better learn from each other than geographi-cally isolated firms, while local competition provides themincentives to adopt new technologies, innovate and becomemore productive (Glaeser, Kallal, Scheinkman, & Shleifer,1992). The concentration (or, else, the inverse of diversifi-cation) of economic activities (HH ) in some region has aless straightforward and a more context-specific impact
on productivity compared with the other factors includedin our model. Jacobs (1969) argued that industrial diversity(or variety) promotes innovation and productivity growthdue to the interaction of firms from different sectors insearch of productive and competitive sources. This processmay support the knowledge transfers among dissimilarindustries and the increasing returns in regional production(Frenken, Van Oort, & Verburg, 2007; Siegel, Johnson, &Alwang, 1995). Holmes and Stevens (2002) and Bun SongLee et al. (2010) argued that the geographical concen-tration of economic activity may increase productivityonly where the concentrated industries are mature and ofrelatively large size. It is argued that diversification extern-alities are more pronounced in densely populated regions,whose economies have possibly entered the later stages ofdevelopment and integration, and mostly encompasshigh-technology, more adaptive and innovative industriesinteracting with each other through well-definedcooperation networks (Prager & Thisse, 2012; van derPanne, 2004).
Appendix B in the supplemental data online provides adescription of the construction of variables. Table C1 inAppendix C, also online, summarizes the main descriptivestatistics of the variables of the production function and thetechnical inefficiency model. Table C2, again online, illus-trates the correlation coefficients between the explanatoryvariables of the inefficiency model, reassuring us that noserious multicollinearity problem exists in the data set.
Equation (2) also includes a vector of region-specificdummies (Π) to account for time-invariant unobservedheterogeneity. We further include time dummies (T) tocontrol for common production shocks. The term wit
expresses a random variable, defined by the truncation ofthe normal distribution. All the parameters included inthe log-linear production function (1) and the inefficiencymodel (2), along with the model variances s2 = s2
v + s2u
and g = s2u/(s
2v + s2
u), are jointly estimated at one stageby using the maximum likelihood estimator. Based on alikelihood ratio test, rejection of the hypothesis g = 0 infavour of g . 0 would imply that deviations from the fron-tier are due to inefficiency effects.
MODEL ESTIMATION AND RESULTS
Econometric estimatesTable 1 reports baseline maximum-likelihood estimates ofequations (1) and (2). The inefficiency equation is simul-taneously estimated with the production function usingall regional explanatory variables in a stepwise manner.We can distinguish a significantly positive effect of physicalcapital and labour on output. This outcome is plausible andcompares well with the results of the relevant literature. Inorder to determine whether deviations from the estimatedfrontier are due to inefficiency effects, we test the hypoth-esis g = 0 against g . 0. The parameter g is significantlydifferent from zero, which implies that inefficiency effectsare present and that we should proceed with the estimationof parameters related to the sources of inefficiency.
1400 Theodore Tsekeris and Sotiris Papaioannou
REGIONAL STUDIES
Column 1 reports the results when the identificationstrategy involves only the variable of human capital. Theresults are in favour of a negative but statistically insignifi-cant influence of human capital on technical inefficiency.Columns 2–5 present econometric estimates after includ-ing successively in the model the variables of internal mar-ket potential and its dispersion (column 2), external marketpotential (column 3), interregional specialization (column4) and sectoral concentration (column 5). The majority ofeconometric estimates shown in columns 2–5 suggest anegative impact of the aforementioned variables on techni-cal inefficiency, which is also statistically significant.
Existing theoretical considerations are verified asregards the efficiency-enhancing impact of urbanization
economies, measured here by the average internal marketpotential index. This finding implies that there exist posi-tive externalities associated with the agglomeration of thetotal regional economy, arising from the average increaseof prefectures’ urban size and the average reduction oftransport costs, as proxied by the distance between prefec-tures’ urban centres within a region.
The outcome that the dispersion of urbanization econ-omies, as measured by the standard deviation of internalmarket potential across the prefectures of a region, signifi-cantly reduces inefficiency suggests that a polycentric(rather than a monocentric or primate city-based) structureof regional economies favours their efficiency. Similarly,improvement of interregional market access significantly
Table 1. Regional-level econometric estimates (baseline model specification).1 2 3 4 5
Production function
β0 0.89***
(3.70)
3.08***
(16.06)
1.49***
(24.64)
10.20***
(4.56)
2.84*
(1.76)
ln (Hours worked) 0.81***
(9.02)
0.62***
(13.20)
0.63***
(8.08)
0.35***
(4.78)
0.33***
(7.65)
ln (Total physical capital) 0.24***
(3.65)
0.23***
(4.29)
0.32***
(3.83)
0.26***
(4.54)
0.57***
(23.30)
Time trend 0.003
(0.72)
0.0008
(0.18)
–0.03***
(–5.52)
–0.01***
(–4.16)
–0.02***
(–6.10)
Inefficiency model
δ0 0.33**
(2.07)
0.27***
(5.70)
0.87***
(4.47)
1.75***
(5.78)
1.38***
(7.66)
Hours worked by highly skilled persons –0.74
(–0.88)
–0.30
(–1.49)
–0.40***
(–4.54)
–0.29**
(–2.17)
–0.40***
(–3.17)
Internal market potential –0.04***
(–3.10)
–0.03***
(–2.89)
–0.05***
(–6.10)
–0.05***
(–5.54)
Internal market potential dispersion –0.001
(–1.23)
–0.002
(–1.33)
–0.001*
(–1.92)
–0.001***
(–4.46)
External market potential –0.06***
(–3.19)
–0.06
(–1.40)
–0.09**
(–2.02)
Interregional specialization –0.87***
(–5.18)
–0.40***
(–2.75)
Sectoral concentration –0.94***
(–6.58)
Time effects Included Included Included Included Included
Region effects Included Included Included Included Included
σ2 (p-value) 0.01***
(6.56)
0.0005***
(5.35)
0.001***
(6.60)
0.0004***
(6.82)
0.0004***
(10.94)
γ (p-value) 0.05***
(4.50)
0.00007***
(3.29)
0.19***
(10.89)
0.01
(0.46)
0.06**
(2.20)
Log-likelihood 179.14 387.64 356.07 410.45 404.90
Observations 169 169 169 169 169
Notes: t-statistics are given in parentheses.***, ** and *Significance at 1%, 5% and 10% levels respectively.
Regional determinants of technical efficiency: evidence from the Greek economy 1401
REGIONAL STUDIES
contributes to the reduction of inefficiencies in Greekregions. This finding provides support to policy decisionson large-scale infrastructure investments, aiming to pro-mote interregional connectivity and accessibility acrossthe country, as these decisions are associated with increasedeconomies of scale and reduced transport costs.
The results provide supporting empirical evidence onthe significantly positive influence of specialization on effi-ciency. This outcome can be related to the better adoptionof new technologies, innovation and knowledge spilloversoriginating from the geographical concentration of firmswithin a sector. Moreover, the results suggest that sectoralconcentration (rather than diversification) within a regionhelps to reduce inefficiencies. This outcome contrastswith evidence already established for more advanced econ-omies, wherein a positive association between diversifica-tion and efficiency exists (Prager & Thisse, 2012; Siegelet al., 1995; van der Panne, 2004). This finding can beexplained by the intrinsic characteristics of the domesticproduction system and the highly heterogeneous spatialstructure of the Greek economy.
Specifically, although an increasing diversificationcould potentially produce favourable economic outcomes,most regions have an industrial base with limited varietyso that substantial diversification economies cannot berealized (Petrakos, Fotopoulos, & Kallioras, 2012).The most diversified region of Attiki is the most devel-oped and persistently the most efficient one. Nonethe-less, peripheral regions (especially Notio Aigaio, VoreioAigaio, Kriti and Dytiki Makedonia) whose economiesare traditionally concentrated in sectors which are natu-ral resource-based and outward-looking, in terms ofbeing more open and more integrated into the globaleconomy (such as agriculture, as long as it relates toagricultural exports, and tourism), have been found tobe more resilient to the crisis (Petrakos & Psycharis,2016; Psycharis et al., 2014) and had an overall betterefficiency performance than other regions with morediversified productive base (see also the results of thecomparative efficiency scores across regions below).Therefore, the sectoral concentration of local economiescan possibly allow regions to adjust better to the chan-ging competition conditions and create or expand theirnetworks to international markets, according to theirlocal comparative advantages.
As far as the impact of human capital is concerned,the results verify that it contributes significantly to redu-cing inefficiencies of Greek regions. Estimates that referto year- and region-specific dummy variables are shownin Table D1 of Appendix D in the supplemental dataonline. These estimates rely on the full baseline modelspecification shown in column 5 of Table 1. All esti-mated coefficients of these dummy variables (except foryear 2011) are statistically significant and reassure usthat time- and region-specific effects are correctlyincluded in the inefficiency model specification. The esti-mated coefficients of the time-dummy variables have anegative sign up to the year 2010, suggesting thatyear-specific factors exerted a positive influence on the
efficiency of the Greek economy up to 2010. However,these coefficient estimates switch to positive from 2011onwards, suggesting that the advent of the crisis hadnegative consequences on the efficiency of Greek regions.The statistical insignificance of the 2011 dummy coeffi-cient and the weak significance of the 2012 dummycoefficient can be justified on the basis that 2011–12 isa rather short period not allowing one truly to measurethe damaging effect of the Greek crisis.2 The fact thatall coefficient estimates of regional dummies enter witha positive and a statistically significant sign implies theexistence of region-specific effects (e.g., related to geo-graphical constraints), not captured by our model speci-fication, which exert a negative influence on technicalefficiency.
Robustness analysisNext, we check the robustness of baseline econometric esti-mates (Table 1) by proposing a number of alternative spe-cifications. To do so, we employ a set of additional variablesto represent the level and dispersion of urbanization econ-omies and to account for the influence of intra-regionalspecialization (see Appendix B in the supplemental dataonline).
Table 2 shows estimates of alternative model specifi-cations. The results verify the significantly positiveimpact on technical efficiency of urbanization and its dis-persion, as proxied by the regional population density andits generalized entropy index respectively (column 2).The statistical significance of these two variables denotesthe robustness of the estimated efficiency-enhancingimpact of urbanization economies and its dispersion.Conversely, intra-regional specialization does not seemto exert any significant influence on technical efficiency(columns 1–3). This outcome suggests that the geo-graphical concentration of economic activity within aregion does not considerably affect its efficiency, in con-trast to the geographical concentration in the interregio-nal scale (Table 1). Econometric estimates of allremaining variables in our model remain practicallyunchanged and retain their sign and statisticalsignificance.
For comparison purposes, the fourth column of Table 2reports the results of a translog econometric specification.These estimates suggest that the variables of external andinternal market potentials weakly affect regional efficiency.However, the g-coefficient of this specification lacks stat-istical significance and, hence, econometric estimates ofthe explanatory variables of this model should be con-sidered with caution.
Efficiency scores and the contribution ofregional variablesWe obtain predictions of efficiency scores by using the con-ditional expectation of technical efficiency of region i attime t, i.e., TEit¼exp {−Uit}. Based on Battese and Coelli(1995), TEit is obtained by estimating the conditionalexpectation of −Uit given the random variable
1402 Theodore Tsekeris and Sotiris Papaioannou
REGIONAL STUDIES
Table 2. Regional-level econometric estimates (alternative model specifications).Panel A: Alternative econometric models Panel B: Modelling for crisis effects
1 2 3 4 5 6 7 8 9 10
Production function
β0 5.71***
(3.06)
1.76***
(3.47)
3.57
(1.05)
0.91
(0.76)
1.94***
(6.64)
10.61***
(8.16)
8.81***
(6.59)
1.19***
(6.37)
1.61***
(7.21)
–0.91***
(–5.44)
ln (Hours worked) 0.23***
(3.63)
0.31***
(5.83)
0.30***
(5.28)
0.40***
(6.85)
0.32***
(6.07)
0.20***
(4.36)
0.29***
(8.66)
0.52
(1.36)
0.30***
(5.87)
0.85
(1.17)
ln (Total physical capital) 0.55***
(9.06)
0.63***
(14.41)
0.57***
(5.24)
0.59***
(26.12)
0.61***
(12.95)
0.38***
(7.38)
0.36***
(6.88)
0.47
(1.62)
0.64***
(13.85)
0.98***
(14.67)
ln (Hours worked) × ln (Hours
worked)
1.43
(0.41)
ln (Total physical capital) ×
ln (Total physical capital)
–1.78
(–0.36)
ln (Hours worked) × ln (Total
physical capital)
0.92
(1.37)
Time trend –0.02***
(–5.22)
–0.02***
(–6.25)
–0.02***
(–6.86)
–0.01
(–0.18)
–0.03***
(–6.12)
0.05***
(12.14)
0.01***
(4.11)
–0.02***
(–3.43)
–0.02***
(–2.50)
0.001
(0.43)
Inefficiency model
δ0 1.62***
(5.47)
0.83***
(4.56)
1.39***
(3.59)
1.13***
(14.41)
1.33***
(6.27)
2.05***
(10.88)
1.52***
(9.96)
0.19
(0.57)
1.09***
(5.63)
1.14***
(3.98)
Hours worked by highly skilled
persons
–0.12
(–0.89)
–0.41***
(–3.15)
–0.56***
(–4.45)
–0.26
(–0.21)
–0.32**
(–2.21)
–0.22*
(–1.91)
–0.33***
(–2.99)
–0.04
(–0.04)
–0.29*
(–2.37)
–0.26
(–1.07)
Hours worked by highly skilled
persons × Crisis
0.01
(0.14)
Population density –0.002***
(–2.82)
Population density dispersion –1.78**
(–2.05)
Interregional specialization –0.76***
(–4.87)
–0.02
(–0.01)
–0.33**
(–2.03)
–0.63***
(–4.06)
–0.41***
(–2.87)
–0.04
(–0.04)
–0.30*
(–1.90)
–0.16
(–0.53)
(Continued )
Regionaldeterminants
oftechnicalefficiency:evidence
fromthe
Greek
economy
1403
REGIONALSTU
DIES
Table 2. Continued.Panel A: Alternative econometric models Panel B: Modelling for crisis effects
1 2 3 4 5 6 7 8 9 10
Interregional specialization ×
Crisis
0.12***
(3.00)
Intra-regional specialization –0.01
(–0.02)
–0.07
(–0.22)
–0.27
(–0.77)
Sectoral concentration –1.07***
(–4.78)
–1.31***
(–5.26)
–1.31
(–0.50)
–1.07***
(–4.08)
–0.97***
(–3.92)
–0.78***
(–3.27)
–0.01
(–0.01)
–0.95***
(–3.76)
–0.37
(–0.62)
Sectoral concentration × Crisis 0.53
(1.48)
External market potential –0.09**
(–2.11)
0.03
(0.77)
–0.10**
(–2.32)
–0.10*
(–1.91)
–0.10**
(–2.20)
–0.10***
(–3.20)
–0.09***
(–3.04)
–0.003
(–0.03)
–0.07*
(–1.72)
–0.17***
(–2.97)
External market potential ×
Crisis
0.002
(0.03)
Internal market potential –0.04***
(–4.99)
–0.05***
(–4.66)
–0.04*
(–1.90)
–0.05***
(–5.09)
–0.06***
(–8.10)
–0.05***
(–8.20)
–0.01
(–0.49)
–0.04***
(–5.31)
–0.04***
(–3.23)
Internal market potential ×
Crisis
0.001*
(1.78)
Internal market potential
dispersion
–0.002***
(–2.88)
–0.0007
(–0.79)
–0.003
(–1.09)
–0.002**
(–2.29)
–0.001
(–1.20)
–0.003***
(–4.36)
–0.0002
(–0.05)
–0.002**
(–2.41)
–0.002
(–1.22)
Internal market potential
dispersion × Crisis
–0.001***
(–4.84)
Time effects Included Included Included Included Included Included Included Included Included Included
Region effects Included Included Included Included Included Included Included Included Included Included
σ2 (p-value) 0.0005***
(9.00)
0.0006***
(9.13)
0.0004***
(5.31)
0.0005***
(5.71)
0.0005***
(8.23)
0.0003***
(9.26)
0.0003***
(7.31)
0.01***
(4.45)
0.0005***
(7.93)
0.0007***
(5.43)
γ (p-value) 0.02
(0.51)
0.000006*
(1.67)
0.02
(0.12)
0.01
(0.02)
0.16***
(7.40)
0.91***
(81.52)
0.00
(0.02)
0.18
(1.45)
0.12***
(3.84)
0.31***
(12.84)
Log-likelihood 402.67 384.74 395.42 412.08 404.66 427.26 434.67 134.52 408.66 346.63
Observations 169 169 169 169 169 169 169 169 169 169
Notes: t-statistics are given in parentheses.***, ** and *Significance at 1%, 5% and 10% levels respectively.
1404Theodore
Tsekerisand
SotirisPapaioannou
REGIONALSTU
DIES
1it = Vit − Uit :
TÊit = E [ exp (−Uit)|1it]
= exp −mit +1
2�s2
[ ]{ }
· Fmit
�s− �s
[ ]/F
mit
�s
[ ]{ }, (3)
where F(.) is the distribution function of the standard nor-mal random variable 1it ,
mit = (1− g) · [d0 + dHCHCit + dIMPIMPit
+ dDIMPDIMPit + dEMPEMPit + dSSit
+ dHHHHit + dPPi + dTT ]− g 1it ,
and �s2 = g(1− g)s2.Table 3 provides a comparison of efficiency scores
between regions across the study period. It is clear thatthere exist significant and persistent disparities in the levelsof efficiency across Greek regions. The by far most efficientregion is Attiki, with an average score of 98.5%, followed(in order) by Notio Aigaio and Dytiki Ellada, with averageefficiencies of 73.5% and 66.6% respectively. In contrast,Sterea Ellada and Peloponnisos are the least efficientregions, with average scores of 47.7% and 52.7% respect-ively. This outcome is consistent with the fact that Greeceis characterized by significant economic disparities amongthe core region of Attiki and peripheral areas and by aslow interregional (β) convergence process (see AppendixE in the supplemental data online).
Up to 2007, efficiency scores were constantly risingacross all regions. With the exception of Attiki, whichexploits its advantageous central position as the largestdominant market to create economies of scale, all otherregions witnessed a drop in their efficiency performancefrom 2008 onwards, following the crisis outbreak. Onlythe island regions of Notio Aigaio and Ionia Nisia showeda negligible recovery in their efficiency score in 2012. Suchoutcomes signify the remarkably adverse impact of theeconomic crisis on productivity of the peripheral areas ofthe country.
We further explore to what extent each regional variablecontributes to the improvement of efficiency across regionsof the Greek economy.3 To do so, we first evaluate effi-ciency levels of each region by successively clearing outthe influence of each factor. To obtain such measures ofnet technical efficiency (net of factor influences), we replacethe term:
[d0 + dHCHCit + dIMPIMPit + dDIMPDIMPit
+ dEMPEMPit + dSSit + dHHHHit + dPPi + dTT ]
in equation (3), with the following term:
[d0 + dHCHCit + dIMPIMPit + dDIMPDIMPit
+ dEMPEMPit + dSSit + dHHHHit + dPPi + dTT
− dZZit],
where Z stands for the regional explanatory variable whose
influence is measured. The differences between gross andnet efficiency scores represent the contribution of eachregional variable to technical efficiency and are shown inTable 3.
On average, the variable of external market potentialhas the largest contribution to technical efficiency (29%),followed by the variable of internal market potential(14.4%), whose impact, however, exhibits the highest vari-ation among regions. This outcome stresses the relativelyincreased importance of the interregional market accessand urbanization economies on the efficiency of Greekregions. By and large, the best-performing regions ofAttiki, Notio Aigaio and Dytiki Ellada as well as theregions of Dytiki Makedonia and Kentriki Makedoniareceive the highest contribution. Attiki exploits the highestbenefits from use of human capital (8.3%) and internalmarket access (90.6%). This is because it comprises a singlehighly urbanized prefecture and its human capital is thehighest among Greek regions.
The variables of external market potential and sectoralconcentration exert their highest influence on the islandregion of Notio Aigaio (46.8% and 16.4% respectively).This finding underlines the increased importance for thespecific region’s efficiency of the accessibility to/fromother regions and its reliance on a limited range of econ-omic activities (mostly related to the sector of trade, trans-port and tourism).
The second largest contribution (following Attiki) ofthe variable of internal market potential is observed in theregion of Kentriki Makedonia (17.8%). This is mainlydue to the increased urban concentration in the prefectureof Thessaloniki. Also, the dispersion of internal marketpotential has the largest impact on this region (8.2%), ascompared with all other regions, due to the relativelybalanced development of urban concentrations in the pre-fectures around the prefecture of Thessaloniki. Finally, weobserve the largest influence of the variable of interregionalspecialization on efficiency (13%) in Dytiki Makedonia,due to its increased specialization (especially of the prefec-ture of Kozani) in the production of electricity.
Region- and time-specific effects on efficiency are alsoplotted in the last three columns of Table 3. In sum, theresults imply that there exist considerable regional andtime effects, which are larger for the year 2012 and overallcontribute to a lower level of technical efficiency. Havingshown that the time dummies for the years 2011 and2012 exert a positive influence on inefficiency (see TableD1 in Appendix D in the supplemental data online), weextend the analysis to explore whether the crisis has affectedthe influence of our baseline variables. To do so, we intro-duce interaction terms between these determinants and atime dummy proxying for the period 2011–12 (this variablereceives ones for the years 2011 and 2012, and zero other-wise). The right part of Table 2 reports econometric resultswith these interaction terms, which indicate that most ofthe baseline variables still exert a significantly negativeinfluence on regional inefficiency. They also demonstratethat the crisis counters the negative influence on ineffi-ciency of internal market potential (at the 10% significance
Regional determinants of technical efficiency: evidence from the Greek economy 1405
REGIONAL STUDIES
Table 3. Comparative technical efficiency scores (%) and relative contribution (%) of each determinant.
Region
Panel A: Efficiency scores across regions andyears (%) Panel B: Relative contribution of each determinant on technical efficiency (%)
Efficiency2000
Efficiency2008
Efficiency2012
Averageefficiency(2000–12)
Hoursworked
byhighlyskilledpersons
Interregionalspecialization
Sectoralconcentration
Externalmarketpotential
Internalmarketpotential
Dispersionof internalmarketpotential
Year2011
Year2012
Regionaleffect
Anatoliki
Makedonia-
Thraki
53.77 63.96 55.57 59.62 4.84 1.29 8.47 17.53 6.37 1.34 –1.37 –2.65
Kentriki
Makedonia
55.32 67.42 55.25 61.97 6.92 0.52 8.60 23.43 17.83 8.23 –1.37 –2.64 –33.59
Dytiki
Makedonia
55.86 67.00 55.66 63.58 5.42 13.03 10.18 31.71 4.93 1.91 –1.38 –2.66 –33.97
Ipeiros 53.35 60.87 51.89 57.65 4.89 1.14 8.29 25.15 4.42 0.86 –1.27 –2.48 –14.19
Thessalia 56.56 66.55 55.93 62.20 5.69 1.71 8.27 32.17 8.49 3.08 –1.36 –2.67 –31.24
Ionia Nisia 59.69 67.53 54.36 63.15 4.11 3.48 13.41 27.05 7.68 0.67 –1.30 –2.59 –19.13
Dytiki Ellada 60.29 71.79 60.41 66.60 4.73 1.22 8.74 37.73 9.46 2.31 –1.48 –2.88 –30.46
Sterea Ellada 45.34 50.46 42.36 47.70 3.31 3.25 6.83 28.70 5.79 1.36 –1.03 –2.02 –52.12
Attiki 90.96 100.00 100.00 98.50 8.29 0.56 5.66 26.11 90.59 0.03 0.04 0.03
Peloponnisos 47.72 56.20 47.37 52.67 3.61 1.12 5.59 28.78 5.81 1.27 –1.17 –2.26 –35.06
Voreio
Aigaio
57.70 71.12 58.35 64.17 5.86 1.67 10.74 27.05 4.70 0.45 –1.42 –2.79 –8.42
Notio Aigaio 67.85 80.76 66.79 73.52 5.51 4.31 16.43 46.83 10.37 0.18 –1.58 –3.19 –36.87
Kriti 58.45 70.22 56.92 64.75 5.51 1.40 9.77 24.79 10.70 2.63 –1.40 –2.72 –11.09
Mean
contribution
5.28 2.67 9.31 29.00 14.40 1.87 –1.24 –2.42 –26.93
Note: Regional effects could not be computed for Attiki as it is the region with the highest efficiency score. They could not also be calculated for Anatoliki Makedonia-Thraki because it is used in the analysis as the base region.
1406Theodore
Tsekerisand
SotirisPapaioannou
REGIONALSTU
DIES
level) and interregional specialization (at the 1% signifi-cance level). However, the crisis entails a significantly (atthe 1% level) negative impact of internal market potentialdispersion on technical inefficiency.
FURTHER DISCUSSION
The regression estimates shown in Tables 1 and 2 implythat regional production exhibits decreasing returns toscale, as the sum of production function coefficients (capitaland labour) is lower than 1 (close to 0.9). Though this esti-mate is not too far from 1, it is in favour of a less pro-portional increase of output after an increase inproduction inputs. Despite that this outcome is observedfor a period during which Greece achieved relatively highgrowth rates, it can be attributed to misallocation of labourand capital inputs towards less productive sectors of theeconomy (Kollintzas, Papageorgiou, Tsionas, & Vassilatos,2016), which, in turn, led to a lower-than-optimal exploi-tation of productive capacity.
It is also noteworthy that the time-trend coefficientestimates are negative and statistically significant in mostregression results shown in Tables 1 and 2. Although –to the best of our knowledge – empirical evidence in favourof technological backwardness of Greek regions is notavailable, the period under investigation is characterizedby a significant drop of TFP growth rates from 2008onwards. Gogos, Mylonidis, Papageorgiou, and Vassilatos(2014) showed that TFP growth of the Greek economycollapsed from 3.1% in the period from 2001 to 2007 to–5.97% during 2007–12.
Finally, the efficiency scores reported in Table 3 showthat Attiki remained the most efficient region, while itsefficiency score amounted to 100% for a large number ofyears, even during the crisis. Although this is a realisticcase if seen in comparison with other regions of Greece,the zero inefficiency score for so many consecutive yearsis likely to be attributed to inherent assumptions pertainingto our stochastic frontier model. Kumbhakar, Parmeter,and Tsionas (2013) recently constructed a stochastic fron-tier model with inhomogeneous distribution of ineffi-ciency, which realistically assumes the existence of a largenumber of production units that exhibit zero inefficiency.
CONCLUSIONS
In this paper we estimated technical efficiency scores andidentified sources of inefficiency of Greek regions. Thefindings revealed the existence of significant disparities inthe levels of efficiency across regions. While the efficiencyscores were constantly rising across all regions up to2007, a considerable drop is observed from 2008 onwards,with the exception of the leading metropolitan region ofAttiki, signifying the adverse impact of the economic crisison the productivity of peripheral areas.
It was shown that spatial variations in efficiency can beattributed to factors related to the interregional marketaccess, level and dispersion of urbanization economies,specialization, sectoral concentration and human capital in
each region. It should be further stressed that, in additionto physical and human resource-factor endowments andgeographical characteristics, institutional and political fac-tors may have also led to differences in regional efficiency.As in the case of many other countries worldwide (Acemo-glu, 2008; Rodríguez-Posé, 2013; Rodríguez-Posé & DiCataldo, 2015; Rodríguez-Posé & Garcilazo, 2015), thepoor quality of government institutions and policies inGreece at the national and regional levels (e.g., see theEuro-pean Quality of Government Index; Charron et al., 2016)had an adverse impact on the efficacy of factors influencingthe economic performance in the long run (Alogoskoufis,1995).
Moreover, investment decisions in the country arehighly centralized, with sometimes conflicting objectivesand failures to coordinate policies and exploit synergiesacross tiers of government, deviating from criteria of econ-omic efficiency, as they often embrace equity and politicalrent-seeking considerations (Rodríguez-Posé, Psycharis,& Tselios, 2016; Tsekeris, 2014). Note that a decentraliza-tion of management to regional authorities was legislated in2011, but the overlapping of responsibilities, lack ofadministrative/enforcement capacity and absence of avail-able resources due to fiscal constraints deterred regionsfrom exercising local-specific policies. In the frameworkof national reform policies currently made to promote anew growth model, Greek regions should undertake insti-tutional arrangements suitable to their own needs and com-parative advantages to prioritize and coordinate theallocation of investment resources according to wherethey can provide the largest efficiency gains.
Our estimates show that enhanced interregional marketaccess exerts a notably high and positive contribution totechnical efficiency of all regions, but those mostly affectedrefer to the most developed peripheral regions (NotioAigaio and Dytiki Ellada). It is shown that the contri-bution of the intra-regional market access, which proxiesfor the urbanization economies, is lower in the peripheralregions, as compared with the influence of the externalmarket access. Therefore, in terms of transport investmentpolicy, it seems that interregional transport improvementsshould be prioritized over intra-regional transport improve-ments, as the former affect more decisively the regional effi-ciencies and disparities of the Greek economy.
The contribution of the dispersion of internal marketaccess mostly affects the inefficiency of regions pertainingto some kind of polycentric development, especially theregion of Kentriki Makedonia. The positive impact of thelatter factor on technical efficiency denotes the importanceof strategic regional planning to promote a more balancedspatial development, which is not dominated by a singlelarge urban agglomeration but is functionally organizedin a polycentric shape, through a territorial network ofurban centres around the principal city of each region.The upcoming completion of the main highway and rail-way network will enhance both the inter- and intra-regional market potential and the polycentric developmentof each region’s urban system. Regarding the island areas,the development of a hub-and-spoke connectivity system
Regional determinants of technical efficiency: evidence from the Greek economy 1407
REGIONAL STUDIES
around the principal island of each region will foster theintra-regional market potential, reducing the reliance onthe mainland metropolitan-area ports (in Attiki and Thes-saloniki) and the cost inefficiencies of the part of theregional transport system operating public service-obli-gation routes.
The econometric results also illustrated the substantialinfluence of productive specialization among regions andof sectoral concentration (instead of diversification) withina region. This outcome suggests the significant role of thedevelopment of local activity clusters, in the form of indus-trial areas, of science and technology parks, and of logisticsparks to promote innovation through knowledge spilloversand to create productivity gains through increasing returnsto scale in production originating from the specializationand concentration of high value-added goods and services.The fact that government policies in favour of a more ‘even’regional development, through either the establishment ofindustrial areas or the provision of investment grants andother allowances, were not successful can be largely attrib-uted to the dominance of market or efficiency-orientedfirm location processes (Labrianidis & Papamichos,1990). A combination of investment incentives andenvironmental restrictions, as well as lower land rent andrelatively easy access to amenities, contributed to the con-centration of industrial activities in the metropolitanregions of Attiki and Thessaloniki, and their neighbouringprefectures. Therefore, appropriate policies to stimulate thedevelopment of local activity clusters should be bettermatched with local entrepreneurial activities and with thecomparative advantages of each region. Based on currentfindings, the efficient development of industrial areasshould rely on the type and degree of specialization, givingpriority to central locations of mainland peripheral areaswith increased intra-regional accessibility and urbanizationeconomies which diminish technical inefficiencies.
We further demonstrated that human capital exerts aconsiderable influence on reducing inefficiencies of Greekregions. The positive association between the increasededucational attainment levels of the labour force andregional efficiency suggests that higher productivity is likelyto be achieved through investment in education and train-ing. It should be emphasized that human capital enhance-ment involves not only the upgrading of the highereducation system but also regionally targeted actions togenerate, receive and absorb a high-quality labour forceand innovation according to local conditions. The regionswhere human capital improvements would bring aboutthe largest benefits in efficiency are those of Attiki andKentriki Makedonia, where the majority of universitiesand research institutes are located. In the current circum-stances of fiscal austerity, emphasis should be given toqualitative features of education (e.g., cooperation betweenresearch institutes and local companies, and lifelong learn-ing) and spatial spillovers, which are both regarded asimportant determinants of the human capital impact onregional productivity (Benos & Karagiannis, 2016), ratherthan increasing the number of higher education enrolmentsand building new universities or university departments to
serve local needs. Finally, the outcome that region-specificeffects do significantly affect inefficiency indicates the needto emphasize the exploitation of the particular local com-parative advantages, or to eliminate possible drawbacks toincrease efficiency performance.
All results mentioned above highlight the need for thecomprehensive treatment of a wide range of spatial sourcesof inefficiency to support the fast recovery and sustainablegrowth of Greek regions. Especially, they underline theneed to take up immediate and targeted measures, suchas effective regional-sectoral policies to increase efficiencyand to reduce the widening productivity gaps betweenthe core (Attiki) and peripheral regions. Given the severeresource limitations due to the austerity measures and fiscalconstraints, policies that enhance the efficient organizationof economic activities in space and the regional reallocationof public funds are indispensable to increase the rate ofreturn on investment.
ACKNOWLEDGEMENTS
The authors thank an associate editor and two anonymousreferees for their valuable comments and suggestions.
DISCLOSURE STATEMENT
No potential conflict of interest was reported by theauthors.
SUPPLEMENTAL DATA
Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/00343404.2017.1390312.
NOTES
1. NUTS ¼ Nomenclature des Unités TerritorialesStatistiques.2. Table 3 illustrates the adverse influence of the recessionon the efficiency performance of Greek regions.3. Similar attempts can be found in the microeconomicliterature (e.g., Coelli, Perelman, & Romano, 1999).
ORCID
Theodore Tsekeris http://orcid.org/0000-0002-7231-6077Sotiris Papaioannou http://orcid.org/0000-0003-1862-9694
REFERENCES
Acemoglu, D. (2008). Introduction to modern economic growth.Princeton: Princeton University Press.
Aiginger, K., & Davies, S. W. (2004). Industrial specialization andgeographic concentration: Two sides for the same coin? Not forthe European Union. Journal of Applied Economics, 7(2), 231–248.
1408 Theodore Tsekeris and Sotiris Papaioannou
REGIONAL STUDIES
Alogoskoufis, G. (1995). The two faces of Janus: Institutions, policyregimes and macroeconomic performance in Greece. EconomicPolicy, 10(20), 147–192. doi:10.2307/1344540
Battese, G. E., & Coelli, T. J. (1995). A model for technical ineffi-ciency effects in a stochastic frontier production function forpanel data. Empirical Economics, 20(2), 325–332. doi:10.1007/BF01205442
Becker,G. (1993).Human capital: A theoretical and empirical analysis,with
special reference to education. Chicago: University of Chicago Press.Beeson, P. E., &Husted, S. (1989). Patterns and determinants of pro-
ductive efficiency in the state manufacturing. Journal of RegionalScience, 29(1), 15–28. doi:10.1111/j.1467-9787.1989.tb01219.x
Benos, N., & Karagiannis, S. (2016). Do education quality and spil-lovers matter? Evidence on human capital and productivity inGreece. Economic Modelling, 54, 563–573. doi:10.1016/j.econmod.2016.01.015
Charron, N., Dahlberg, S., Holmberg, S., Rothstein, B., Khomenko,A., & Svensson, R. (2016). The quality of government EU regional
dataset (September 16 Version). Gothenburg: The Quality ofGovernment Institute, University of Gothenburg. Retrievedfrom http://www.qog.pol.gu.se
Coelli, T., Perelman, S., & Romano, E. (1999). Accounting forenvironmental influences in stochastic frontier models: Withapplication to international airlines. Journal of Productivity
Analysis, 11(3), 251–273. doi:10.1023/A:1007794121363Driffield, N., & Munday, M. (2001). Foreign manufacturing,
regional agglomeration and technical efficiency in UK industries:A stochastic production frontier approach. Regional Studies,35(5), 391–399. doi:10.1080/713693833
Floros, C., Voulgaris, Z., & Lemonakis, C. (2014). Regional firmperformance: The case of Greece. Procedia Economics and
Finance, 14, 210–219. doi:10.1016/S2212-5671(14)00704-7Frenken, K., Van Oort, F., & Verburg, T. (2007). Related variety,
unrelated variety and regional economic growth. Regional
Studies, 41(5), 685–697. doi:10.1080/00343400601120296Fritsch, M., & Slavtchev, V. (2011). Determinants of the efficiency of
regional innovation systems. Regional Studies, 45(7), 905–918.doi:10.1080/00343400802251494
Glaeser, E. L., Kallal, H. D., Scheinkman, J. A., & Shleifer, A.(1992). Growth in cities. Journal of Political Economy, 100(6),1126–1152. doi:10.1086/261856
Gogos, S. G., Mylonidis, N., Papageorgiou, D., & Vassilatos, V.(2014). 1979–2001: A Greek great depression through the lensof neoclassical growth theory. Economic Modelling, 36, 316–331.doi:10.1016/j.econmod.2013.09.031
Haughwout, A. F. (1998). Aggregate production functions, interre-gional equilibrium, and the measurement of infrastructure pro-ductivity. Journal of Urban Economics, 44(2), 216–227. doi:10.1006/juec.1997.2066
Holmes, T. J., & Stevens, J. J. (2002). Geographic concentration andestablishment scale. Review of Economics and Statistics, 84(4),682–690. doi:10.1162/003465302760556495
Jacobs, J. (1969). The economy of cities. New York: Vintage.Ke, S., & Yu, Y. (2014). The pathways from industrial agglomeration
to TFP growth – The experience of Chinese cities for 2001–2010. Journal of the Asia Pacific Economy, 19(2), 310–332.doi:10.1080/13547860.2014.880286
Kollintzas, T., Papageorgiou, D., Tsionas, E., &Vassilatos, V. (2016).Market and political power interactions in Greece: An empirical
investigation (Economics Department Working Paper No. 6-2016). Athens: Athens University of Economics and Business.
Kumbhakar, S. C., Parmeter, C. F., & Tsionas, E. G. (2013). A zeroinefficiency stochastic frontier model. Journal of Econometrics, 172(1), 66–76. doi:10.1016/j.jeconom.2012.08.021
Labrianidis,L.,&Papamichos,N. (1990).Regionaldistributionof indus-try and the role of the state in Greece. Environment and Planning C:Government and Policy, 8(4), 455–476. doi:10.1068/c080455
Lee, B. S., Jang, S., & Hong, S. H. (2010). Marshall’s scale econom-ies and Jacobs’ externality in Korea: The role of age, size and thelegal form of organisation of establishments. Urban Studies, 47(14), 3131–3156. doi:10.1177/0042098009359953
Louri, H. (1988). Urban growth and productivity: The case of Greece.Urban Studies, 25(5), 433–438. doi:10.1080/00420988820080551
McCoy, K., &Moomaw, R. L. (1995). Determinants of manufactur-ing efficiency in Canadian cities: A stochastic frontier approach.Review of Regional Studies, 25(3), 317–330.
Mikelbank, B. A., & Jackson, R.W. (2000). The role of space in pub-lic capital research. International Regional Science Review, 23(3),235–258. doi:10.1177/016001700761012774
Mitra, A. (1999). Agglomeration economies as manifested in techni-cal efficiency at the firm level. Journal of Urban Economics, 45(3),490–500. doi:10.1006/juec.1998.2100
Mitra, A. (2000). Total factor productivity growth and urbanizationeconomies: A case of Indian industries. Review of Urban and
Regional Development Studies, 12(2), 97–108. doi:10.1111/1467-940X.00023
Otsuka, A., & Goto, M. (2015). Regional policy and the productiveefficiency of Japanese industries. Regional Studies, 49(4), 518–531. doi:10.1080/00343404.2013.795285
Otsuka, A., Goto, M., & Sueyoshi, T. (2010). Industrial agglomera-tion effects in Japan: Productive efficiency, market access, andpublic fiscal transfer. Papers in Regional Science, 89(4), 819–840.doi:10.1111/j.1435-5957.2010.00286.x
Petrakos, G., Fotopoulos, G., & Kallioras, D. (2012). Peripheralityand integration: Industrial growth and decline in the Greekregions. Environment and Planning C: Government and Policy,30(2), 347–361. doi:10.1068/c10171r
Petrakos, G., & Psycharis, Y. (2016). The spatial aspects of economiccrisis in Greece. Cambridge Journal of Regions, Economy and
Society, 9(1), 137–152. doi:10.1093/cjres/rsv028Prager, J. C., & Thisse, J. F. (2012). Economic geography and the
unequal development of regions. Oxford: Routledge.Psycharis, Y., Rovolis, A., Tselios, V., & Pantazis, P. (2014).
Economic crisis and regional development in Greece. Région et
Développement, 39, 67–85.Rodríguez-Posé, A. (2013). Do institutions matter for regional devel-
opment? Regional Studies, 47(7), 1034–1047. doi:10.1080/00343404.2012.748978
Rodríguez-Posé, A., &DiCataldo,M. (2015).Quality of governmentand innovative performance in the regions of Europe. Journal ofEconomic Geography, 15(4), 673–706. doi:10.1093/jeg/lbu023
Rodríguez-Posé, A., & Garcilazo, E. (2015). Quality of governmentand the returns of investment: Examining the impact of cohesionexpenditure in European regions. Regional Studies, 49(8), 1274–1290. doi:10.1080/00343404.2015.1007933
Rodríguez-Posé, A., Psycharis, Y., & Tselios, V. (2016). Politics andinvestment: Examining the territorial allocation of public invest-ment in Greece. Regional Studies, 50(7), 1097–1112. doi:10.1080/00343404.2015.1009027
Romp, W., & De Haan, J. (2007). Public capital and economicgrowth: A critical survey. Perspektiven der Wirtschaftspolitik,8(S1), 6–52. doi:10.1111/j.1468-2516.2007.00242.x
Siegel, P. B., Johnson, T. G., & Alwang, J. (1995). Regional econ-omic diversity and diversification. Growth and Change, 26(2),261–284. doi:10.1111/j.1468-2257.1995.tb00171.x
Tsekeris, T. (2014). Multi-sectoral interdependencies of regionalpublic infrastructure investments. Socio-Economic Planning
Sciences, 48(4), 263–272. doi:10.1016/j.seps.2014.06.001Vagionis, N., & Spence, N. (1994). Total factor regional productivity
in Greece. Environment and Planning C: Government and Policy,12(4), 383–407. doi:10.1068/c120383
Van der Panne, G. (2004). Agglomeration externalities: Marshall ver-sus Jacobs. Journal of Evolutionary Economics, 14(5), 593–604.doi:10.1007/s00191-004-0232-x
Regional determinants of technical efficiency: evidence from the Greek economy 1409
REGIONAL STUDIES
Abstract
INTRODUCTION
RELATED LITERATURE
ECONOMETRIC METHODOLOGY
MODEL ESTIMATION AND RESULTS
Econometric estimates
Robustness analysis
Efficiency scores and the contribution of regional variables
FURTHER DISCUSSION
CONCLUSIONS
ACKNOWLEDGEMENTS
DISCLOSURE STATEMENT
SUPPLEMENTAL DATA
NOTES
ORCID
REFERENCES
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles false /AutoRotatePages /PageByPage /Binding /Left /CalGrayProfile () /CalRGBProfile (Adobe RGB 50199851) /CalCMYKProfile (U.S. Web Coated 50SWOP51 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.1000 /ColorConversionStrategy /sRGB /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments false /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo false /PreserveFlatness true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings false /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict <> /ColorImageDict <> /JPEG2000ColorACSImageDict <> /JPEG2000ColorImageDict <> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict <> /GrayImageDict <> /JPEG2000GrayACSImageDict <> /JPEG2000GrayImageDict <> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Average /MonoImageResolution 300 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict <> /AllowPSXObjects true /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /Description <>>> setdistillerparams<> setpagedevice
Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=cgse20
Globalisation, Societies and Education
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/cgse20
Quality assurance in Greek Higher Education andthe imperative to use English
Areti Vogopoulou, Antigone Sarakinioti & Anna Tsatsaroni
To cite this article: Areti Vogopoulou, Antigone Sarakinioti & Anna Tsatsaroni (2022): Qualityassurance in Greek Higher Education and the imperative to use English, Globalisation, Societiesand Education, DOI: 10.1080/14767724.2022.2048797
To link to this article: https://doi.org/10.1080/14767724.2022.2048797
Published online: 08 Mar 2022.
Submit your article to this journal
Article views: 106
View related articles
View Crossmark data
Quality assurance in Greek Higher Education and the imperativeto use EnglishAreti Vogopoulou a, Antigone Sarakiniotib and Anna Tsatsaronic
aDepartment of Food Science and Technology, University of the Peloponnese, Kalamata, Greece; bSchool ofPhilosophy and Education, Aristotle University of Thessaloniki, Thessaloniki, Greece; cDepartment of Social andEducation Policy, University of the Peloponnese, Corinth, Greece
ABSTRACTThis article examines how dominant discourses disseminated throughquality assurance processes in Greek Higher Education (re)defineacademic work and language use in a globalised HE field. Drawing onBernstein’s theory of pedagogic discourse, we approach government-led external evaluations of university departments as official policy textsthat encode the evaluative criteria of quality institutions and valorisespecific forms of knowledge and conduct. Our analysis of empiricalmaterial derived from selected HE departments illustrates thatperformance measurement systems interlink English as a defaultlanguage with mobility and research activity in a unified discourse ofquality constructing a selective understanding of how Higher EducationInstitutions and academics are expected to attain international visibility.
ARTICLE HISTORYReceived 27 July 2021Accepted 26 February 2022
KEYWORDSEuropean higher educationarea; mobility;internationalisation; Englishlanguage dominance;pedagogic discourse;academic practices andidentities
Introduction
European policies and priorities on higher education (HE), constructed and disseminated mainlywithin the European Higher Education Area (EHEA), encompass quality assurance (QA) processesas essential activities for Higher Education Institutions (HEIs). The declared aim is to enhance qual-ity in teaching and learning and establish new forms of accountability mechanisms (Brennan andShah 2000) assumed to be efficient and transparent (Shore and Wright 2015). To this end, common‘standards’ and ‘guidelines’ included in the Standards and Guidelines for Quality Assurance in theEuropean Higher Education Area (2015) present a powerful governance tool through the EuropeanAssociation for Quality in Higher Education (ENQA) which is coupled with the rise and institutio-nalisation of formalised evaluation agencies and processes at the national level of member states(Sarakinioti and Philippou 2020).
As Gornitzka and Stensaker (2014) argue, a regime in the regulation of quality has emergedtaking the form of an evolutionary process involving multiple agents and experts (Grek 2013), con-tributing to discourses on what education is and how it should function. Within this regime, QAand its related mechanisms of ‘steering from a distance’ through metrics and evidence-basedmodes of accountability, have been introduced and consolidated as a central element of the ‘edu-cational governance’ turn in the European education space (Ozga 2009). What is significant, how-ever, is that performance regimes do not simply refer to the implementation of specific tools formeasuring performances but mainly reshape the organisations’ and agents’ practices and identitiesaccording to their own expectations and demands (Shore and Wright 2015). From a Bernsteinianperspective, these mechanisms can be seen as embodying the rules of legitimate forms of academic
© 2022 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Areti Vogopoulou a.vogopoulou@go.uop.gr Department of Food Science and Technology, University of thePeloponnese, Kalamata, Greece
GLOBALISATION, SOCIETIES AND EDUCATIONhttps://doi.org/10.1080/14767724.2022.2048797
work (Bernstein 2000). Therefore, they contribute to discourses on the role and mission of HE by‘pedagogising’ academics into new forms of conduct (Middleton 2008; Beck and Young 2005)related to the reshaping of central aspects of academic work such as the nature of academic research(Elton 2000; Hazelkorn 2015) and the form of academic identities (Middleton 2008; Harley 2002).
In light of the above, the present article focuses on the ‘evaluation turn’ in Greek HE, initiated in2005 (Law 3374/2005) and accelerated by the introduction of a neoliberally oriented legal frame-work in 2011 (Gouvias 2012; Gounari 2012) entitled ‘Structure, Operation, Quality Assurance ofStudies and the Internationalisation of Institutions of Higher Education’. Our aim is to analysehow the discourses articulated in and through the implementation of institutionalised QA mech-anisms (re-)define academic work. In particular, we focus on the imperative to use English forteaching and research purposes externally imposed on Greek academics by the quality assuranceprocesses enacted in the Greek context.
In our study we analyse external evaluations reports drafted by independent committees of inter-national experts acting as external evaluators of HE institutions in the framework of QA. At theor-etical level, we approach these texts produced within the context of QA processes as discursiverelays of the rules for achieving high levels of ‘quality’ in Greek HEIs and at the same time as devicesfunctioning ‘pedagogically’ to orient academics and institutions towards what counts as legitimateacademic and pedagogic practice in HEIs (Bernstein 2000). In particular, we approach the finalreports of external evaluation processes as official documents that co-articulate the dominant dis-courses on HE reform and quality enhancement, focusing on the power/knowledge relationsinscribed in and through these discourses as well as the extent to which they signal significantchanges in perceptions of the role and mission of HE.
Following a short introduction on QA processes in Greek HEIs, we present the theoretical rootsof our study, the methodology and the observations of our empirical research resulting from theanalysis of external evaluation reports produced for fifteen departments of five Greek HE insti-tution. In the discussion of our findings, it becomes evident that the discourse on language usearticulated in and through the external evaluation reports brings together elements of dominantdiscourses on mobility, research activities and international visibility through high impact publi-cations while the shifts in knowledge/power relations that occur orient academics to new formsof practice and to ‘market-driven’ forms of identity.
Quality Assurance and HE transformations: where do language policies andpractices stand?
Quality and quality assurance processes are a fundamental component of the Bologna discourse onthe construction and maintenance of the EHEA. In addition, they occupy a central position in thenarratives about the role of HE in contemporary society, configured around the idea that HEIsshould contribute to the development of both national and regional economies, and ‘trainable’human capital as well as accrue economic and symbolic benefits to institutions themselves (Slaugh-ter and Rhoades 2004). In this understanding of the mission and role of HE, profit and prestigeappear as significant driving forces for change in the field of HE (Marginson 2007, 2016), introdu-cing thus the market imperatives of competition, efficiency and excellence (Naidoo 2016) alongwith the values and practices of New Public Management (Ball 2008). Connected to these mar-ket-related imperatives is the emphasis placed on performance measurement and its accompanyingmechanisms which are assumed to produce objective representations of the quality of both individ-uals’ and organisations’ outputs (Ball 2008). In accordance, European countries have developed HEsystems of evaluation and quality assurance as a means of promoting soft governance towardsaccountability (Ozga et al. 2011; Ozga 2012) and the prevailing culture of performativity (Ball 2012).
This managerialist view of quality is distorting traditional notions of responsibility in HE whilethe autonomy of universities is being eroded by the introduction of external official monitoring.Several studies have documented these developments by illustrating how QA processes function
2 A. VOGOPOULOU ET AL.
as restructuring mechanisms that transform university governance patterns, re-establish powerrelations (Landri 2014; Grimaldi 2020; Sarakinioti and Philippou 2020), and change the conditionsof academic work in HE institutions (Filippakou and Tapper 2008; Salter and Tapper 2000). In thisrespect, the evaluation regime triggers processes of re-structuring of academic and professionalidentities (Harley 2002; Beck and Young 2005; Middleton 2008) since constant external monitoringresults in the ‘erosion of academic professionalism as a group of highly trained individuals whoexercise control over their own institutions and conditions of work’ (Olssen 2016, 142). In otherwords, by participating in the social practice of evaluation academics are positioned as acquirersof new forms of conduct, knowledge and practice, and ultimately of the desirable ‘official’ identitythrough the criteria encoded in the evaluation rules. This of course does not mean that academicsdo not try to manage their conduct (Ball 2003) since they actually become involved in processes ofaccommodating, challenging or resisting dominant discourses and constantly negotiate their iden-tities (Middleton 2008). However, academics’ views are not within the scope of the present study.
The literature also highlights the dire effects of the evaluation mechanisms on academic researchactivity, as what is valorised in this regime of performativity is short term, outcomes-drivenresearch (Craig, Amernic, and Tourish 2014; Hazelkorn 2015). This orients research activity toan instrumental model of conduct driven by ‘market-forces’ (Ziman 2000, 173) and correlatesresearch output and recognition with publications in specific high-ranking, English-speaking out-lets (Flowerdew 2008; Ferguson 2007; Bocanegra-Valle 2014). Thus, citations and research impact,recognised as the main criteria of the performance of individuals and institutions, accentuate thedominance of the English language in the academia.
Although several contributions have been made on various issues related to language use in HE,ranging from the spread of English taught programmes in HE (Wächter and Maiworm 2014;Coleman 2006; Kedzierski 2016) and language–related problems in student exchanges (Teichler2004) to academics’ publishing in English (Curry and Lillis 2004; Flowerdew 2008; Ferguson,Perez-Llantada, and Plo 2011) the relation between quality assurance processes and the emergingdynamics for adjustments and reforms concerning the use of English in HE contexts is an under-researched area. The linguistic dimension of institutionalised evaluation processes however is partof the multilevel and multidimensional reach of the evaluation regime. Research selectivity is aterm that describes the emerging dynamics of this regime regarding institutional strategies and indi-vidual practices (Warren et al. 2021). Specifically, the main argument is that within the evaluationregime institutional strategies and individual practices tend to be selectively oriented, at the linguisticlevel towards the strengthening of the dominance of English, at the disciplinary level towards forms ofknowledge that may generate tangible research outputs, and at the epistemic level towards sanctionedapproaches leading to the marginalisation of others. In this respect, the present study focuses andgives evidence on how external quality assurance processes conduct academics to exercise practicesdriven by the principles of research selectivity, a phenomenon that as we argue has far reachingimplications for HE governance and the societies which they are supposed to serve.
Quality Assurance in Greek Higher Education
The system of Quality assurance and accreditation (QAA) in the Greek HE system consists of amultilayered structure of agencies located at the national, institutional and departmental level.At the national level and in response to the Bologna Process at an early phase of its development(Asderaki 2009), the Hellenic Quality Assurance and Accreditation Agency (HQA) was first estab-lished in 2005 as an independent quality assurance agency operating under the supervision of theMinistry of Education. Later on, law 4009/2011 upgraded the role of the HQA from an agency ofquality assurance, as established by the legal framework of 2005, to a quality assurance and accred-itation agency of HE institutions and their programmes of study. The HQAwas recently renamed asHellenic Authority for Higher Education (HAHE) (Law 4653/2020). HAHE preserves, as its mainmission, the authority to evaluate and accredit the quality of the operation of HEIs and the
GLOBALISATION, SOCIETIES AND EDUCATION 3
programmes they offer, as established by the legal framework of 2011. In this paper we use the acro-nym HQA since this was valid during the period of our research.
Being a member of the European Association for Quality Assurance in Higher Education(ENQA) of the EHEA, the HQA complies with the general set of guidelines and proceduresincluded in the European Standards and Guidelines for quality (ESG 2015). In response to that,QAA procedures of academic units take place both at the level of departments and institutions.
The QAA system was fully implemented in the Greek HE sector during the last decade followinga slow and fragmented process of implementation resulting from serious opposition from differentacademic, social and political stakeholders (Stamelos and Kavasalakis 2011). The pressures forbroad reforms of all public sectors related to the recent period of the fiscal crisis in Greece accel-erated the institutionalisation of QA mechanisms in Greek HEIs, leading to the completion ofthe first round of evaluation procedures ever implemented in Greece in all HE institutions andall departments by 2016.
Evaluation procedures are structured into two successive phases: an internal self-evaluation car-ried out by the academic staff of each unit and then an external evaluation conducted by a commit-tee of independent academics with experience in evaluating universities, appointed by the HQA.The internal evaluation process concludes with the production of an internal evaluation reportthat conveys the unit’s profile and documents its positive and negative areas of operation as wellas its aspirations. On completion of the internal evaluation process, the external evaluation processbegins. External evaluation committees, based both on quantitative data (e.g., mobility rates) andother evidence of qualitative form provided in the internal evaluation reports or/and gatheredduring the external evaluation committees’ site visits (e.g., information frommeetings with teachingand administrative staff, students and social partners), produce external evaluation reports (EERs)highlighting ‘best practices’ and making recommendations for improvement. The EERs which werecompleted by 2016 were written in English and made public on the websites of both the HQA andthe units under evaluation.
Despite participating in both internal and external cycles of evaluation processes, academics’roles, either as evaluators or those being evaluated, are restricted since the HQA keeps a strong reg-ulative role in all phases of the evaluation. That is, evaluations follow a standardised procedurestrictly monitored by the HQA which determines the stages, criteria and indicators of academicquality, all reflecting the guidelines and principles of the ESG. The requirement is that internaland external evaluation reports should be compiled in specific format following common templatesand pre-set criteria written in the form of focus questions addressing issues of curriculum, researchquality and quantity as well as administration and other services.
It is worth mentioning that, despite differences that are likely to occur in each EER owing to thespecial characteristics of each unit, these reports judge and conduct HEIs on the basis of commonprocedures and formalised criteria while mediating global and European discourses on what isexpected of a high-standard HE institution. Members of the evaluation committees are powerfulagents in institutionalised QA processes who have significant space to exercise their own powerwhen ‘translating’ official policies into evaluation reports. However, and despite differences intheir own understandings, the fact that they are expected to function within a strictly regulated fra-mework using the same criteria as prescribed by the top national agency of QA creates the con-ditions for the emergence of a concrete and rather unidirectional discourse. That is why,although differences do appear in the external evaluation reports, the general tendency is to praiseor advise against the same practices – at least as far as language issues are concerned.
Theoretical background of the analysis of the dynamics and discourses on QA inGreek Higher Education
In order to investigate whether the institutionalised external evaluation processes operate as relaysof discourses that reconfigure the role of language in teaching and research practices in HE and
4 A. VOGOPOULOU ET AL.
direct academics’ work towards certain forms of activities, we draw on Bernstein’s theoretical fra-mework and the concept of the pedagogic device. This theory directs our analysis to the processes ofpedagogisation carried out through QA processes which can be seen as mediating dominant Euro-pean/global discourses and reproducing ‘“regimes of truth” through which people govern them-selves and others’ (Ball 1994, 22).
The pedagogic device, defined as an ensemble of specific distributive, recontextualising and eva-luative rules by which specific forms of knowledge are selected, organised and distributed shapingpower and control relations in the field, in our case in HE (Bernstein 2000, 180), is significant inrealising the multiple agents and discourses contributing to dominant forms of subjects’ regulation.The pedagogic device, which generates ‘a symbolic ruler of consciousness’ (Bernstein 1990, 180),provides a language of description. That is, it provides a set of micro-tools enabling us to identifythe boundaries that demarcate the forms of the knowledge selected, transmitted and evaluated bythe committees of international experts when conducting the external evaluations and reporting theresults in the EERs as well as the forms of social control exercised over the field of HE.
In Bernstein’s theory, the metaphoric use of boundaries refers to power that regulates therelations between categories and to control which is exercised through such relations within cat-egories, corresponding to classification and framing respectively. Classification and framing helpus to account for the preservation or change of each category’s specialisation in the texts of externalevaluation reports. Furthermore, the concept of ‘meaning orientation’, distinguished as ‘introjected’or ‘projected’, explains the generation of inner commitments to intrinsic values (i.e., disciplinaryknowledge) or the reverse, that is the tendency to adopt outwardness primarily shaped in responseto external demands and contingencies of the ‘market’ (Bernstein 2000; Beck and Young 2005; Sar-akinioti, Tsatsaroni, and Stamelos 2011). Any strengthening or weakening of boundaries createsspaces in which new forms of knowledge and power relations are created leading to processes ofidentity (re)formation based on specific categories and related contexts (Bernstein 1990). Addition-ally, Bernstein argues, embedded in the pedagogic device is a regulative discourse operating as a‘moral discourse, which creates order, relations and identity’ (Bernstein 2000, 32). The latter con-cept is at the centre of our present analysis since ‘moral discourses regulate what knowledge isselected and how it is organised to produce selective orientations in meaning’ (Singh, Thomas,and Harris 2013, 5).
The Bernsteinian concepts outlined above form the basis of our analysis of the evaluation pro-cesses from the perspective of the pedagogisation of knowledge. This allows us to identify howknowledge and social control are embedded in the devices of evaluation processes producingnew regulative discourses for HE and academics. In particular, the forms of knowledge andpower that promote transformations linked to high ‘quality’ institutions are identified in the eva-luative rules of the external evaluation reports in order to analyse what is required from HEIsand academics in terms of the nature of their work and the knowledge forms that they shouldvalue. Although pedagogic discourse is often characterised by contradictions, and what is trans-mitted may not be always acquired, it still aims at the production, reproduction and transformationof culture projecting legitimate forms of knowledge and identities, ultimately the maintenance ortransformation of social order.
Our study
Τhe findings presented in this paper are part of a research study conducted during the academicyears 2015–2018 (Vogopoulou 2020) and are derived from the analysis of external evaluationreports (EERs) of fifteen departments from three different HEIs of the country. During that period,the Greek HE system comprised Universities and (Higher) Technological Educational Institutions(TEI) which had full university status but a technical/applied orientation and almost all weremerged with universities in 2018–2019. Both types of institutions were aligned to the same legalframework and equally dependent on the Ministry of Education and Religious Affairs (MoERA).
GLOBALISATION, SOCIETIES AND EDUCATION 5
Therefore, they were subjected to the same system of QA following the common framework ofevaluation designed by the national agency, that is HQA, for all the HEIs of the country.
Since the aim of our study was to capture the characteristics of the discourse generated throughthe QA processes and documented in the EERs irrespectively of the different scientific fields of HEIdepartments, we selected our sample from what may be considered as typical universities anddepartments of the country. Mainly, we considered the status, history and tradition, student prefer-ence as well as orientation, being either practical or theoretical, of Greek HEIs and departments. Atthis point, it should be noted that HE in the Greek system is public according to the Greek Con-stitution. Universities are self-administered entities under public law and state-funded but at thesame time remain closely supervised by the MofERA in all aspects of their function. The numberof new entries in universities follows the principle of numerus clausus which is determined by theMofERA annually. Upper secondary school graduates participate in extremely demanding nation-wide examinations known as Panhellenic exams and, depending on their scores, they compete for aposition in the departments of the country’s HEIs. This system results in a type of national hierar-chy of HEIs by which high-rank institutions considered prestigious are centrally located with alongstanding tradition as well as job and career promising departments that attract and admit stu-dents of high performance. HEIs on the other hand, located on the periphery or/and newly estab-lished are considered second in the ranking.
Taking the above into account our final selection included departments of HEIs located inSouthern Greece which may be considered typical of the Greek HE sector: a high-ranked, reputableuniversity and TEI with a long tradition of admitting students with high performance and a largenumber of academic and administrative staff; a well-established university and TEI in a semi-cen-tral geographic area with a significantly large student body and academic staff; and finally, a morerecently established University with departments located in different capitals of the same region,fewer human resources and smaller student body, clearly positioned lower in the hierarchy ofGreek HEIs. We focused on three departments from each institution (N = 15, in total) and froma wide range of disciplines: engineering, so-called hard sciences, technology, administrativesciences, social sciences, education and humanities. Deriving data from different disciplines servedour aim to trace the elements that appear to be dominant in the discourse articulated in QA pro-cesses in relation to language use and its links to wider policies of reform. Providing a comparativestudy between institutions of different rank or departments of different disciplines was beyond thescope of the present study.
As already mentioned, methodologically we approach EERs as documents mediating Europeanand global discourses on HE quality and therefore are appropriate for analysing shifts in the form ofknowledge and power relations taking place in the Greek context. Our research questions focus firston how language policies/practices are articulated in relation to quality and secondly, on whatpower/knowledge relations become inscribed in the rules of good academic practice and in the pro-jected ‘official identities’. To analyse the texts we coded the EERs with the primary objective of iden-tifying issues related to language use and how they were brought together with HE practices whichwere either recommended or confirmed as ‘best practice’ according to the centrally pre-set criteriaof the evaluation process. The emerging themes were then analysed based on a discourse analysistool developed by operationalising the theoretical view that boundaries as ‘silent punctuations ofsocial space’ demarcate relations between and within categories and are relayed via evaluation pro-cesses (Bernstein 2000, xiii). In particular, we identify boundary changes between and within cat-egories by recording shifts in classification and framing regarding the social interactions as well asthe knowledge resources inscribed in those categories. In terms of issues of social control, we exam-ine the conditions and the extent of autonomy of academics while in relation to the knowledgeresources we look into the strength of boundaries around specific forms of knowledge regardedlegitimate for informing academics on how to keep in line with what is considered as qualitywork in HE.
6 A. VOGOPOULOU ET AL.
English and its links to mobility and international visibility in the EERs
This section presents and discusses the findings of our analysis and is divided into three subsections.As illustrated below the evaluators in the EERs of our sample perceive the adoption of English asbeing embedded within policies and practices aiming at: (1) increased mobility mainly in the frame-work of the Erasmus+ exchange programme; and (2) international visibility through publications ofacademic research in high-rank scientific journals. Each of these themes is presented in the two firstsubsections and in the third they are discussed against the axes of social control and the forms ofknowledge considered legitimate as described in the methodology section.
Limitations to mobility and recommendations to adopt English-taught courses
In the EERs analysed for this study, departments are urged to take advantage of the opportunitiesprovided by the Erasmus framework for student and staff mobility and remove any obstacles ham-pering mobility. Factors that negatively affect the participation in the Erasmus programme (i.e.,language barriers, financial constraints, administrative obstacles) are identified in the EERs as pro-blems that should be fixed. The following extract is an example of the evaluators strongly criticisingthe department’s failure to eliminate practical problems hindering the successful implementation ofthe Erasmus programme:
The committee takes exception to the apparent fact that the department does not pay due attention towardssome practical problems related to the Erasmus exchange program. (External evaluation report, Institution 1,Department B)
The benefits of mobility appear as a recurrent theme in the EERs expressed in terms of two ratio-nales: one emphasising the opportunities for educational/professional advancement and the secondrevolving around the opportunities for personal development through intercultural experiences.The recommendations informed by the former build on the idea of empowering academics and stu-dents in their scholarly/academic development and trajectories. Hence, mobility is linked to thescope and extent of exposure to other institutions abroad in terms of academic scholarship:
To date, participation in the Erasmus programme by the Department faculty and students has been very lim-ited. The external evaluation committee strongly encourages the Department members to utilise this mech-anism for fostering faculty and student development in research and scholarship. (External evaluationreport, Institution 5, department A)
On the other hand, arguments emphasising the benefits of being exposed to different cultures,academic systems and languages point to the social significance of mobility. In this understanding,the ‘pedagogising’ discourse (Singh 2015; Tsatsaroni, Sifakakis, and Sarakinioti 2015) in the EERsdevelops the idea that first-hand experience of different cultures, lifestyles and educational contextsleads to the cultivation of international understanding and language competency along with aca-demic development:
Student mobility is very limited […]. We believe there is significant scope for expanding student exchangesand strongly urge the Department to move in that direction by establishing more agreements and encouragingstudents to participate. The benefits of student exposure to other cultures, language and educational systemsare immeasurable. (External evaluation report, Institution 2, Department A)
Despite differences when arguing in favour of mobility, English courses constitute a key elementin the process of attracting foreign students in the EERs. This reflects the growing tendency ofadopting English in universities in order to increase numbers of foreign student enrolments aspart of a university’s internationalisation strategies even in European countries with powerfullanguages and long traditions in HE (Kerklaan, Moreira, and Boersma 2008; Mosneaga and Ager-gaard 2012; Wächter and Maiworm 2014). The dominance of English in HE has been characterisedas one of the misconceptions of internationalisation (de Wit 2011) and an unintended consequence
GLOBALISATION, SOCIETIES AND EDUCATION 7
of European mobility. Nevertheless, in the following extract the well-established notion that inter-nationalisation of a programme requires the adoption of English is reproduced in a taken forgranted manner:
Mobility is quite unbalanced in terms of inflows and outflows, a situation that seems primarily due to theshortage of the department courses taught in English. Hence, the Committee suggests that the Departmentought to address the issue of the internationalisation of its programme. (External evaluation report, Institution1, Department A)
International visibility, research and the need for English scientific publications
The volume of research and the number of publications in international journals are includedamong the basic indicators of individual academics as well as the whole departments’ performance.The evaluation committees clearly demarcate the range of activities that may potentially lead toincreasing international visibility and provide the rules to be applied in order to fulfil the require-ment of an internationally regognised institution. These rules can be summarised into three cat-egories: developing strategic research plans and collaborations with foreign scholars as well astargeting international publications in English-speaking high-impact journals. We will now presentand comment on each one of the above.
Strategic research plans and research collaborationsStarting with the recommendations to strategically plan research activities so as to yield results wefind that throughout the extracts of the EERs outcome-driven research is prioritised and at the sametime co-related with gaining prestige and status through publications in highly-acknowledged Eng-lish-speaking outlets. Research driven institutions are obviously presented as exemplars of ‘bestpractice’ which underscores the significance attributed to knowledge production in comparisonto knowledge reproduction occurring in teaching. The committees’ statements on the significanceof the research are grounded on the assumption that research typifies prestigious university insti-tutions. This is demonstrated in the following extract which states what a university is worth of:
Research is of paramount importance in a university worth its salt. Visibility of the department can be raisedby decisively intensifying the collaborative research efforts with foreign scholars. The department shouldenhance its research status spectacularly and move towards placing itself by the side of great European insti-tutions. (External Evaluation Report, Institution 1, Department A)
In order to achieve higher yields in research, recommendations on how to improve research out-put are clearly articulated in the EERs. Intensifying research collaborations is mentioned in theextract above as a method of raising the university’s visibility. In other extracts collaborationsare strongly advised since they may result in cultivating ‘a network with external partners: otherUniversities, public and private institutions in order to raise the Department’s visibility and status’(Evaluation Report, Institution 2, Department C).
In addition, academics are oriented towards adopting a new set of social and scientific conven-tions that prioritise the structuring of regulatory frameworks for scientific research, constant moni-toring and the need to produce research outputs readily disseminated. The following extractillustrates well the need for strategic frameworks that valorise a shift from individually-based initiat-ives to comprehensive plans of research work, monitored by the institution on a set of predefinedcriteria. In this understanding, individual achievements need to become part of a collective effort ofa research group which implies interdisciplinary or/and multidisciplinary teams. Furthermore, pro-gress should be monitored and measured:
Whilst the Department has planned some actions to improve its research (especially the creation of a researchlaboratory), they are not included in a comprehensive plan which outlines the priority research lines, estab-lishes research groups and defines clearly indicators to measure the progress of research. The EEC did not
8 A. VOGOPOULOU ET AL.
observe evidence of an infrastructure at institutional or departmental level to develop such a framework.(External Evaluation Report, Institution 4, Department B)
In a similar manner, the external evaluators advance the view that target setting and monitoringbased on performance indicators are likely to improve the department’s research outcomes andraise its visibility:
The EEC did not see a single coherent strategic plan including targets, implementation plans and performanceindicators. The EEC suggests that the Department may wish to consider developing such a document whichcovers all aspects of its activities: teaching, research, external collaboration and national and international visi-bility. (External Evaluation Report, Institution 4, Department B)
International publications in high-ranking journalsStatus and international recognition are connected to the ‘quality’ of publishing outlets. Therequirement, articulated explicitly in the EERs, is that research activities should be directed tothe production of results readily disseminated through high-ranking international outlets whichtranslates into English language publications:
The international journal quality lists usually rank academic journals in four or five groups, with the first twobeing noted as ‘outstanding by international standards’ and ‘internationally recognized’, respectively, or simi-lar. The faculty members of the Department published many times during the last six years in internationallyrecognised journals and this should be continued. (External Evaluation Report, Institution 3, Department C)
What is more, expectations about the language of publications is overtly expressed rather thanimplied in several EERs:
We appreciate that in some contexts writing in Greek may be more appropriate than writing in English. Butthe two should not be seen as mutually exclusive. Internationalisation of research is the way to go particularlywith younger academics, who are building their research profiles. (External Evaluation Report, Institution 1,Department B)
Furthermore, the comparison is the mode through which the EERs mediate dominant discoursesthat co-articulate research activities, quantitative measurement of research output and the value ofinternational publications. In the extract below, the evaluators compare the performance of thedepartment based on what is considered as the ‘international standard’ set by prestigious,highly-ranked institutions – however vague this may be. Metrics are used to convey what isexpected from the department and academic staff while clear directions are given on what shouldbe done to improve their performance. The department’s research profile, it is suggested, is whatprovides visibility, therefore recognition, in the international arena, which can be achieved byrecruiting new members that have ‘strong research potential’:
With some noticeable exceptions past research output is found to be well below the internationally recognisedexpectations for analogous programmes, operating within good foreign universities (an average of 2 high-quality journal papers and several well-regarded conference publications per year per faculty member).This points to the need for the development of research criteria in the evaluations of academic staff pro-motions and the recruitment of new candidates […] with strong research potential and expertise relevantto the main research thrust of the unit, which is yet to be defined. (External Evaluation Report, Institution2, Department C)
Also, academics are meant to be classified according to their performance, based on the criterionof productivity, indicated by the number of publications they have achieved (usually) within a year.In the extract below, the Department is praised for achieving high metrics, which is equated withincreased quality of its scientific work overtime:
It is noteworthy, that the Department is apparently the most productive within the University as indicated bythe number of publications/per faculty member. An increasing number of publications in journals with animpact factor >3 during the last six years indicates furthermore an overall increasing quality of the scientificwork carried out. (External Evaluation Report, Institution 2, Department A)
GLOBALISATION, SOCIETIES AND EDUCATION 9
Furthermore, the positioning of internationally recognised outlets high in the hierarchy isamplified in some evaluation reports which in a direct way advise against publishing in ‘lower qual-ity journals’ or obscure outlets. The extract below is a case in point:
Further, faculty members have also published numerous times in lower quality journals. Given the success inpublishing in internationally recognized journals, we would urge the faculty members to proceed mainly inthis direction and enhance their publication records in such journals in the future. (External EvaluationReport, Institution 2, Department C)
EERs as a pedagogic device and the emerging regulative discourse for HE
We have argued that the external evaluation processes have a pedagogical function since they con-vey the criteria of reaching high levels of quality in teaching, research and other educational/admin-istrative services. We have also argued that the EERs reflect specific understandings of HEIs missionand role that recontextualise global discourses on HE. These understandings are made visible to theevaluated units through reviewing, comparing and recommending ‘best practices’ that make up aquality HE. In terms of what is relayed in the EERs concerning language use the previous sectionillustrated that mobility, research activity and the publication of research outputs in specific jour-nals presuppose the knowledge and use of English as the default language. In this section, ourfindings are discussed through the operationalisation of the theoretical concepts outlined pre-viously, aiming to map the form of social relations and the knowledge resources underpinningthe EERs given the fact that our main finding indicates that the seemingly disparate policies andpractices of language, mobility and research activity become interlinked due to the demand todevelop and project an international profile.
First, through a visible pedagogy (Bernstein 1977) consisting of clearly articulated guidelines inthe EERs, departments are urged to become more ‘active’ in the Erasmus programme, eliminatepractical problems that might hinder its implementation such as the absence of English taughtcourses, and also to establish collaborations via networking with foreign scholars and other insti-tutions. By directing the departments under evaluation to the practices above, the communicationcontext and its social relations are structured within contexts that demonstrate stronger framingconditions aiming to affect the decision-making processes of departments/academics. In particular,it is well-articulated that for the evaluators, increasing mobility is an essential element of HEIswhich requires compliance with external rules, fulfilment of prerequisites, adoption of a specificlanguage, sequencing and pacing of activities resulting from structural rules (i.e., the success ofthe Erasmus pogramme, the rise of inbound mobility through English-medium-instructioncourses) as well as negotiations over differences and interpretations between academics workingjointly.
In terms of the knowledge sources that support the EERs, recommendations regarding the pro-motion of mobility derive legitimation by prioritising either educational/academic goals or per-sonal/social development through exposure to cultures and languages. In the first case,developing academic scholarship is kept strongly insulated from other forms of benefits whilemeaning orientation is introverted since mobility is linked mainly to academic achievement. Inthe second case, the classifications between forms of knowledge that are sanctioned are weakeningsince personal/social development is regarded to be equally vital to academic development whilemeaning orientation projects benefits both in education and social life. In either case, the usefulnessof mobility in obtaining some sort of benefits is emphasised, illustrating an instrumental under-standing of its value while universities are made responsible for providing the opportunity of gain-ing these benefits. Concerning issues connected to linguistic communication, the adoption ofEnglish is seen as a proxy of good levels of performance in mobility.
Regarding the evaluation rules on academic research activity, the EERs clearly outline that theorganisation and development of research should be strategically planned while the disseminationof research output should be done through publications in internationally recognised journals. In
10 A. VOGOPOULOU ET AL.
these recommendations boundaries are redrawn around what constitutes academic knowledge pro-duction by adding a clear outward-oriented instrumental dimension. It also conveys the idea thatknowledge production, detached from any inner commitments (Bernstein 2000) is an issue ofrational planning and absolute determination to win in the race and gain both tangible and intan-gible profits. The recommended shift valorises comprehensive, well-planned and organised activi-ties, guided by continuous institutionalised monitoring. Therefore, while the boundary betweenresearch and other academic activities is strengthened, the generic criteria used for the evaluationof research productions indicate a weakening of the specialisations separating and distinguishingthe different academic disciplines. That is, under the umbrella of research development, weakerclassifications in knowledge call for collaborations, networking and prioritisation of result-orientedresearch. The idea of collaborations further implies interdisciplinarity or/and pluridisciplinarityand moves boundaries within disciplines whilst framing of the communication context strengthensas academics are invited to work in academic programmes that need to be compromised among allcontributors and targeted to raising visibility. Similarly, networking figures high in the suggestionsof the EERs which potentially leads to the weakening of classifications between knowledge formsand the strengthening of framing since agents are integrated and the communication context isshaped in externally formed spaces.
Furthermore, in relation to the linguistic issue academics are expected to target specific outletsand simultaneously draw a line between the powerful standardised English language as produced inthe inner cycle of Anglophones (Kachru 1990) and the non-powerful, less prestigious Greeklanguage. In line with the demands posed by the evaluation processes, specialised knowledgeneeds to be transmitted, communicated through the medium considered in our globalised timesas ‘common-sense’ knowledge bringing thus to an equally important status specialised scientificknowledge with English language skills.
Overall, the rules codified in the EERs are expressions of the principles of performativity andmanagerial accountability, orienting academics towards adopting an ethos which sanctions measur-able outcomes as the only worthwhile purpose of research. In essence, the regulative discourse ofperformativity, which is embodied in the evaluation rules articulated in the EERs, tailors academiceffort and the crucial mission of academic/scientific research to measurement requirements. Pub-lications, collaborations and research activities become a benchmark for improvement. Under theseconditions, the harsh reality of ‘publish or perish’ transforms the English language used by highimpact journals into the most powerful means for raising visibility and demonstrating quality inthe Greek HE sector.
Discussion
In this final section, we present and discuss our conclusions drawing on the findings presented inthe preceding section and in relation to the research questions of our study. The latter aimed toidentify coarticulations of language practices and dominant discourses on the nature, the knowl-edge resources and the power relations of the practices projected as legitimate in a reformed, qualityinstitution and their potential implications for academic work.
In approaching QA processes as a pedagogic device we argue that it was possible to identify thesocial relations and hierarchies that emerge (i.e., who evaluates whom on what criteria), the knowl-edge resources (i.e., forms of knowledge, expertise) as well as elements of discourses of differentorigins (i.e., quality, mobility and research publications, symbolic capital of the English languageas a proxy of international status and visibility) that have been selected, resulting in a new articula-tion relevant to the specific national setting. In particular, discourses articulated in the externalevaluation reports bring the English language issue to the fore as it is linked to practices and policiesincluded in what the literature characterises as internationalisation strategies and tend to includepromotion of mobility, English-taught courses to attract international students as well as targetinginternational visibility through publications. Although the origins and the structures of the above
GLOBALISATION, SOCIETIES AND EDUCATION 11
discourses are different, QA mechanisms promote a discourse which selects the elements of the dis-courses that are relevant for this specific national context (but perhaps in other contexts too) andbring dominant elements into a new articulation. In this way these mechanisms function as a ped-agogic device by (politically and pedagogically) selecting and re-organising elements, giving newmeanings to them. Therefore, our initial question of ‘where does language stand in the discourseon quality?’ has led us to reveal that language, mobility and research performance are unified ina discourse of quality giving a selective understanding of what it means for an institution or an aca-demic to be ‘internationalising’.
Regarding power relations, our analysis reveals that performance monitoring has become morecentralised due to the implementation of external evaluation processes. It appears that through thepeer-reviewed evaluation system, carried out by an independent body operating under the auspicesof the Ministry of Education, the state has strengthened its capacity to ‘steer’ from a distance andintroduce in the HE sector the state’s official recontextualisations of an international quality univer-sity discourse. This denotes the erosion of the traditional autonomy of HEIs over their recontextua-lisation fields (Bernstein 2000) and the enhancement of the state’s regulative power throughcontrolling the QA processes which define standards and evaluate the extent to which each depart-ment has managed to reach them. Positioned in the evaluative field, QA technologies have sur-passed the boundaries distinguishing academics as a highly specialised category of professionalsthat can determine their working conditions, potentially leading to their de-professionalisation(Olssen 2016).
Of significance is the fact that the pedagogic discourse produced through the QA processes ori-ents institutions to the development and maintenance of a dynamic and internationally recognisedresearch sector. This specific gaze cast over academic work and the role of HEIs aligns with neolib-eral understandings which discursively position HEIs as contributors to global economic compe-tition. Furthermore, the restructuring of academic work promoted through QA introducesmarket and neo-managerial frameworks that favour the principles of performativity and the instru-mental use of knowledge. Boundaries around knowledge and its value are apparently (re)drawn,shifting to a utilitarian and market-driven understanding of knowledge and its production whichalienates it from inward orientations. Scholars such as Muller (2009), Beck (2002, 2009) andBeck and Young (2005), inspired by Bernstein’s (2000) approach to the sociology of knowledge,have stressed the changing nature of knowledge and of the relation between knowledge andknowers and users of knowledge. Linked to this change is the idea that knowledge should createadvantage and profit (Bernstein 2000) as well as the general tendency of capitalising knowledge pro-duction (Shattock 2009).
In this context, the discourse of the EERs, introduces in the Greek HE sector a set of strategiesthat can potentially lead to high levels of performance in academic research, measured by citationsin internationally recognised outlets, the number of international collaborations and the extent ofnetworking and mobility. The regulative dimension of this discourse clearly embodies the principleof performativity (Ball 2008, 2012). Also, it becomes apparent that a strong tendency of boundaryblurring exists between traditionally well-demarcated fields of intellect and fields of practice as wellas strengthening of the control over the communication context, which clearly advocates the use ofEnglish.
In addition, the emphasis placed on the production and dissemination of research outputsthrough specific kinds of publications transforms communication in the academia and particularlyacademic writing into a high-stakes activity linked to the institutions’ and individuals’ aspirations tobecome visible and recognised at the international level. Institutions and academics are forced torethink their language policies and practices since the benefits of using English in academic researchdissemination and mobility are strongly emphasised.
Reaching beyond national borders is entangled with the notion of ‘openness’ to a rather unrest-ricted market, requiring evidence-based performance and English as the default language. Theselogics sustain discourses of the merits of monolingualism as a cost-effective means of
12 A. VOGOPOULOU ET AL.
communication in the academia and simultaneously of English language dominance since Englishfigures as the only choice. Thus, scientific knowledge and expertise demand English language com-petency, demonstrating another significant displacement of boundaries to accommodate the coex-istence of the verticality of scientific knowledge (Muller 2006) with a horizontal set of skills, i.e.,English language use. The latter may contribute significantly to a competitive status in the globalscene, an assumption which is taken as a given by the weakening of classifications between disci-pline-based knowledge and hybrid formations of agents/agencies cooperating for market-likerealisations and recognition of knowledge production.
Furthermore, legitimate criteria of HE performance project a certain official market-orientedidentity justified by reference to its potential contribution to a competitive HE field, demarcatedby high and low-performing institutions. This identity type ‘advocates’ a knowledge-base that pur-sues practical realisations and both material and status gains (Warren et al. 2021), marking thus ashift towards objectives and values of the economic field. Even when slippages in policy discourseretreat to academic scholarship and research driven by inner commitments, the value of English isstill reconfigured as a significant form of linguistic capital in the English dominated academic com-munity (Ammon 2001) with far reaching implications as outlined by our analysis. One of the mostimportant consequences however, revealed by our analysis, is that the imperative to use Englisharticulated in the discourse on quality and internationalisation functions as a powerful regulativeforce – essential aspect of performativity – which can have an effect on individuals’ behaviourand linguistic trajectories (Martín-Rojo 2018). It also entails the demand for a continuous effortto invest in developing language capabilities and resources since, contrary to the dominant dis-course on the internationality of English, it is not easily and universally accessible by all – a situationthat surely aggravates already existing asymmetries.
Disclosure statement
No potential conflict of interest was reported by the author(s).
ORCID
Areti Vogopoulou http://orcid.org/0000-0001-6690-1969
References
Ammon, U., ed. 2001. The Dominance of English as a Language of Science. Effects on Other Languages and LanguageCommunities. Berlin: Mouton de Gruyter.
Asderaki, F. 2009. “The Impact of the Bologna Process on the Development of the Greek Quality Assurance System.”Quality in Higher Education 15 (2): 105–122.
Ball, S. J. 1994. Education Reform. A Critical and Post-Structural Approach. Buckingham: Open University Press.Ball, S. J. 2003. “The Teacher’s Soul and the Terrors of Performativity.” Journal of Educational Policy 18 (2): 215–228.Ball, S. J. 2008. The Education Debate. Bristol: The Policy Press.Ball, S. J. 2012. “Performativity, Commodification and Commitment: An I-Spy Guide to the Neoliberal University.”
British Journal of Educational Studies 60 (1): 17–28. doi:10.1080/00071005.2011.650940.Beck, J. 2002. “The Sacred and the Profane in Recent Struggles to Promote Official Pedagogic Identities.” British
Journal of Sociology of Education 23 (4): 617–626.Beck, J. 2009. “Appropriating Professionalism: Restructuring the Official Knowledge Base of England’s ‘Modernised’
Teaching Profession.” British Journal of Education 30 (1): 3–14.Beck, J., and M. F. D. Young. 2005. “The Assault on the Professions and the Restructuring of Academic and
Professional Identities: A Bernsteinian Analysis.” British Journal of Sociology of Education 26 (2): 183–197.Bernstein, B. 1977. Class, Codes and Control, vol. 3. Towards a Theory of Educational Transmissions. 2nd ed. London:
Routledge & Kegan Paul.Bernstein, B. 1990. Class, Codes and Control, Vol. 4: The Structuring of Pedagogic Discourse. London: Routledge.Bernstein, B. 2000. Pedagogy, Symbolic Control and Identity: Theory, Research, Critique. Revised Edition. New York,
NY: Rowman & Littlefield.
GLOBALISATION, SOCIETIES AND EDUCATION 13
Bocanegra-Valle, A. 2014. “‘English is my Default Language’: Voices from LSP Scholars Publishing in a MultilingualJournal.” Journal of English for Academic Purposes 13: 65–77.
Brennan, J., and T. Shah. 2000. “Quality Assessment and Institutional Change: Experiences from 14 Countries.”Higher Education 40: 331–349.
Coleman, J. A. 2006. “English-medium Teaching in European Higher Education.” Language Teaching 39 (1): 1–14.Craig, R., J. Amernic, and D. Tourish. 2014. “Perverse Audit Culture and Accountability of the Modern Public
University.” Financial Accountability & Management 30 (1): 1–24.Curry, M. J., and T. Lillis. 2004. “Multilingual Scholars and the Imperative to Publish in English: Negotiating
Interests, Demands, and Rewards.” TESOL Quarterly 38: 666–688.de Wit, H. 2011. “Internationalisation of Higher Education: Nine Misconceptions.” International Higher Education
64: 6–7.Elton, L. 2000. “The UK Research Assessment Exercise: Unintended Consequences.” Higher Education Quarterly 54
(3): 274–283.Ferguson, G. 2007. “The Global Spread of English, Scientific Communication and ESP: Questions of Equity, Access
and Domain Loss.” Ibιrica 13: 7–38.Ferguson, G., C. Perez-Llantada, and R. Plo. 2011. “English as an International Language of Scientific Publication: A
Study of Attitudes.” World Englishes 30: 41–59.Filippakou, O., and T. Tapper. 2008. “Quality Assurance and Quality Enhancement in Higher Education: Contested
Territories?” Higher Education Quarterly 62 (1-2): 84–100.Flowerdew, J. 2008. “Scholarly Writers who use English as an Additional Language: What Can Goffman’s ‘‘Stigma’’
Tell us?” Journal of English for Academic Purposes 7: 77–86.Gornitzka, A., and B. Stensaker. 2014. “The Dynamics of European Regulatory Regimes in Higher Education —
Challenged Prerogatives and Evolutionary Change.” Policy and Society 33 (3): 177–188. doi:10.1016/j.polsoc.2014.08.002.
Gounari, P. 2012. “Neoliberalizing Higher Education in Greece: New Laws, Old Free-Market Tricks.” Power &Education 4 (3): 277–288.
Gouvias, D. 2012. “The Post-Modern Rhetoric of Recent Reforms in Greek Higher Education.” Journal for CriticalEducation Policy Studies 10 (2): 282–313.
Grek, S. 2013. “Expert Moves: International Comparative Testing and the Rise of Expertocracy.” Journal of EducationPolicy 28 (5): 659–709.
Grimaldi, E. 2020. An Archaeology of Educational Evaluation: Epistemological Spaces and Political Paradoxes. Oxon:Routledge.
Harley, S. 2002. “The Impact of Research Selectivity on Academic Work and Identity in UK Universities.” Studies inHigher Education 27 (2): 187–205.
Hazelkorn, E. 2015. Rankings and the Reshaping of Higher Education: The Battle for World-Class Excellence. London:Palgrave.
Kachru, B. 1990. “World Englishes and Applied Linguistics.” World Englishes 9: 3–20.Kedzierski, M. 2016. “English as a Medium of Instruction in East Asia’s Higher Education Sector: A Critical Realist
Cultural Political Economy Analysis of Underlying Logics.” Comparative Education 52 (3): 375–391.Kerklaan, V., G. Moreira, and K. Boersma. 2008. “The Role of Language in the Internationalisation of Higher
Education: An Example from Portugal.” European Journal of Education 43 (2): 241–255.Landri, P. 2014. “Governing by Standards; the Fabrication of Austerity in the Italian Education System.” Education
Inquiry 5 (1): 25–41. doi:10.3402/edui.v5.24057.Marginson, S. 2007. “Global Position and Position Taking: The Case of Australia.” Journal of Studies in International
Education 11: 5–31.Marginson, S. 2016. Higher Education and the Common Good. Melbourne: Melbourne University Press.Martín-Rojo, L. 2018. “Neoliberalism and Linguistic Governmentality.” In The Oxford Handbook of Language Policy
and Planning, edited by J. W. Tollefson and M. Pérez- Milans, 544–567. Oxford: Oxford University Press.Middleton, S. 2008. “Research Assessment as a Pedagogical Device: Bernstein, Professional Identity and Education in
New Zealand.” British Journal of Sociology of Education 29 (2): 125–136.Mosneaga, A., and J. Agergaard. 2012. “Agents of Internationalization? Danish Universities’ Practices for Attracting
International Students.” Globalisation, Societies and Education 10 (4): 519–538.Muller, J. 2006. “On the Shoulders of Giants: Verticality of Knowledge and the School Curriculum.” In Knowledge,
Power and Educational Reform, edited by R. Moore, M. Arnot, J. Beck and H. Daniels, 11–27. London: Routledge.Muller, J. 2009. “Forms of Knowledge and Curriculum Coherence.” Journal of Education and Work 22 (3): 205–226.Naidoo, R. 2016. “The Competition Fetish in Higher Education: Varieties, Animators and Consequences.” British
Journal of the Sociology of Education 37 (1): 1–10.Olssen, M. 2016. “Neoliberal Competition in Higher Education Today: Research, Accountability and Impact.” British
Journal of Sociology of Education 37 (1): 129–148.Ozga, J. 2009. “Governing Education Through Data in England: From Regulation to Self-Evaluation.” Journal of
Education Policy 24 (2): 149–162.
14 A. VOGOPOULOU ET AL.
Ozga, J. 2012. “Governing Knowledge: Data, Inspection and Education Policy in Europe.” Globalisation, Societies andEducation 10 (4): 439–455.
Ozga, J., P. Dahler-Larsen, C. Segerholm, and H. Simola, eds. 2011. Fabricating Quality in Education: Data andGovernance in Europe. Abingdon, U.K: Routledge.
Salter, B., and T. Tapper. 2000. “The Politics of Governance in Higher Education.” Political Studies 48 (1): 66–87.Sarakinioti, A., and S. Philippou. 2020. “European Discourse on Higher Education Quality Assurance and
Accreditation: Recontextualizations in Greece and Cyprus at Times of ‘Crisis’.” European Education 52 (2):132–145.
Sarakinioti, A., A. Tsatsaroni, and G. Stamelos. 2011. “Changing Knowledge in Higher Education.” In Knowledge andIdentity: Concepts and Applications in Bernstein’s Sociology, edited by G. Ivinson, G. B. Davies and J. Fitz, 69–89.London: Routledge.
Shattock, M., ed. 2009. Entrepreneurialism in Universities and the Knowledge Economy: Diversification andOrganizational Change in European Higher Education. Maidenhead, UK: Society for Research in to HigherEducation & Open University Press.
Shore, C., and S. Wright. 2015. “Audit Culture Revisited. Rankings, Ratings, and the Reassembling of Society.”Current Anthropology 56 (3): 421–444.
Singh, P. 2015. “Performativity and Pedagogising Knowledge: Globalising Educational Policy Formation,Dissemination, and Enactment.” Journal of Education Policy 30 (3): 363–384.
Singh, P., S. Thomas, and J. Harris. 2013. “Recontextualising Policy Discourses: A Bernsteinian Perspective on PolicyInterpretation, Translation, Enactment.” Journal of Education Policy 28 (4): 465–480.
Slaughter, S., and G. Rhoades. 2004. Academic Capitalism and the New Economy: Markets, State, and HigherEducation. Baltimore: Johns Hopkins University Press.
Stamelos, G., and A. Kavasalakis. 2011. “The Public Debate on a Quality Assurance System for Greek Universities.”Quality in Higher Education 17 (3): 353–368.
Standards and Guidelines for Quality Assurance in the European Higher Education Area (ESG). 2015. Brussels,Belgium.
Teichler, U. 2004. “The Changing Debate on Internationalization of Higher Education.” Higher Education 48: 5–26.Tsatsaroni, A., P. Sifakakis, and A. Sarakinioti. 2015. “Transformations in the Field of Symbolic Control and Their
Implications for the Greek Educational Administration.” European Educational Research Journal 14 (6): 508–530.Vogopoulou, A. 2020. “Internationalisation, Language Policies/practices and the Hegemony of English: A Study of
New Forms of Governance, Knowledge and Identities in Greek Higher Education Institutions.” PhD diss.,University of the Peloponnese.
Wächter, B., and F. Maiworm, eds. 2014. English-Taught Programmes in European Higher Education. The State ofPlay in 2014. ACA Papers on International Corporation IN Education. Bonn: Lemmens Medien GmbH.
Warren, S., M. Starnawski, A. Tsatsaroni, A. Vogopoulou, and P. Zgaga. 2021. “How Does Research PerformativityImpact on the Non-Core Regions of Europe? The Case for a New Research Agenda.” Higher Education 81: 607–622.
Ziman, J. 2000. Real Science. Cambridge: Cambridge University Press.
GLOBALISATION, SOCIETIES AND EDUCATION 15
Abstract
Introduction
Quality Assurance and HE transformations: where do language policies and practices stand?
Quality Assurance in Greek Higher Education
Theoretical background of the analysis of the dynamics and discourses on QA in Greek Higher Education
Our study
English and its links to mobility and international visibility in the EERs
Limitations to mobility and recommendations to adopt English-taught courses
International visibility, research and the need for English scientific publications
Strategic research plans and research collaborations
International publications in high-ranking journals
EERs as a pedagogic device and the emerging regulative discourse for HE
Discussion
Disclosure statement
ORCID
References
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles false /AutoRotatePages /PageByPage /Binding /Left /CalGrayProfile () /CalRGBProfile (Adobe RGB 50199851) /CalCMYKProfile (U.S. Web Coated 50SWOP51 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.1000 /ColorConversionStrategy /sRGB /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments false /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo false /PreserveFlatness true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings false /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict <> /ColorImageDict <> /JPEG2000ColorACSImageDict <> /JPEG2000ColorImageDict <> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict <> /GrayImageDict <> /JPEG2000GrayACSImageDict <> /JPEG2000GrayImageDict <> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Average /MonoImageResolution 300 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict <> /AllowPSXObjects true /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /Description <>>> setdistillerparams<> setpagedevice
Greek Life, Academics, and Earnings William E. Even, Austin C. Smith
Journal of Human Resources, Volume 57, Number 3, May 2022, pp.998-1032 (Article)
Published by University of Wisconsin Press
For additional information about this article
[ Access provided at 15 Nov 2022 18:36 GMT from Liberty University ]
https://muse.jhu.edu/article/853093
Greek Life, Academics,and Earnings
William E. EvenAustin C. Smith
ABSTRACT
Using records from a large public university, we examine the impact of Greeksocial organizations on academic performance and labor market outcomes.To isolate the causal effect of Greek life, we exploit a university policyprohibiting students from joining a Greek organization during their firstsemester and requiring a minimum grade point average (GPA) for eligibility.Regression discontinuity and panel methods reveal that Greek affiliationreduces student grades by 0.1–0.3 standard deviations. Greek effects arelargest during the semester of pledging, semesters of increased socialactivities, and for males. We find no evidence that Greek affiliation improveslabor market outcomes for marginally eligible students.
I. Introduction
Social fraternities and sororities play a central role in the lives of manyU.S. college students. According to the 2009 College Senior Survey, 17.5 percent ofgraduating college seniors report having joined aGreek organization during their college
William Even is a professor emeritus of economics at Miami University and a research fellow at IZA.Austin Smith is an associate professor of economics at Miami University and a research affiliate at IZA(smitha83@miamioh.edu). The authors thank Deborah Fletcher, Jason Lindo, Greg Madonia, GregNiemesh, Nicholas Sanders, Glen Waddell, and three anonymous referees, as well as seminar participantsat the University of Maine, Miami University, and the 2019 ASSA annual meetings for helpful comments.All remaining errors are those of the authors. The authors have no relevant or material financial intereststhat relate to the research described in this paper. This study relies on transcript-level student data thatcontain sensitive information on student outcomes. Anyone interested in acquiring the data must contact theMiami University Office of Institutional Research. Additional replication materials are provided in theOnline Appendix of Replication Materials.[Submitted October 2018; accepted February 2020]; doi:10.3368/jhr.57.3.1018-9814R3JEL Classification: I23, I21, and I26ISSN 0022-166X E-ISSN 1548-8004ª 2022 by the Board of Regents of the University of Wisconsin System
Supplementary materials are freely available online at: http://uwpress.wisc.edu/journals/journals/jhr-supplementary.htmlWilliam Even https://orcid.org/0000-0002-4116-2044Austin Smith https://orcid.org/0000-0002-7425-6645
THE JOURNA L OF HUMAN RE SOURCE S � 5 7 � 3
careers (Franke et al. 2010).1 The national organizations representing fraternities andsororities outline several laudable goals for its membership, including the intellectualdevelopment of members, the cultivation of leadership skills, and citizenship throughservice and philanthropic efforts. Evidence of success on these goals includes the factthat, in 2014 alone, undergraduate fraternity members provided 3.8 million hours ofcommunity service and raised $20.3 million for philanthropic causes.2 Although Greekorganizations pursue positive goals, their role in campus life is frequently called intoquestion when there are high profile cases of binge drinking, hazing, and risky behaviorthat has occasionally resulted in death.3 While low-frequency catastrophic events likemortality draw national headlines, less is known about the impact of fraternities andsororities on important dimensions of college life—academic performance and post-graduate career outcomes.The problematic behaviors sometimes associated with Greek organizations suggest
that their effect on student performance could be substantial. For example, DeSimone(2009) finds that Greek members consume more alcohol than non-Greeks, even aftercontrolling for prior drinking behavior. An increasing body of evidence demonstrates acausal link between alcohol consumption and reduced academic performance of collegestudents (Carrell, Hoekstra, andWest 2011; Lindo, Swensen, andWaddell 2013; Ha andSmith 2019). Additionally, social events associated with binge drinking (for example,major college sports games) reduce academic performance (Lindo, Swensen, andWaddell 2012). Countering potential downsides, Greek organizations often incentivizeacademic performance by awarding scholarships and providing academic support totheir members (for example, tutoring).Estimating the causal impact of Greek affiliation on academics has proven chal-
lenging. While Greek organizations are quick to point out that their members havegrade point averages (GPAs) above their institution specific averages, this simple sta-tistic does not demonstrate that membership improves academic performance. A keyfactor driving these differences is thatmanyGreek organizations imposeminimumGPArequirements for joining a chapter.4 Additionally, previous studies document significantdifferences in background characteristics betweenGreeks and non-Greeks. For instance,Sacerdote (2001) finds strong evidence of selection into Greek organizations at Dart-mouth based on above average alcohol consumption during high school. Greeks alsotend to come frommore affluent families and aremore likely to bewhite (DeDonato andThomas 2017; Routon andWalker 2014). These observable differences raise the concern
1. Throughout the paper, “Greek” refers to social fraternities and sororities, not Greek professional organi-zations or honor societies.2. These facts are available for fraternities from the North-American Interfraternity Conference (nicindy.org)and for sororities from the National Panhellenic Conference (https://npcwomen.dynamic.omegafi.com/)(accessed October 14, 2021).3. During 2017, pledges died at Pennsylvania State University, Louisiana State University, Florida StateUniversity, and Texas State University.4. In fact, the North-American Interfraternity Conference (NIC), the governing body for many fraternities inthe United States, requires that potential newmembers with established collegiate GPAs have at least a 2.5GPAto join. The NIC also requires that each chapter maintain a GPA that is at or above 2.7. If the male GPA at theinstitution is below 2.7, the chapter must maintain a GPA that equals or exceeds the male GPA at the institution.The National Panhellenic Conference (NPC) that serves as the governing body for most sororities requires thateach chapter meet or exceed the campus all-women’s GPA.
Even and Smith 999
that the two groups may also differ in unobservable ways (for example, ambition,motivation, social skills, etc.). Given these systematic differences between Greek andnon-Greek students, simple comparisons of academic outcomes for the two groupsshould not be interpreted as causal.The existing literature on Greek affiliation’s impact on academic performance has
attempted to sort out the causal effect but remains conflicted. A pair of studies using alarge set of individual characteristics in a propensity score matching framework findnegligible negative and some positive effects of Greek affiliation on academics (RoutonandWalker 2014;Walker,Martin, andHussey 2015). In contrast, DeDonato and Thomas(2017) use an individual fixed-effects approach and find substantial negative effects ofGreek affiliation using the same data as Walker, Martin, and Hussey (2015).5 Theseconflicting results highlight the importance of addressing unobservable differencesbetween Greek and non-Greek students.Characteristics of our data and the institutional setting allow us to contribute to the
existing literature in several ways. The public university we consider is broadly rep-resentative of four-year public institutions and has a large Greek community, making itan ideal setting to estimate the impact of Greek life.6 The granularity of our data allowsus to track a student’s Greek status by semester, from unaffiliated, to pledge, to activemember, and when applicable, to inactive due to chapter suspension or individualmembers departing. This allows us to document the impact of Greek life on academicsacross the Greek life cycle. By linking our transcript data with survey data on labormarket outcomes for six cohorts of graduates, we are able to investigate whether effectspersist after graduation. Finally, our results are supported by two separate identificationstrategies that exploit university-specific eligibility requirements for joining a Greekorganization.Our first identification strategy is a regression discontinuity (RD) approach that relies
on an institutional GPA requirement for joining Greek organizations. The university inour study requires students to attain a cumulative GPA of 2.5 prior to rushing. Our RDresearch design is motivated by the intuition that students with GPAs just below the 2.5eligibility threshold entering their second semester provide a good counterfactual forstudents who have aGPA just above the threshold. Aswewill show, students who barelyexceed the eligibility requirement are much more likely to join a Greek organizationthan those who barely fall short. Further, students with GPAs close to the eligibilitythreshold are very similar in terms of their observable characteristics. This supports theassertion that the subsequent difference in academic performance of students just aboveand just below the threshold captures a causal impact of Greekmembership on academicachievement.Our second identification strategy builds onDeDonato and Thomas (2017) and relies
on the longitudinal nature of our data.7 The university in our study requires students to
5. Using retrospective survey data and an instrumental variables approach, Mara, Davis, and Schmidt (2018)find that students induced to join fraternities based on university housing policies earn lower grades.6. Previous studies have analyzed data from an elite private university (Duke) (Walker, Martin, and Hussey2015; De Donato and Thomas 2017) and an unnamed formerly all-male private university (Mara, Davis, andSchmidt 2018).7. De Donato and Thomas (2017) use an individual fixed-effects model with data from two cohorts at DukeUniversity. One limitation of their data is that it only provides information on Greek status at one point during astudent’s academic career.
1000 The Journal of Human Resources
complete at least one semester as a full-time student prior to rushing a Greek organi-zation and to earn a GPA of 2.5 or better. In our fixed-effects models, we identify theimpact of Greek affiliation on academic performance by comparing students’ post-affiliation performance to their own pre-affiliation performance. This approach differ-ences out any unobservable characteristics about the student that are time invariant.Additionally, we control for cumulative college experience to address the possibilitythat students may systematically improve or shirk as they approach graduation.Results from both specifications indicate that, following affiliation, students’ grades
fall significantly below the level expected based on their prior performance and accu-mulated experience. Results from the panel model reveal that the detrimental impact islargest during the semester of pledging and for males. Increased course withdrawalsand selection into easier courses during the pledging semester suggest that focusingsolely on course grades understates the detrimental impact of affiliation on academics.Speaking to mechanisms, declines in grades for active members are significantly largerin spring semesters, which coincide with additional social events but fewer reportedservice hours per member.Despite reductions in academic performance, it is possible that other aspects of Greek
life, such as the fostering of leadership skills and professional networks, could servegraduates well in the labor market. To complement our analysis on academic perfor-mance, we use secondary data from a post-graduation survey to estimate the impact ofGreek affiliation on employment status and starting salaries. While a Greek salarypremium is evident in the cross-section, we present evidence that cross-sectional esti-mates are positively biased. Using the RD approach, we find no evidence of a salarypremium for Greek students with marginal academic credentials upon affiliating with afraternity or sorority, and in several specifications we can rule out the large positiveeffects found in prior studies.The rest of the paper is organized as follows. Section II describes the university, its
Greek life policies, and our transcript-level data. In Section III, we detail our RDapproach and present the associated results. In Section IV, we describe the individualfixed-effects model, present our results of impacts across the Greek life cycle, anddiscuss heterogeneity in the Greek effects. Section V presents estimates of the impact ofGreek life on long-run educational outcomes, and Section VI presents results for labormarket outcomes. Section VII discusses the implications of our findings.
II. Background and Data
Our data come from a large public university located in the midwesternUnited States. Based on the 25th and 75th percentile of ACT scores of incomingstudents, this university is similar to four-year public institutions such as IndianaUniversity–Bloomington, University of California–Santa Barbara, and the Universityof Vermont. Importantly for our study, the university has a richGreek heritage, andmorethan 38 percent of the students in our sample affiliate with a Greek organization at somepoint in their undergraduate careers, about twice the national average of 17.5 percent.While Greek affiliation is more common at this university, the Greek life cycle is
typical of that at many institutions. The vast majority of students (99 percent) start their
Even and Smith 1001
college careers during the fall semester. Students are prohibited from affiliating with afraternity or sorority during their first semester, as they adjust to campus life. Studentswho complete a minimum of 12 credit hours are eligible for Greek recruitment duringtheir second semester.8 Beginning in the fall of 2009, the university added the re-quirement that a student must have a GPA of at least 2.5 to be eligible for recruitment.9
Online Appendix Table A1 outlines the traditional timeline for a Greek student anddefines key terms used throughout the paper. Recruitment (sometimes known as “rush”)is the process by which Greek organizations introduce themselves to prospective mem-bers by hosting a series of open events. After the brief recruitment period lasting ap-proximately one week, existing members vote on whether to extend a bid to prospectivemembers, who in turn decide whether to accept the bid. Matched students becomepledges, undergoing new member education that lasts for much of the semester. Anec-dotal evidence suggests pledges are sometimes subjected to hazing and that the “socialactivities” during the pledge semester may divert attention from a student’s studies. Uponsuccessfully completing the new-member education process, pledges become activemembers. Barring any voluntary or forced disaffiliation (for example, suspension), stu-dents remain active members for the remainder of their undergraduate careers.Our primary data track academic performance at the student–course level from fall
2007 through spring 2018. We focus on performance during the traditional fall andspring semesters.10 Because our identification strategies exploit a university policyprohibiting students from joining a Greek organization during their first semester, wedrop students missing first-semester records (that is, those enrolling prior to fall 2007).Finally, we restrict our analysis to domestic students because the majority of interna-tional students are missing ACT scores, and the international students in our samplealmost never affiliate. After dropping incomplete records, the resulting sample contains1,065,129 student–course observations for 34,505 students.Table 1 provides summary statistics stratified by gender and Greek status. Uncon-
ditionally, bothmale and female Greeks perform better in the classroom than non-Greekcounterparts. This holds whether the comparison is based on raw GPA or a GPA basedon grades normalized by course–section.11 For the remainder of the paper, a “course”refers to a specific section of a course in a given semester unless specifically notedotherwise.While the above summary statistics reveal that Greek membership is associated with
slightly higher grades, the simple statistics donot imply a causal relationship. The primary
8. Transfer students are allowed to be part of the Greek recruitment process in the first semester they enroll ifthey have completed at least 12 hours with at least a 2.5 GPA at their former institutions. If transfer studentsweremembers of Greek organizations at their prior institutions, they could affiliatewith the same chapters upontransfer and bypass the recruiting process.9. While the university had no GPA requirement for pledging prior to 2009, individual organizations wereallowed to impose a GPA requirement. We do not have information on chapter-specific GPA requirements.10. We exclude grades from condensed terms (summer and January terms) because these terms are substan-tially different in structure, offerings, class size, and teacher and student characteristics.11. The normalized grade is calculated as the difference between the student’s grade and the average grade forthe course–section divided by the standard deviation of grades in the course–section. The grade scale is thestandard four-point scale, with both anA andA+ generating four points and a failing grade yielding zero points.A given course (for example, Principles of Microeconomics) can have many sections taught at different timesby different instructors. Our normalization is done by course–section.
1002 The Journal of Human Resources
empirical challenge to estimating the causal impact of Greek affiliation is that mem-bership is endogenous because there are GPA requirements for admission and studentschoose whether to affiliate. As shown in Table 1, relative to non-Greeks, students whoaffiliate are less likely to be a domestic minority, a state resident, a student athlete, orto have medium or high financial need. Differences in these characteristics are likelyto exert bias on estimates that compare the academic performance of Greeks to non-Greeks. Moreover, it is conceivable that there is selection on unobservables (for exam-ple, work ethic, personality traits, etc.) that would cause Greek status to be endogenousin a grade regression that controlled for observable characteristics. To overcome theendogeneity in Greek status, we rely on two different identification strategies: regressiondiscontinuity and panel data approaches.
III. Regression Discontinuity Approach
In order to be eligible to rush a Greek organization, a student must havecompleted at least one full-time semester with a cumulative GPA of at least 2.5. The vastmajority of students (88 percent of women and 80 percent of men) that affiliate duringtheir academic careers pledge during the spring semester of their first year. The 2.5cumulative GPA requirement is imposed at the time of affiliation, and if a student’scumulative GPA drops below 2.5 after entry, there is no consequence from the university.
Table 1Summary Statistics
FemaleNon-Greek
FemaleGreek Difference
MaleNon-Greek
MaleGreek Difference
Student characteristicsGPA 3.29 3.35 0.06 3.03 3.09 0.06Normalized course grade 0.15 0.21 0.06 -0.05 0.01 0.06ACT composite score 26.87 26.79 -0.08 27.65 27.47 -0.18Credit hours per semester 15.09 14.97 -0.11 14.62 14.45 -0.16State resident 0.73 0.55 -0.18 0.70 0.54 -0.16High or mediumfinancial need
0.35 0.17 -0.18 0.28 0.16 -0.12
Domestic minority 0.16 0.09 -0.07 0.14 0.09 -0.05Athlete 0.06 0.01 -0.05 0.05 0.01 -0.04Greek-only characteristicsPledge second semester 0.88 0.80Ever suspended 0.05 0.11Ever departed 0.11 0.10Observations 9,888 8,701 9,681 6,235
Notes: Sample data consist of domestic undergraduate students from fall 2007 through spring 2018 who initiallyenrolled in fall 2007 or later.
Even and Smith 1003
Thus, we focus on cumulative GPA upon entering a student’s second semester as the keyeligibility criteria.12 For this RD specification, we restrict the sample to students enrollingin fall 2009 and beyond, which is the time frame inwhich the university GPA requirementwas in effect.To identify a plausible comparison group for Greeks, we take advantage of the fact
that eligibility to join a Greek organization is suddenly activated with a first-semesterGPA of 2.5. Students with GPAs just above 2.5 are free to affiliate, while students withGPAs just below 2.5 are not. As wewill show, the marginally eligible students are muchmore likely to join a Greek organization, but the two groups are otherwise quite similarin their academic performance through their first college semester and in their ob-servable characteristics. Comparing subsequent academic outcomes of the marginallyeligible to the marginally ineligible provides an unbiased estimate of the impact ofGreek affiliation for students with qualifying GPAs near the threshold.The first empirical question is whether the university’s eligibility policy binds. Does
attaining a first-semester GPA of 2.5 lead to a discontinuity in the probability of affil-iating with a Greek organization? Figure 1 illustrates the discontinuity in second-semester affiliation. It shows the probability of affiliating during second semester (theoutcome) on the vertical access and the number of GPA points above or below the 2.5GPA threshold (the running variable) on the horizontal access. Each dot represents theshare of students within that GPA categorywho affiliate. EachGPA category has awidthof 0.05 GPA points. For instance, the dot just to the right of the cutoff represents thepercentage of all students with a GPA between 2.5 and 2.549 who pledge during secondsemester. All subsequent RD figures take the same form.Visually, it is apparent that being just above the eligibility threshold leads to ap-
proximately a 25 percentage point increase in the probability of pledging during secondsemester.13 If other factors that might influence GPA do not change discretely at the 2.5threshold, any discrete change in subsequent academic performance can plausibly beattributed to Greek affiliation.To estimate the magnitude of the discrete jump in Greek affiliation and any corre-
sponding discontinuity in subsequent academic performance, we estimate the followingreduced-form RD model:
(1) yijt =b0 +b1Eligiblei + f (CenterCumGPAi)
+ f (CenterCumGPAi)�Eligiblei + ht + eijtwhere yijt is an outcome for student i in course j in semester t, such as normalizedcourse grade.14 Eligiblei is an indicator equal to one if student i attained a first-semester
12. The university introduced a winter session in January 2014. If students enroll in the winter term after theirfall semester, the GPA at the end of the winter term is used as the qualifying GPA. For ease of exposition,students’ GPA at the end of their first winter term is referred to as the first-semester GPA.13. There are a few students without qualifying GPAs (those to the left of the threshold) who nonetheless appearto affiliate. These students may be transfer students who are eligible to affiliate based on GPAs at their originalinstitutions or could reflect downward grade changes that occur after eligibility is determined (for example, anincomplete grade in first semester would not impact eligibility. Once an incomplete course is completed, thegrade is retroactively replaced in official records to reflect the ultimate course grade that is earned. Courses thatare never completed are eventually replaced with an F, which could make a cumulative GPA that was eligible atthe time of pledging later appear ineligible on the basis of the student’s cumulative GPA.)14. Normalized course grades are the deviation of the student’s grade from the course–section–semester meandivided by the course–section–semester standard deviation.
1004 The Journal of Human Resources
cumulative GPA of at least 2.5, and f(CenterCumGPAi) is a function of the runningvariable, which is the student’s first-semester GPA centered at the 2.5 eligibilitythreshold. Finally, yt is a fixed effect for each academic semester. The coefficient ofinterest, b1, is the effect of being eligible to affiliate with a Greek organization duringsecond semester on the outcome of interest.15
Following Imbens and Lemieux (2008), we estimate the discontinuity using locallinear regression and uniform kernel weights.16 For consistency across outcomes andspecifications, we present themain results using a bandwidth of 0.4 grade points (that is,2.1–2.9). In prior versions of the paper, we have shown similar results using bandwidthsof 0.5 or 0.3 grade points, and we continue to show how the main estimates vary acrossbandwidths.17 Since our running variable of cumulative GPA entering second semester
Figure 1Discontinuity in Second-Semester PledgingNotes: Each circle is the mean of the outcome in an interval of –0.025 around the point (including the lower butnot upper boundary). Sample restricted to students first enrolling in fall 2009 through fall 2017. Figure includesseparate linear fits on either side of the threshold using a bandwidth of 0.4 GPA points.
15. Our approachmost closely follows the regression discontinuity approach employed byLindo, Sanders, andOreopoulos (2010), which uses a GPA cutoff for academic probation to estimate the impact of probation onsubsequent academic performance.16. Imbens and Lemieux (2008) argue that differences in estimates when using more complex weightingschemes are primarily indicative of sensitivity to bandwidth choice.17. In Online Appendix Table A2, we show that results are similar when using the mean square error optimalbandwidth from Calonico, Cattaneo, and Titiunik (2014) and when using triangular kernel weighting, whichapplies more weight to observations closer to the threshold.
Even and Smith 1005
is discrete (measured to the nearest hundredth), we cluster the standard errors at the GPAlevel as recommended by Lee and Card (2008).
A. Evidence of the Validity of the RD Design
Nonrandom sorting is the central threat to any RD design in which individuals whocould be affected by the policy may know the cutoff. In our setting, strategic sortingcould occur if students otherwise just below the eligibility cutoff were actively influ-encing their GPAs to become eligible. This could take the form of students lobbyingteachers to award them higher grades or students exerting precisely enough effort toachieve cumulative GPAs of 2.5.By focusing on eligibility status entering second semester, we decrease the likelihood
of strategic manipulation. Students in their first semester may be unaware of universitypolicies and are less familiar with college grading standards than more experiencedstudents. According to university statistics, for 2013–2018, more than 1,000 studentswho were academically ineligible registered for Greek recruitment before their firstsemester GPA was known, suggesting that they were unaware of the requirement oroverly optimistic about their GPAs. Further, it would be difficult for students to hitspecific cumulative GPAs because incoming students must complete at least 12 credithours to become Greek-eligible, so their cumulative GPAs represent grades acrossseveral courses taken concurrently.As noted by McCrary (2008), if subjects were able to manipulate the outcome to just
meet the eligibility requirement, wewould expect a disproportionate number of studentsto have cumulative GPAs falling just above the 2.5 threshold. Online Appendix FigureA1 shows the distribution of observations around the 2.5 GPA threshold. Visually, thedistribution is smooth across the threshold, consistent with no manipulation.18 Esti-mating Equation 1 using the number of observations at each GPA level as the dependentvariable yields a statistically insignificant discontinuity of 5.6 (p-value of 0.54).The validity of the RD design also hinges on the assumption that additional char-
acteristics related to academic performance other than Greek affiliation vary continu-ously at the cutoff.19 To further probe this assumption, Table 2 reports the extent towhich several observable variables that could plausibly affect academic outcomes arecontinuous across the cutoff. Columns 1–5 of Table 2 contain a measure of academicaptitude (ACT score) and several demographic characteristics: minority status, age atuniversity entry, state resident status, and financial need. We do not find statisticallysignificant (at the 0.10 level) discontinuities in any of these variables. Further, themagnitude of the point estimates are negligible, suggesting that individuals on eitherside of the cutoff have similar background characteristics.20
18. We present Online Appendix Figure A1 using bins of 0.1 GPA points rather than 0.05 GPA points toaddress mechanical heaping that occurs at increments of 0.1 GPA points. For example, more students will havea first-semester GPA of 2.70 than 2.71, because more combinations of course GPAs average to an even tenth.Bins of 0.1 include one “heaped” GPA in each bin. Figure 2 demonstrates that after accounting for thismechanical heaping, the distribution of observations is smooth around the threshold. See Barreca, Lindo, andWaddell (2016) for a further discussion of heaping in RD settings.19. The primary concern is not that observable characteristics are smooth across the threshold, as we cancontrol for observables directly. However, discontinuities in observable characteristics may suggest that thereare also discontinuities in unobservable determinants of academic performance.20. Conclusions are the same (none significant at the 0.1 level) for bandwidths of 0.3, 0.35, 0.4, 0.45, and 0.50.
1006 The Journal of Human Resources
Tab
le2
Discontinuitiesin
ObservableCharacteristics
IndividualCharacteristics
First-Sem
esterOutcomes
BestACT
Equivalent
Dom
estic
Minority
Medium
orHigh
FinancialNeed
State
Resident
Age
atEntry
With
draw
from
Any
Course
CourseCredits
Com
pleted
Levelof
Course
Average
Grade
forCourse
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Discontinuity
-0.136
-0.010
0.005
0.013
-0.011
0.014
0.078
0.008
0.008
(0.205)
(0.022)
(0.020)
(0.021)
(0.021)
(0.025)
(0.124)
(0.008)
(0.013)
Observatio
ns5835
5835
5835
5835
5835
5835
5835
5835
5835
Meanof
dep.
var.
26.06
0.18
0.28
0.64
18.05
0.14
14.79
1.13
3.05
Regressions
testfordiscontin
uity
inobservablecharacteristicsatthe2.5GPA
eligibility
requirem
entforfirstsemester.Allmodelsarelin
earin
therunningvariable(G
PA),allowfor
differentslopeson
each
side
ofthethreshold,anduseabandwidthof
0.4GPA
pointsandauniformkernel.O
bservatio
nsareatthestudentlevel.T
hesampleisrestricted
tostudentsfirst
enrolling
infall2009
throughfall2017.L
evelof
course
representsthesuggestedyearthecourse
betaken(1=firsty
ear,2=second
year,etc.).S
tandarderrorsclusteredon
GPA
(tothe
nearest1,000th)
arein
parentheses.*p
<0.10,*
*p<0.05,*
**p<0.01.
1007
In Columns 6–9 we explore whether students just above the threshold exhibitdifferent course-taking behavior during first semester. Attempts to manipulate first-semester GPA could involve strategically withdrawing from challenging courses orselectively enrolling in easier coursework. Again, we do not find any significant (at the0.10 level) discontinuities in coursewithdrawals, completed credits, or course difficulty,as measured by either the suggested year of enrollment or the average grade assigned tostudents in a course–section excluding the student themselves. Collectively the resultspresented in Table 2 suggest that students just on either side of the threshold have verysimilar characteristics upon enrollment and exhibit similar first-semester course-takingbehavior, which supports the validity of the RD design.Finally, the university that we study has no other policies that require a 2.5 GPA at the
end of the first semester. One common university policy with a GPA threshold issatisfactory academic progress—a set of requirements for students receiving federalfinancial aid. Undergraduate students at this institution are required to achieve a cu-mulative GPA of at least 2.0 following four semesters of enrollment. A cumulative GPAof 2.0 is also required to avoid academic probation. To avoid confounding our estimateswith the effect of these other treatments, we restrict our RD analysis to bandwidths of 0.5or less (that is, qualifying GPAs of 2.0–3.0).
B. RD Estimates for Academic Performance
If Greek affiliation affects academic performance, a discontinuity in academic outcomesshould appear at the eligibility threshold analogous to the discontinuity in Greek af-filiation. Figure 2 shows the discontinuity in several measures of course grades forall coursework following first semester (hereafter referred to as subsequent grades).Regardless of how course grades are measured, there is a visual drop in academic per-formance at the eligibility threshold, suggesting that Greek affiliation hinders classroomperformance.Table 3 shows the estimated impact of affiliating with a Greek organization second
semester on subsequent academic performance. To calculate the impact of Greek af-filiation on academic performance, the discontinuity in subsequent grades (Figure 2)must be reweighted by the discontinuity in pledge probability (Figure 1). Following thestandard approach for fuzzyRDdesigns (Hahn, Todd, andVan derKlaauw2001),we dothis using two-stage least squares as described below:
(2) Gradeijt = b0 +b1 dAffiliateSecSemi + f (CenterCumGPAi)
+ f (CenterCumGPAi)�Eligiblei + ht + eijtwhere second-semester affiliation is instrumented using the 2.5 eligibility requirement.Since the first-stage discontinuity in affiliation always has a first-stageF-statistic greaterthan 100, we need not be concerned about weak instruments. The corresponding fuzzyRD estimate shown inColumn 1 of Table 3 implies thatGreek affiliation reducesGPA insubsequent coursework by about 0.26 standard deviations.In order to better understand our central result, we explore the impact of Greek
affiliation on several course grade outcomes. In Columns 2–5 of Table 3 we presentresults for grade point and binary indicators for earning specific letter grades. To makethe results more comparablewith those presented inColumn 1,which address differences
1008 The Journal of Human Resources
in course-taking behavior by rescaling a student’s grade relative to their classmates, weinclude course fixed effects.21
Figure 2Discontinuities in Subsequent Course GradesNotes: Each circle is the mean of the outcome in an interval of –0.025 around the point (including the lower but notupper boundary). Sample restricted to students first enrolling in fall 2009 through fall 2017. Each observation is anindividual course grade for courses taken in the second semester or beyond and capped at 12 semesters of experience.Observations are weighted by credit hour. Figure includes separate linear fits on either side of the threshold using abandwidth of 0.4 GPA points.
21. One complication with using course fixed effects in the RD design is that not all students who take a givencourse have qualifying GPAs that are within the relevant bandwidth. Because our goal in using this control is toaddress the average difficulty in the class across all students, we adopt the two-stage augmented local linear
Even and Smith 1009
Tab
le3
RegressionDiscontinuityEstimates
oftheEffectof
Greek
Affilia
tionon
CourseGradesandSelection
Variables
Normalized
CourseGrades
Grade
Point
(Unnormalized)
Grade
ofA
orB
Grade
ofC
Grade
ofD,F
,or
With
draw
alLevelof
Course
Average
Grade
inCourse
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Affiliatesecond
semester
-0.264**
-0.236**
-0.0918**
0.0238
0.0657**
-0.148**
-0.0577
(0.126)
(0.098)
(0.0391)
(0.0244)
(0.0264)
(0.0679)
(0.0474)
Coursefixedeffects
No
Yes
Yes
Yes
Yes
No
No
Observatio
ns137,429
137,094
142,424
142,424
142,424
142,424
137,108
Notes:D
ependentvariablelistedattopof
each
column.Allmodelsarelin
earintherunningvariable(G
PA),allowford
ifferentslopes
oneach
side
ofthethreshold,andusea
bandwidthof
0.4GPA
pointsandauniformkernel.A
dditionalcontrolsincluded
ineach
regression
areentering
age;ACTscore;indicatorsforfem
ale,domestic
minority,state
resident,and
mediumor
high
financialneed;andfixedeffectsforsem
estersof
experience
andacadem
icsemester.The
sampleisrestricted
tostudentsfirstenrollin
ginfall2009
throughfall2017,cappedat12
semestersof
experience.C
olum
ns2–5includecourse
fixedeffectsusingthetwo-stageaugm
entedlocallinearapproachoutlinedinHausm
anandRapson(2018).L
evelof
course
representsthesuggestedyearthecourse
betaken(1=firsty
ear,2=second
year,etc.).A
verage
gradein
course
istheaverageletterg
rade
received
bystudents(excluding
thestudenth
imor
herself)in
thesamesectionof
thecourse.S
tandarderrorsarein
parentheses.In
Colum
ns1,
6,and7standard
errorsare
clusteredon
qualifying
GPA
(tothenearest100th).In
Colum
ns2–
5standard
errorsarederivedfrom
200bootstrapped
samples
toallowthefirst-stagevariance
tobe
reflected
inthesecond
stage.*p
<0.10,*
*p<0.05,*
**p<0.01.
1010 The Journal of Human Resources
Columns 2–5 of Table 3 present evidence that overall grade point falls substantiallydue to a shift down the grade distribution. Grade point falls by 0.24 GPA points on thetypical 0–4 scale (Column 2 of Table 3). The probability of attaining an A or B falls bynine percentage points, there is an insignificant increase in the probability of attaining aC, and there is a six percentage point increase in the probability of attaining a D or F orwithdrawing from the course.22
Columns 6 and 7 of Table 3 present the estimated impact of affiliation on subsequentcourse-taking behavior. While Columns 1–5 address selection into courses by rescalinggrades within a course or through the use of course fixed effects, course selection itselfcould be affected. For instance, if Greeks expect that the time commitment associatedwith Greek life will reduce their time available for coursework, they may strategicallychoose easier courses. In Column 6, we present the effect of Greek affiliation on thesuggested year of enrollment for a course (for example, 1 for 100 level courses and 4 for400 level courses). The results imply that Greeks take courses that on average have alower suggested year of enrollment, perhaps reflecting easier coursework. In Column 7,we test whether affiliation causes students to select course–sections with higher averagegrades excluding the student themselves. If Greek affiliation causes students to seek outeasier course–sections with higher grades, we should find a positive effect of affiliationon the average course grade. However, we find that affiliation has no statistically sig-nificant effect on the average course grade of courses chosen. Because the intensity ofGreek activities varies across semesters, we explore the dynamic effects of affiliation oncourse-taking behavior later in the panel analysis.In Online Appendix Table A2 we provide evidence that the central course grade
estimates are robust to several alternative specifications. Consistent with the identifyingassumption that other determinants of academic performance vary smoothly around theeligibility threshold, the omission of additional controls leaves estimates largely un-changed (Column 2 of Online Appendix Table A2). Results are also similar withoutweighting by credit hour (Column 3) orwhen omitting studentswhohad qualifyingGPAsof exactly 2.5 (Column 4). Finally, estimates are fairly stable across bandwidths, in-cluding the bandwidth suggested by the mean-squared-error optimal bandwidth selectorproposed by Calonico, Cattaneo, and Titiunik (2014) (Online Appendix Figure A2).One additional concern is that ourmain specification could overstate or understate the
effect of Greek affiliation if students just on one side of the cutoff disproportionatelyaffiliate later in their academic careers. For instance, if all students who just miss thecutoff affiliate the following spring, our model estimates the effect of affiliating one yearearly. In Online Appendix Table A3 we present evidence that there are minimal dif-ferences in late-joining behavior around the cutoff. Accounting for late joiners under theassumption that treatment effects are independent of when a student joins yields similarbut slightly larger estimates than those presented in the main text (see Online AppendixTable A4).
approach suggested by Hausman and Rapson (2018). In the first stage, we use the full sample to residualize theoutcome (for example, course grade point), partialling out the course fixed effects and other controls. In thesecond stage, we run the local linear RD regression on the residualized data. Finally we bootstrap the standarderrors to allow the first-stage variance to be accounted for in the second stage.22. Estimating the analogous RD regression for cohorts entering in 2007 and 2008 (pre-policy) reveals nodecrease in subsequent grades for students just above the 2.5 threshold, assuaging concerns that some otherpolicy associated with a 2.5 threshold is causing the reduction in grades (Online Appendix Table A5).
Even and Smith 1011
To provide some perspective on the size of the estimatedGreek effects on grades (0.26standard deviations or 0.24 grade points), Carrell, Fullerton, and West (2009) find thatbeing randomly assigned to a peer group with SAT verbal scores one standard deviationabove the mean increases course grades by about 0.08 standard deviations. BecauseGreek affiliation alters peer groups and involves a substantial time commitment out-side of class, a better analog may be participation in collegiate athletics. Maloney andMcCormick (1993) find that athletes in revenue-generating sports perform about 0.4grade points worse in season than they do out of season. In short, the estimated decline inacademic performance associated with Greek affiliation is large, but consistent withmechanisms identified in prior work.These estimates represent a local average treatment effect: the impact of joining a
Greek organization during second semester of the first year on subsequent grades forstudents who marginally exceed the eligibility threshold. One concern is that thesemarginally eligible pledges may be atypical in someway or may only represent pledgesat a small share of chapters that find the 2.5 threshold binding. Reassuringly, 54 of the 61chapters (89 percent) in the sample admitted at least one new pledge with a qualifyingGPA between 2.5 and 2.6, and 60 of the 61 chapters (98 percent) admitted students withqualifying GPAs between 2.5 and 2.9 that would be within our primary bandwidth forthe RD. Therefore, the RD captures an effect of Greek affiliation derived frommarginalpledges from the vast majority of chapters. Speaking to the generalizability of theseestimates, a GPA requirement of 2.5 is quite common nationally.23 Any policy changethat adjusts this eligibility threshold, for instance, by increasing its stringency, wouldimpact these marginal pledges. To understand whether Greek students with GPAsmarginally above the threshold experience a similar treatment effect to those furtherfrom the cutoff, we consider the fixed-effects model stratified by qualifying GPA in thenext section.
IV. Individual Fixed-Effects Approach
While the RD analysis provides compelling evidence that Greek af-filiation has a causal negative effect on academic performance, the RD design is limitingin several ways. As we describe below, exploiting the panel feature of our data to correctfor the endogeneity of Greek affiliation allows for a more in-depth analysis of howGreek affiliation affects academic performance over the life cycle of the student, as wellas how the effects vary across different types of students. The fixed-effects regressionspecification we use for the analysis can be described as the following difference-in-difference model that compares the time-path of grades for those who do versus do notaffiliate with Greek life:
(3) Gradeijt =Greekitb +Xitc +ai + ht + lijt
23. The North-American Interfraternity Conference (NIC), which is the governing body for many fraternitiesin the United States, requires that a potential new member who has established a collegiate GPA have at least a2.5 GPA to join. Some institutions allow incoming students to affiliate during their first semester, prior toestablishing a collegiate GPA.
1012 The Journal of Human Resources
where Gradeijt is the normalized course grade for student i in course j in semester t.Greekit is a dummy variable that indicates whether student i in semester t is currentlypledging or previously joined a Greek organization. Individual fixed effects, ai, controlfor unobserved time-invariant characteristics that potentially affect a student’s grades,such as innate ability. The yt represent fixed effects for academic terms, and Xit is avector of controls for individual characteristics that vary over semesters. Our preferredspecification includes fixed effects for experience (that is, semesters since initial en-rollment) to address the possibility that students may systematically improve or weakenas they approach graduation. Standard errors are corrected for clustering at the studentlevel.24
In the fixed-effectsmodel, the identification of theGreek effect on grades relies on thefact that students are not allowed to be part of the Greek system in their first semester ofenrollment. In the second semester, students who complete 12 credit hours andmeet thenecessary 2.5 GPA requirement may choose to rush and pledge. If there is a Greek effecton grades, we should see the grade–experience profiles of Greek and non-Greek stu-dents begin to diverge at the second semester.Figure 3 compares the GPA profile of students who affiliate with Greek organizations
at some point during their time at the university (“ever-Greeks”) with two comparisongroups: (i) the full sample of students who never affiliate with Greek organizations(“never-Greeks”) and (ii) the never-Greeks who were eligible to affiliate with a Greekorganization during their second semester. Notably, the average GPA for the full sampleof never-Greeks starts much lower than ever-Greeks and exhibits a consistently positivetrend. In contrast, eligible non-Greeks (that is, thosewith at least a 2.5GPA and 12 credithours at the end of their first semester) have a very similar starting GPA to ever-Greeks,but experience a second semester decline in GPA that is only about one-third of thatexperienced by ever-Greeks (0.05 versus 0.14). For 85 percent of ever-Greeks, pledgingoccurs in the second semester, and this is wherewewould first expect to see a divergencein grade profiles if affiliation affects performance. It is worth noting that the ever-Greekgrade disadvantage that first appears in the second semester does not disappear in theremainder of the traditional four years of college.
A. Fixed-Effects Results
In Table 4 we present our main results for the difference-in-difference specification. Inthis analysis, the Greek dummy variable (the treatment variable) turns to one in the firstsemester that a student pledges or joins Greek life and is unchanged for the remainder ofthe student’s college career. In later analysis, we consider a more detailed breakdown ofthe effect of a person’s Greek involvement.Gender imbalances in college enrollment and success arewell documented, with 1.34
female bachelors’ degree recipients for every male recipient (U.S. Department ofEducation 2018, Table 322.20). One narrative for the relative success of females is thatbehavioral and developmental advantages facilitate improved preparation prior to en-tering college (Goldin, Katz, and Kuziemko 2006). Recent studies find that collegeinterventions are oftenmore effective for females (Angrist, Lang, andOreopoulos 2009;Carrell and Sacerdote 2017), and there is some evidence of gender-based heterogeneity
24. The model cannot simultaneously include fixed effects for students, experience, and academic term, sincethey are perfectly collinear. Hence, we do not include fixed effects for academic term in the model.
Even and Smith 1013
in the effect of Greek affiliation on risky behaviors (DeSimone 2009). Motivated by thepossibility of different experiences in fraternities versus sororities and the possibilityof different behavioral responses to similar treatment, we stratify results for the panelanalysis by gender.In Column 1 of Table 4, we show the estimated effect of affiliation from a regression
of a student’s normalized course grade on an indicator for Greek affiliation. To addressthe possibility that students may systematically improve or weaken as they progresstowards degree completion, we control for experience using semester-of-experiencefixed effects. Using the full sample of students, we find that sorority members perform0.14 standard deviations worse, and fraternity members perform 0.17 standard devia-tions worse, relative to what we would expect based on their prior performance andaccumulated experience.While we present estimates relative to all non-Greeks for comparison, our preferred
specification limits the sample to only students eligible to affiliate at the end of their firstsemester. The eligible never-Greeks form a more analogous comparison group becausethey have the same opportunity to affiliate and they exhibit amore similar distribution ofinitial grades. Further, differential dropout rates are likely to overstate the academicimprovement of the full sample of never-Greeks. Students who start with low initialgrades and do not improve may be subject to academic suspension and even dismissal.Because the threshold for academic probation (and eventually academic suspension) at
Figure 3GPA Profiles of Greek and Non-Greek StudentsNotes: Observations areweighted by credit hours. Eligible means the student had a first-semester GPA of 2.5 orabove and completed at least 12 credit hours, enabling them to pledge during second semester.
1014 The Journal of Human Resources
the university is 2.0, students with the highest danger of academic suspension followingfirst semester are only present in the full-sample comparison group. Focusing on stu-dents who are Greek eligible at the end of the first semester mitigates the concern thatcompositional changes between the Greek and comparison group are driving results.Imposing the 2.5GPA restriction for all students reduces the estimated negative effect ofGreek affiliation from -0.14 to -0.08 for women, and from -0.17 to -0.12 for men.Since we believe that the restricted sample is more likely to satisfy the parallel trendsassumption and yield an unbiased estimate of the Greek effect, the remainder of ouranalysis with the panel data will impose that restriction.In Specifications 3 and 4, nonnormalized numerical grades (ranging from 0 for F to
4.0 for A) are used as the measure of academic performance. Course-specific fixedeffects are added to the model to control for the level of difficulty in the course. Thepattern of results is similar to that obtained with normalized grades. Namely, Greekaffiliation has a statistically significant negative effect on grades for both sexes, theeffects are larger for men than women, and the effects are smaller when the sample isrestricted to those eligible to affiliate at the end of the first semester.25
Table 4Effect of Greek Affiliation on Grades
Full Sample,NormalizedGrades
2.5+ GPA inFirst Semester,
Normalized GradesFull Sample,Actual Grades
2.5+ GPA inFirst Semester,Actual Grades
(1) (2) (3) (4)
Panel A: Females
Greek -0.137*** -0.083*** -0.099*** -0.053***(0.006) (0.006) (0.005) (0.005)
Observations 589,442 521,883 584,942 516,820
Panel B: Males
Greek -0.171*** -0.117*** -0.136*** -0.085***(0.008) (0.008) (0.006) (0.007)
Observations 520,045 426,754 514,161 419,806
Notes: Dependent variable is normalized course grade in Specifications 1 and 2 and the actual grade (rangingfrom 0 to 4) in Specifications 3 and 4. Greek dummy is first set to one in the semester the student pledges and inall subsequent semesters. Specifications 2 and 4 restrict the sample to students with a GPA of 2.5 or above atthe end of the first semester since this is the sample of students eligible to pledge in the second semester. Allregressions are weighted by credit hour and include student and semesters of experience fixed effects. Themodels using actual grades also include course fixed effects. Standard errors corrected for clustering at thestudent level are given in parentheses. *p< 0.10, **p< 0.05, ***p< 0.01.
25. Given that the standard deviation of grades in the average course is 0.74, adjusting for differences in theunits of measurement implies that the estimated Greek effects on grades are very similar whether usingnormalized grades or actual grades with course fixed effects added as controls.
Even and Smith 1015
B. Evidence on Identifying Assumptions
One potential concern with the above difference-in-difference estimates is that theirvalidity rests on the identifying assumption that, absent Greek affiliation, the grade–experience profiles of Greek and non-Greek students would be identical. This as-sumption would be problematic if those with a desire to affiliate increase their effortlevel in the first semester to improve the odds of being eligible and, after meeting the 2.5GPA requirement, revert to their normal level of effort and have their grades fall back toor below the 2.5 GPA. If such “pre-trends” in the Greek versus non-Greek grade profilesexist, this would cast some doubt on the validity of the identifying assumption.Since most students pledge in their second semester, we have limited information on
pre-trends. However, about 15 percent of thosewho affiliatewith Greek life pledge aftertheir second semester. To test for pre-trends, a model with student and experience fixedeffects is estimated with interactions between an ever-Greek dummy variable and a“time-to-pledge” variable. The ever-Greek dummy is set to one for all students who everaffiliate with Greek life and is zero otherwise. For ever-Greeks, the time-to-pledgevariable is the number of semesters relative to the pledge semester. For those who neveraffiliate, the interaction is set to zero in all periods. We normalize the coefficient on theinteraction to zero for the pledge semester. If there is a pre-trend in the data where thetreated (ever-Greek) display a different grade profile than the nontreated, the coefficientsin semesters prior to the pledge semester would be nonzero.Figure 4 plots the coefficients on the interaction terms along with 95 percent confi-
dence intervals. There is no evidence that, prior to affiliation, those who eventuallyaffiliate have grade profiles that differ from those who never affiliate. This supports ouridentifying assumption that, absent Greek affiliation, the treated and untreated groupswould have parallel grade–experience profiles.The plot also reveals several patterns in the data that are consistent with a decrease in
academic performance caused by Greek affiliation. First, there is a stark relative drop inGreek normalized grades starting in the semester of pledging and persisting for severalsemesters. Second, the magnitude of the reduction in grades varies predictably with theintensity of Greek life. The largest gap in performance occurs during the student’spledge semester (0.15 standard deviations), which is typically thought of as the mostchallenging and time-intensive semester of Greek life. There is some recovery in gradesafter the pledge semester, but the effects remain negative and statistically significant(0.05 level) for four semesters after pledging. Moreover, there is a sawtooth pattern tothe effects that is consistent with the negative effects being larger in the spring semesterof every year—when pledging occurs.26 For active members, recruitment during thespring semester involves hosting activities for prospective pledges followed by theprocess of deciding on bids, educating new members, and attending or organizingsocial activities following initiation. During the last six years of our sample (the farthestback that such records exist), Greek organizations registered 24 percent more socialevents with alcohol during spring than in fall semesters. While social activities are more
26. Online Appendix Table A7 presents the semester-by-semester dynamics using the RD approach andreveals a similar sawtooth pattern, with large effects occurring most spring semesters and the largest effectoccurring in the third semester of affiliation. Lack of power prohibits us from concluding that any semestereffect is significantly different from the pledge semester.
1016 The Journal of Human Resources
time-consuming for Greeks in the spring semester, the opposite is true for time spent incommunity service. The data available for the last four years of our sample show thathours of community service per member is 16 percent lower (about 0.5 hours) in thespring than the fall semester. It appears that the pledge semesters are not only hard on therecruits, but on the active members as well.27
C. Heterogeneous Effects of Greek Affiliation
One strength of our administrative data set relative to that used in previous studies is thatit tracks students’ Greek status by semester. This allows us to document the impact ofGreek life on academics across the Greek life cycle—from unaffiliated, to pledge, toactive member, and when applicable, to inactive member. To examine effects across theGreek life cycle, we replace the single Greek indicator used in our earlier analysis withthree mutually exclusive dummy variables: Greek pledge, Greek active, and former
Figure 4Relationship between Greek Affiliation and Normalized GradesNotes: Figure provides point estimates (and 95 percent confidence intervals) for difference between ever-Greekand non-Greek normalized grades. Estimates represent coefficients on interaction between semesters relative topledge semester and a dummy variable for whether a student ever affiliates with a Greek chapter. Interaction forsemester prior to pledging (t = –1) is the omitted dummy variable. The regression model also includes fixedeffects for each student, semesters of experience, and academic term.
27. De Donato and Thomas (2017) also find larger negative effects of Greek life in the spring semester whenpledging occurs at Duke University.
Even and Smith 1017
Greek. As noted, for most students who affiliate, pledging occurs during the secondsemester. However, about 15 percent of those who affiliate pledge after their secondsemester. Those who remain affiliated with Greek life after pledging are classified asGreek active, and those who leave Greek life are classified as former Greek. We alsoexamine the impact of Greek affiliation on other measures of academic activity beyondnormalized grades, including the effect on the number of course credits completed ineach semester, the level of difficulty of courses taken as measured by the average gradegiven in the course, and the suggested year of enrollment for the course.The results presented in Table 5 suggest that, for both sexes, the negative effect of
Greek affiliation on normalized grades is largest during the pledge semester, persistsafter pledging, and even after the student leaves Greek life. Continuing negative effectsfor former members could be due to permanent changes in behavior. For instance,Carrell, Fullerton, and West (2009) find that peer effects on academic performancefrom randomly assigned squadrons of freshman at the Air Force Academy persist ata diminished rate in subsequent years of college.28 Moreover, it is highly likely thatmembers of a suspended fraternity or sorority continue living and socializingwithmanyof their former brothers/sisters, and some of the behavior that was detrimental to theiracademic performance may persist.Greek life also reduces the number of course credits completed for both men and
women. The effect is largest in the pledge semester, drops but remains statisticallysignificant after pledging, and disappears if the student leaves Greek life. The effect ofGreek affiliation on course difficulty (as measured by average grade given in the course)and course level suggests that students try to lighten their course load during their pledgesemester by taking lower-level courses and courses that award higher-than-averagegrades. The effect on course difficulty disappears once they become active or formermembers, and the effect on course level is reversed after they become active members.Overall, the results suggest that the pledge semester has the most detrimental effect onacademic performance for Greek members and that the students try to lighten this effectby taking easier courses and fewer credit hours.From the above,we know that the pledge semester, which always occurs in the spring,
is the most challenging for new recruits. The sawtooth pattern in Figure 4 also illustratesthat the pledge (spring) semester may do harm to the active members involved in therecruiting process. By adding an interaction between a spring semester dummy and theGreek active and former Greek dummies to the models in Table 5, we can test fordifferential Greek effects in the spring versus fall semesters on the various academicoutcomes. The estimates (available in OnlineAppendix Table A8) indicate that being anactive Greekmember has larger negative effects on course grades and credits completedduring the spring semesters for both men and women. These effects are statisticallysignificant at the 0.05 level. There is also evidence that the increased probability ofwithdrawing from a course and the decrease in credit hours associated with being anactive Greek is greater in the spring semester for both sexes. Apparently, the pledgeprocess amplifies the negative effect of Greek affiliation for both the new recruits andthe active members.
28. In our setting, one potential channel for persistent declines in academic performance is alcohol use.Previous studies find that Greek affiliation causes increased alcohol consumption (DeSimone 2007, 2009) andthat excessive alcohol consumption in college can establish a pattern of overuse that extends later in life(Borsari and Carey 2001).
1018 The Journal of Human Resources
The negative impact of Greek life on academic performance described above is anaverage across a wide range of students, but the effects could differ depending on thecharacteristics of the student or chapter. If this is true, programs could be added to helpmitigate the effects for the types of students or chapters where the negative effects aregreatest. The regressions presented in Table 6 shed some light on this. In the first fourspecifications, we investigate whether the negative impact of Greek life is lower amongstudents who are better prepared academically upon entry into Greek life. Grade regres-sions are estimated separately for subsamples stratified by the student’s first-semesterGPA. The grade groupings divide the sample of those eligible to pledge at the end of thefirst semester (that is, those with a GPA of 2.5 or above) into approximate quartiles. For
Table 5Effects of Greek Affiliation on Grades, Course Credits, and Course Difficultyover the Greek Life Cycle
NormalizedGrades
Average Gradefor Course
Level ofCourse Withdraw
Course CreditsCompleted
(1) (2) (3) (4) (5)
Panel A: Females
Greek pledge -0.122*** 0.018*** -0.018*** 0.005*** -0.307***(0.007) (0.004) (0.005) (0.001) (0.038)
Greek active -0.069*** -0.000 0.012** -0.001 -0.179***(0.007) (0.004) (0.006) (0.001) (0.038)
Former Greek -0.044*** -0.001 0.027** -0.001 0.039(0.012) (0.005) (0.011) (0.002) (0.066)
Observations 521,883 521,883 530,500 530,500 104,490
Panel B: Males
Greek pledge -0.189*** 0.046*** -0.016*** 0.017*** -0.802***(0.009) (0.004) (0.006) (0.002) (0.048)
Greek active -0.087*** 0.001 0.018*** 0.004*** -0.118***(0.009) (0.004) (0.007) (0.001) (0.045)
Former Greek -0.091*** -0.006 0.030*** 0.002 -0.007(0.014) (0.006) (0.011) (0.002) (0.075)
Observations 426,754 426,754 438,263 438,263 86,649
Notes: All observations are at the student–course level. Dependent variable is listed at the top of each column.Level of course is the suggested year the course be taken (1 = first year, 2= second year, etc.). Average grade incourse is the average letter grade assigned (excluding the student themselves) in the same section of the course.Standard errors corrected for clustering at student level are in parentheses. All regressions are weighted bycourse credit hours and include student and semesters of experience fixed effects. *p< 0.10, **p< 0.05,***p < 0.01.
Even and Smith 1019
both men and women, the effect of Greek affiliation is negative and statistically sig-nificant for all four quartiles of the first-semester grade distribution. The coefficientsdiffer only slightly across the subsamples. For women, the negative effects are slightlylower among those at the higher end of the first-semester grade distribution; for men,the effects are virtually identical for students in the top and bottom quartile of the first-semester grade distribution.In Specification 5, we introduce an interaction between the Greek dummy and a
dummy indicating whether a person had an ACT below the median for the university.For women, there is no evidence that Greek life has a differential effect for those withhigh versus low ACT scores. For men, the negative effect of Greek life is about 30percent larger among those with low ACT scores. Finally, Specification 6 includesan interaction between the Greek dummy variable and a dummy indicating whether a
Table 6Heterogeneity in Greek Effects on Normalized Grades
Sample Restrictions on First-Semester GPA
[2.5,3.0] (3.0,3.30] (3.30,3.65] (3.65,4.0] [2.5,4.0] [2.5,4.0](1) (2) (3) (4) (5) (6)
Panel A: Females
Greek -0.098*** -0.085*** -0.088*** -0.067*** -0.086*** -0.074***(0.016) (0.015) (0.011) (0.009) (0.008) (0.007)
Greek ·Low ACT 0.011(0.009)
Greek ·Mediumor HighFinancial need
-0.035***(0.012)
Observations 105,356 112,964 145,749 157,814 521,883 521,883
Panel B: Males
Greek -0.126*** -0.104*** -0.115*** -0.132*** -0.103*** -0.110***(0.017) (0.016) (0.014) (0.015) (0.010) (0.009)
Greek ·Low ACT -0.033**(0.013)
Greek ·Mediumor HighFinancial need
-0.037**(0.017)
Observations 118,398 98,093 112,903 97,360 426,754 426,754
Notes: Dependent variables is normalized course grade. Standard errors corrected for clustering at student level are inparentheses. All regressions are weighted by course credit hours and include student and semesters of experience fixedeffects. *p < 0.10, **p< 0.05, ***p< 0.01.
1020 The Journal of Human Resources
student has medium or high financial need when admitted to the university. Affiliatingwith and remaining in a Greek organization can be expensive, with current dues aver-aging near $500 per semester. The negative coefficients on the interaction variable forboth men and women suggest that Greek affiliation has a more detrimental effect on theacademic performance of those with greater financial need. This larger negative effectmay occur because Greek life could place greater pressure on those with financial needand increase the chance that they have to take paid work.The effect of Greek life may also differ across chapters. Anecdotal evidence suggests
that there is variation across chapters in terms of the emphasis placed on the academicperformance of their members that could affect the selection process for new recruits, aswell as the incentives or support provided to improve academic performance of activemembers. To investigate the degree of heterogeneity across chapters, we estimate theGreek effect of each of the 61 chapters on normalized grades.29 The mean of the pointestimates is -0.14, and themedian is-0.10. The estimates range from a low of-0.62 to ahigh of 0.34 and are negative for all but one chapter, and 51 of the 61 point estimates arenegative and statistically significant at the 0.05 level.30 Inspection of the point estimatesreveals significant variation in the impact of affiliation across chapters.Unfortunately, our data provide relatively little information on chapter characteris-
tics that would allow us to explore what causes the effects to vary across chapters.Undoubtedly, much of the variation depends on the culture that a chapter builds overtime, and we have no objective measure of that. However, some chapter-specific in-formation can be constructed, including the number of members in each chapter, theaverage grade of new recruits (as measured by their GPA in the semester prior topledging), and the number of chapter siblings (that is, fraternity brothers or sororitysisters) that each student has in a given course. Chapter size could have a positive ornegative effect on performance. A larger network of siblings could mean students haveaccess to a greater network of support for classes. On the other hand, larger chapterscould result in more distractions for the student. More siblings in a course shouldimprove one’s support network and improve grades. Finally, a higher GPA among newrecruits could be a signal of the chapter’s attention to academic performance and couldresult in positive peer effects.The regression estimates in Table 7 reveal the heterogeneity in effects across chapters
based on the above characteristics. In the first two specifications, the grade effects ofchapter size and the number of siblings in courses are estimated for men and women.The model also controls for class size so as not to conflate the effect of siblings in classwith the size of the course. The results suggest that larger chapters have a more detri-mental effect on normalized grades for both men and women, and the effect is larger formen than women. Having more siblings in a course has a positive effect on grades forboth sexes, though the effect is greater for men than women. In the third and fourthspecification, we add interactions between the Greek dummy and dummies indicatingwhether the chapter’s pledge class was in the middle two or top quartile of the pledgeGPA distribution. For men, a higher-quality pledge class reduces the negative effect of
29. This is accomplished by replacing the single Greek dummy in the models in Table 4 with a chapter dummythat equals one beginning in the semester that a student pledgeswith a chapter that remains one for the rest of theacademic career.30. A plot is provided in Figure A3 of the Online Appendix.
Even and Smith 1021
Greek life. There is a positive effect for women as well, but the effect is not statisticallysignificant at the 0.10 level. Overall, the regressions in Table 7 suggest that largerchapters and chapters with lower standards for new recruits have a more harmful impacton grades. The effect of larger chapters could be partiallymitigated, however, bymakingit more likely that chapter members have siblings in their courses. The impact of chaptersize, siblings in course, and pledge quality is greater in fraternities than sororities. Thismay suggest that both the positive and negative peer effects of Greek life on academicperformance are greater for men than women.
V. Long-Run Educational Outcomes
We now consider some of the long-run impacts of affiliating with aGreek organization during second semester. Given the large role of Greek organizationsin the lives of members, and the substantial impacts we find on classroom performance,it is plausible that longer-run outcomes, such as persistence, degree attainment, and evengraduate school enrollment, could be affected. Since each of these long-run outcomes is
Table 7Greek Effects on Grades: Heterogeneity across Chapters
Females Males Females Males(1) (2) (3) (4)
Greek -0.106*** -0.149*** -0.114*** -0.166***(0.009) (0.010) (0.024) (0.020)
Greek ·Number of membersin chapter (in 100s)
-0.029*** -0.102***(0.006) (0.015)
Greek ·Number of siblingsin course
0.008*** 0.025***(0.002) (0.002)
Number enrolled in course(in 100s)
-0.041*** 0.025***(0.003) (0.003)
Greek ·Pledge GPAin middle two quartiles
0.031 0.041*(0.024) (0.021)
Greek ·Pledge GPAin top quartile
0.042* 0.083***(0.024) (0.022)
Observations 589,442 520,045 521,883 426,754
Notes: Standard errors corrected for clustering at student level are in parentheses. All regressions are weightedby course credit hours and include student and semesters of experience fixed effects. Sample restricted to thosewith GPA of 2.5 or better at end of first semester. Number of members in chapter, number of siblings in course,and number enrolled in course are all measured as deviations from sample means.*p < 0.10, **p < 0.05,***p < 0.01.
1022 The Journal of Human Resources
only realized once per individual, we return to the RD model. Table 8 presents the RDestimates, and Online Appendix Figure A4 displays the corresponding RD figures forselect outcomes.A common narrative about Greek organizations is that they provide a sense of
community to members, which in turn may increase student retention. Column 1 ofTable 8 tests this hypothesis using retention to sophomore year as the outcome. Both thevisual and regression evidence suggests that, if anything, retention declines. The as-sociated point estimate for our preferred specification implies roughly a five percentagepoint decline in retention, though we cannot rule out positive effects. It is important tonote that sophomore retention at this university is quite high relative to national aver-ages, with 92 percent of first-year students returning for sophomore year. The lack of apositive effect on retention could be due in part to the high baseline level of retention,and it is plausible that the effect could differ at institutions with lower retention rates.Columns 2–4 of Table 8 show the estimated impacts on whether or not a student has
graduated within four, five, or six years of their enrollment date. The estimates areimprecise, but consistently negative, which suggests that affiliating with a Greek or-ganization during second semester reduces the probability of graduating for the averageGreek student near the 2.5 qualifying GPA threshold. Column 5 shows that graduationGPA falls, but by less than the estimates using the student–course-level data wouldsuggest (0.093 versus 0.236 grade points). The majority of this difference is due todifferences in the sample of graduates versus all students. Considering an outcome ofcumulative GPA through final semester (Column 6) increases the magnitude of theestimate to -0.16 grade points.31 We also find an imprecise reduction in the probabilityof graduating with at least a 3.0, which is a common admissions requirement forgraduate schools.Columns 8–10 show the estimated impacts on graduation major and enrollment in
graduate school for the sample of graduates. We find imprecise reductions in theprobability of graduating with the initial major declared and the probability of gradu-ating with a STEM degree.32 Finally, Column 10 uses data from the National StudentClearinghouse to estimate the impact on graduate school attendance. The point estimatesuggests a small reduction in the probability of graduate school attendance that is notstatistically significant.Collectively the analysis of long-run educational outcomes provides suggestive ev-
idence that Greek affiliation hinders the academic development of students beyondtransitory reductions in course grades. While the estimates are rarely significant at
31. Effects on cumulative GPA or graduation GPA are also attenuated relative to estimates using the student–course-level microdata mechanically since cumulative GPA averages course grades across a student’s entireacademic career. For Greeks, this includes both pre- (untreated) and post-affiliation (treated) semesters. For thestudent–course-level data we only consider grades earned post-affiliation. Also, the estimates with student–course-level data include course fixed effects and therefore adjust for the difficulty of courses selected. Thecumulative GPA estimates will therefore be attenuated if affiliation leads to a selection of easier course work.32. The definition of STEMmajors is based on CIP (Classification of Instructional Program) codes used by theDepartment of Homeland Security for purposes of the 24-month STEM optional practical training extension.The STEM list is available at https://www.ice.gov/sites/default/files/documents/Document/2016/stem-list.pdf(accessed October 14, 2021). The STEM dummy variable is set to one if any of a student’s majors is in a STEMfield.
Even and Smith 1023
Tab
le8
RegressionDiscontinuityEstimates
ofGreek
Affilia
tionon
Long-Run
Educatio
nalOutcomes
Variables
Sophomore
Retentio
n
Graduate
with
in4Years
Graduate
with
in5Years
Graduate
with
in6Years
Graduation
GPA
Cum
ulative
GPA
through
FinalSem
ester
Graduation
GPA
‡3.0
Graduate
with
Sam
eMajor
STEM
Graduate
Enrolled
inGraduate
School
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
Affiliatesecond
semester
-0.052
-0.143
-0.116
-0.0878
-0.0933
-0.161
-0.0185
-0.211
-0.277*
-0.0223
(0.075)
(0.137)
(0.180)
(0.169)
(0.102)
(0.104)
(0.125)
(0.196)
(0.154)
(0.090)
Observatio
ns5,301
4,234
3,658
3,036
3,130
5,835
3,130
2,520
3,130
3,015
Notes:D
ependentvariablelistedas
columnheader.A
llmodelsarelin
earintherunningvariable(G
PA),allowford
ifferentslopes
oneach
side
ofthethreshold,anduseabandwidthof
0.4
GPA
pointsandauniform
kernel.A
dditionalcontrolsincludeentering
age;ACTscore;indicatorsforfemale,domestic
minority,stateresident,and
medium
orhigh
financialn
eed;
and
fixedeffectsforyear
enrolled.
The
sampleisrestricted
tostudentsfirstenrollin
gin
fall2009
throughfall2017
who
graduatedby
spring
2018.Informationon
enrollm
entin
graduate
school(Colum
n10)isderivedfrom
theNationalS
tudentClearinghouse
andcoversgraduatin
gcohorts,2013–2018.Standarderrorsclusteredon
qualifying
GPA
(tothenearest100th)are
inparentheses.*p
<0.10,*
*p<0.05,*
**p<0.01.
1024
conventional levels, the pattern of results consistently points to negative impacts of Greekaffiliation on academic outcomes. We now turn to examining the impact of affiliation onlabor market outcomes.
VI. Labor Market Outcomes
A central goal for many college students is to attain a desirable andhigh-paying job upon graduation. To the extent that joining a Greek organization altersthe typical college experience, Greek membership could impact job prospects. Ourinitial finding that Greek affiliation reduces classroom performance suggests that Greekmembers may experience worse labor market outcomes than they otherwise would.However, reductions in formal academic achievement may be more than offset by otheraspects of the Greek experience. For example, prior studies note the importance ofclassmate networks in the job referral process (Marmaros and Sacerdote 2006; Zhu2019). Advocates of Greek life suggest that membership improves leadership and socialskills and provides access to professional networks that could serve to improve labormarket outcomes. In fact, using data from the General Social Social Survey, Routon andWalker (2018) find evidence that Greek affiliation is associated with a statisticallysignificant wage premium of approximately 15 percent. A potential concern with theirfinding is that unobservables that affect both earnings and the tendency to affiliate couldbe correlated, thus leading to biased estimates.In this section, we try to shed light on the question of how Greek affiliation impacts
employment status and starting salary of recent graduates using data from a university-issued post-graduation survey administered to students graduating between the 2012–2013 and 2017–2018 academic years. Graduates were contacted during the fall fol-lowing their graduation andwere asked a series of questions regarding their then-currentemployment status.Columns 1–3 of Table 9 present our RD estimates of the impact of Greek affiliation
on employment status.33 The survey response rate for the employment question was ap-proximately 70 percent among graduates not enrolled in continuing education. Column1 of Table 9 shows the estimated effects of being above the 2.5 threshold on the surveyresponse rate to the employment question. There is a marginally significant (0.10 level)increase in the probability of having employment data for students just above the 2.5threshold, perhaps indicating that Greek affiliation increases the probability that stu-dents respond regarding their employment status. Columns 2 and 3 show the estimatedimpact of affiliating with a Greek organization during second semester on employment.The estimates are negative for both employment and full-time employment, suggestingthat Greek affiliation may hinder the student with marginal GPA qualifications uponentry. These estimates should be interpreted cautiously for two reasons. First, estimatesare imprecise, and we are unable to rule out a 5 percent increase in employment atconventional levels. Second, if Greek affiliation increases the probability of respondingto the employment question, as suggested in Column 1, and the marginal respondent isless likely to be employed, results could be partially driven by selection in terms ofresponses.
33. Figure A5 in the Online Appendix displays the corresponding visuals corresponding to the RD regressionsin Table 9.
Even and Smith 1025
Tab
le9
RegressionDiscontinuityEstimates
ofGreek
Affilia
tionon
Labor
MarketOutcomes
EmploymentOutcomes
SalaryOutcomes
Variables
Employment
Survey
Response
Ratea
Employed
Employed
FullTim
e
Salary
Response
Ratea
Ln(Salary)
Ln(Salary)
Ln(Predicted
SalaryBased
onMajor)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Affiliatesecond
semester
0.0725*
-0.0761
-0.0372
0.0373
0.0879***
-0.171
-0.058
(0.038)
(0.094)
(0.111)
(0.037)
(0.009)
(0.136)
(0.056)
Boundingestim
ates
from
equalized
response
ratesb
Average
estim
ate
-0.0774
-0.040
-0.170
95%
rangeof
estim
ates
(-0.145,
-0.001)(-0.116,
0.032)
(-0.247,
–0.100)
Model
RD
RD
RD
RD
OLS
RD
RD
Observatio
ns2,585
1,723
1,323
2,585
4,760
789
2,990
Notes:D
ependentvariablelistedas
columnheader.A
llregression
discontin
uity(RD)m
odelsarelin
earintherunningvariable(G
PA),allowford
ifferentslopes
oneach
side
ofthethreshold,anduseabandwidthof
0.4GPA
pointsandauniformkernel.A
dditionalcontrolsincludeentering
age;ACTscore;indicatorsforfem
ale,domestic
minority,state
resident,and
mediumor
high
financialneed;survey
year;degreeseason
(fall,winter,spring);andentering
area
ofstudy.Colum
ns1–4consistofstudentsgraduatin
gincohorts
from
2013
–2018
who
arenotenrolledin
graduateschool
orcontinuing
undergraduateeducationandwho
respondedto
thepost-graduationsurvey.C
olum
n5and6further
restrictssampleto
only
full-tim
eworkers.C
olum
n7contains
allg
raduatingstudents.S
tandarderrorsclusteredon
GPA
(tothenearest1
00th)arein
parentheses.*p
<0.10,
**p<0.05,*
**p<0.01.
a Estim
ates
inColum
ns1and4estim
atethediscontin
uity
intheprobability
ofhaving
dataon
employment(Colum
n1)
orsalary
(Colum
n4)
atthe2.5threshold.
These
are
reduced-form
estim
ates,n
otIV
estim
ates,w
hich
would
bescaled
totheim
pliedeffectof
Greek
affiliatio
non
survey
response
rates.
bEstim
ates
aregeneratedby
random
lydropping
survey
respondentsuntil
estim
ated
discontin
uityinresponse
rateatthe2.5GPA
thresholdiselim
inated.S
eetextforadditional
detail.
1026 The Journal of Human Resources
Columns 4–7 of Table 9 present our estimates of the impact of Greek affiliation onstarting salaries, and Online Appendix Figure A5 displays the corresponding visuals.The survey response rate for the salary question was 42 percent among graduates notenrolled in continuing education. Each response corresponds to a $10,000 salary range,with top-coding for starting salaries over $90,000. To facilitate analysis, we recodedeach range at its midpoint (for example, $60,000–$69,999.99 becomes $65,000).34 Wefind no evidence of any selection in the response rate to the salary question at the 2.5threshold (Column 4 of Table 9).To compare our sample with prior studies that find a salary premium associated with
Greek affiliation, we begin by presenting ordinary least squares (OLS) estimates. Stu-dents who affiliatewith a Greek organization during second semester report salaries thatare roughly 9 percent higher, after controlling for observable heterogeneity in precol-legiate backgrounds (Column 5 of Table 9).35 This finding is similar to other cross-sectional estimates of the Greek salary premium.36
However, the Greek salary premium estimated using cross-sectional data may notreflect a causal effect, as those who join Greek organizations may be more inherentlysocial or skilled or have a family background that enhances earnings potential. Our datahave limited information about such attributes. Higher starting salaries among suchgraduatesmay simply reflect a return to unobserved characteristics or to preferences thatare correlated with a student’s decision to both join a Greek organization and pursue ahigh-paying job.To estimate a causal impact of Greek affiliation on early career salary, we use the RD
design. After correcting for the endogeneity of Greek affiliation we find a negativeimpact of affiliation on starting salaries (Column 6 of Table 9). While the estimatesare not statistically significant at conventional levels, the stark change relative to OLSestimates provides suggestive evidence that cross-sectional estimates of the Greek sal-ary premium are biased upwards. Coefficient estimates are noisy but uniformly negativeacross several alternative specifications and samples, suggesting that themarginalGreekstudent may suffer a salary penalty.37
34. For the lowest salary category of $0–$20,000, we use $14,500 as the midpoint for full-time workers as itrepresents full-time pay at the federal minimumwage, and we use $7,250 as the midpoint for part-timeworkers(20 hours per week at the federal minimum wage).35. The OLS salary premium for Greek affiliation remains present (statistically significant 7.2 percent pre-mium) when restricting the sample to the 789 observations with first-semester GPAs between 2.1 and 2.9 thatare used in the RD analysis.36. Analyzing data from Dartmouth University, Marmaros and Sacerdote (2006) find a cross-sectional wagepremium of $6,066 on a mean salary of $40,177, a 15 percent salary premium. Routon and Walker (2018) usecross-sectional (or pooled cross-sectional) data from the General Social Survey and provide evidence thatfraternity membership increases earnings for males by approximately 15 percent. Sorority membership has asimilarly sized effect, but it is statistically insignificant at the 0.10 level.Mara, Davis, and Schmidt (2018) find asalary premium of about 20 percent for fraternity members.37. Online Appendix Figure A6 documents RD estimates of the effect of Greek affiliation on starting salariesacross a variety of bandwidths, and Online Appendix Table A9 provides a series of robustness checks. Thecoefficient estimates are noisy but uniformly negative. While few estimates are statistically significant atthe 0.10 level, we are able to rule out modest salary premiums in some specifications. Using the boundariesof the 90 percent confidence intervals, a 10 percent salary premium can be ruled out in 15 of the 31 estimatesusing different bandwidths, and a 15 percent premium can be ruled out using 26 of the 31 bandwidths. Whileimprecise, this analysis is consistent with the narrative that a Greek wage premium is largely a product ofpositive omitted variable bias in the cross-section.
Even and Smith 1027
To enrich our understanding of how Greek affiliation impacts salaries, we considermajor choice. Our prior analysis suggested that Greek affiliation may reduce the prob-ability of attaining a STEM degree. Building on this finding, we investigate how Greekaffiliation impacts the average expected salary on the basis of a student’s major at grad-uation. Predicted salaries for each graduate are formed by taking the average reportedsalary for each individual with that major.38 This approach allows us to consider pre-dicted salaries even for graduates who did not report a starting salary.Column 7 of Table 9 shows the estimated impact of affiliating with a Greek organi-
zation second semester on the natural log of predicted salary for all graduates. Theestimate implies that the average Greek student who was near the 2.5 threshold in theirfirst semester would earn a statistically insignificant 6 percent less, based on their major,suggesting that major choice may be one mechanism through which Greek life influ-ences salary beyond its effect on course performance.One possible concern with our analysis of labor market outcomes is that the em-
ployment response displays a discontinuity at the grade threshold for Greek eligibility,suggesting that Greek affiliation could be related to the probability that a student re-sponds to the employment survey. This could generate a sample selection bias in ourestimates. To examine the potential consequences of this sample selection, we pursue abounding exercise motivated by Lee (2009) and adapted to our RD setting.39 In par-ticular, we randomly drop respondents on the right-hand side of the GPA eligibilitythreshold until the discontinuity in the relevant response rate (Column 1 for employ-ment, Column 4 for salary) is close to zero (that is, absolute value of less than 0.005).Wethen use this sample to estimate the treatment effect with our RDmodel. This process isrepeated 500 times to generate a distribution of treatment effects for samples withoutdifferential response rates.Results from this exercise are presented in Table 9 for the salary, employment, and
full-time employment estimates. Themeans of the 500 estimates are similar to our base-line estimates with the full sample. To understand how sensitive results are to sampleselection, we consider the range of estimates. The upper bounds of the 95 percent rangeof estimates (that is, the 97.5th percentile) for employment, full-time employment, andsalary are -0.001, 0.032, and -0.100, respectively. In other words, Greek affiliation isassociatedwith negligible or negative labormarket outcomes evenwhen considering theupper range of estimates. Consequently, we conclude that the negative labor marketeffects we find are unlikely to be caused by differences in survey response rates betweeneligible and ineligible students.It is important to emphasize that our estimates represent the impact of Greek affili-
ation on starting salaries for students who are just barely academically eligible to affil-iate. It is plausible that the impact of Greek affiliation on earnings varies across thedistribution of academic ability. For instance, if members with high initial grades aremore likely to be ushered into leadership roles within the organization, they may
38. Due to the challenge of attributing earnings power to each major for students who double-major, predictedsalaries for eachmajor are formed using the average salary among all single-major graduates within eachmajor.For double majors, the higher predicted salary of their two majors is used as their predicted salary.39. While Lee (2009) provides some of the background for this exercise, that method was not developed for anRD model and assumes the outcome variable is continuous. Carrell, Hoekstra, and Kuka (2018) provide someof the motivation for the variation of the approach we adopt here.
1028 The Journal of Human Resources
experience a greater increase in soft skills that are valued in the labormarket. Further, wefound a smaller decline in academic performance among sorority members with higherqualifyingGPAs. To the extent that there is a salary penalty formarginal Greeks, it couldbe mitigated (or potentially reversed) for stronger students. A more thorough exami-nation of these possibilities requires a richer data set including a valid instrument forGreek affiliation.
VII. Discussion and Conclusion
An estimated 750,000 undergraduate students across North Americaare currently affiliated with Greek social organizations.40 Despite this prevalence, manyof the impacts of Greek life on members remain unclear.We identify the causal effect ofGreek affiliation on academic performance using a regression discontinuity design thatcompares the subsequent academic outcomes of students who are slightly above versusslightly below the GPA eligibility cutoff imposed at the end of the first semester ofenrollment. To complement our RD approach and extend the analysis beyond studentsat the eligibility cutoff, we use an individual fixed-effects approach that exploits lon-gitudinal variation in a student’s Greek status. Intuitively, this approach compares astudent’s academic performance at different stages in their Greek life cycle to their ownpre-affiliation performance while separately controlling for the impact of accumulatedcollege experience. Both techniques reveal that Greek affiliation causes a statisticallyand economically meaningful reduction in academic performance.The impact of Greek affiliation on academic performance is heterogeneous both
across individuals and within individuals across time. We find that effects tend to belarger formales, for students with lower aptitude or higher financial need, and during thespring semesters when social activities involving alcohol are more frequent. The neg-ative effects are also greater among chapters that have more members and those thathave pledge classes with lower academic aptitude. Academic impacts extend beyondlower course grades to fewer completed credits and selection into easier courses.Contradicting the narrative that Greek affiliation improves student retention, we find nosubstantial effects of Greek affiliation on retention to sophomore year.Our results are important in light of the increasing debate surrounding Greek life on
university campuses. Proponents argue that Greek life delivers clear benefits to the localarea through community service, provides emotional support for members, and fostersthe development of leadership skills and professional networks. Critics often cite high-profile behavioral incidents involving Greek members as evidence of detrimental ef-fects. While several universities have taken action in response to low-frequency buttragic outcomes like fatalities, it is also important to understand the impact of Greekaffiliation on a broader set of individuals. This paper documents a substantial cost for theaverage member—reduced academic performance—which should be considered inconjunction with the other costs and benefits of Greek life.
40. This estimate is widely disseminated on university websites. The original source is the North-AmericanInterfraternity Conference, which currently provides an estimate for fraternity membership (384,193 as of2015–2016) but not for sorority membership.
Even and Smith 1029
It is possible that the decline in formal academic achievement is offset by gains inother areas that would be rewarded in the labor market. To investigate this, we com-plement our main analysis with estimates of the impact of Greek affiliation on thestarting salaries of graduates. While Greek affiliation is correlated with higher earnings,we provide evidence that this is not a causal effect. Using the regression discontinuityapproach, we can rule out the large positive effects on starting salaries found in cross-sectional studies. This finding suggests that Greek affiliation does not provide a sub-stantial labor market payoff for students with marginal academic qualifications at thetime of their Greek affiliation.While our results highlight some of the negative aspects of Greek affiliation, the
abundance of students joining fraternities and sororities suggests that there are benefitsas well. An obvious question for university administrators is what can be done tominimize these negative effects without an outright ban on such organizations. Ourresults imply that interventions targeting the pledging process may offer the greatestscope for impact. The largest decline in academic performance occurs during thepledging semester, consistent with the increased burden of learning the organization’shistory and anecdotal evidence of hazing. Reforms to the pledging process could alsobenefit active members, as we find that they experience larger declines in academicperformance during the spring when the social calendars of Greek organizations arefilled with rush, new member education, and additional social events associated withinitiation. Finally, while the institution studied here has a minimumGPA requirement tobecome a Greek member, once a student is admitted there is no maintenance GPA re-quirement imposed by the university. In terms of external validity, if maintenance GPApolicies mitigate the decline in academic performance among Greeks, then the declineobserved in our setting may be larger than at other universities.While much of the recent attention to Greek life has resulted from tragic incidents
associatedwith hazing, some of the resulting policy changes could alter the Greek effecton academic performance. Some attempted changes include reforming the pledgingprocess and establishing alcohol-free houses. One national fraternity, Sigma AlphaEpsilon, abolished pledging in 2014 and requires that potential members be immedi-ately initiated in an effort to eliminate hazing. In 2000, Phi Delta Theta implemented analcohol-free housing policy, and North Carolina State University implemented a tem-porary ban on alcohol at fraternity events in 2015. Little is known about the efficacy ofsuch policies. Future research on the topic would benefit universities and Greek orga-nizations aiming to enhance the academic performance of students.
References
Angrist, Joshua, Daniel Lang, and Philip Oreopoulos. 2009. “Incentives and Services for CollegeAchievement: Evidence from a Randomized Trial.” American Economic Journal: AppliedEconomics 1(1):136–63.
Barreca, Alan I., Jason M. Lindo, and Glen R. Waddell. 2016. “Heaping-Induced Bias inRegression-Discontinuity Designs.” Economic Inquiry 54(1):268–93.
Borsari, Brian, and Kate B. Carey. 2001. “Peer Influences on College Drinking: A Review ofthe Research.” Journal of Substance Abuse 13(4):391–424.
1030 The Journal of Human Resources
Calonico, Sebastian, Matias D. Cattaneo, and Rocio Titiunik. 2014. “Robust NonparametricConfidence Intervals for Regression Discontinuity Designs.” Econometrica 82(6):2295–326.
Carrell, Scott E., Richard L. Fullerton, and James E. West. 2009. “Does Your Cohort Matter?Measuring Peer Effects in College Achievement.” Journal of Labor Economics 27(3):439–64.
Carrell, Scott E., Mark Hoekstra, and Elira Kuka. 2018. “The Long-Run Effects of DisruptivePeers.” American Economic Review 108(11):3377–415.
Carrell, Scott E., Mark Hoekstra, and James E. West. 2011. “Does Drinking Impair CollegePerformance? Evidence from a Regression Discontinuity Approach.” Journal of PublicEconomics 95(1–2):54–62.
Carrell, Scott, and Bruce Sacerdote. 2017. “Why Do College-Going Interventions Work?”American Economic Journal: Applied Economics 9(3):124–51.
De Donato, Andrew, and James Thomas. 2017. “The Effects of Greek Affiliation on AcademicPerformance.” Economics of Education Review 57:41–51.
DeSimone, Jeff. 2007. “Fraternity Membership and Binge Drinking.” Journal of Health Eco-nomics 26(5):950–67.
———. 2009. “FraternityMembership andDrinking Behavior.”Economic Inquiry 47(2):337–50.Franke, Ray, Sylvia Ruiz, Jessica Sharkness, Linda DeAngelo, and John Pryor. 2010.“Findings from the 2009 Administration of the College Senior Survey (CSS): NationalAggregates.” Los Angeles, CA: Higher Education Research Institute at the University ofCalifornia Los Angeles.
Goldin, Claudia, Lawrence F. Katz, and Ilyana Kuziemko. 2006. “The Homecoming of AmericanCollege Women: The Reversal of the College Gender Gap.” Journal of Economic Perspectives20(4):133–56.
Ha, Joung Yeob, and Austin C. Smith. 2019. “Legal Access to Alcohol and Academic Perfor-mance: Who Is Affected?” Economics of Education Review 72:19–22.
Hahn, Jinyong, Petra Todd, and Wilbert Van der Klaauw. 2001. “Identification and Estimation ofTreatment Effects with a Regression-Discontinuity Design.” Econometrica 69(1):201–9.
Hausman, Catherine, and David S. Rapson. 2018. “Regression Discontinuity in Time: Con-siderations for Empirical Applications.” Annual Review of Resource Economics 10:533–52.
Imbens, Guido W., and Thomas Lemieux. 2008. “Regression Discontinuity Designs: A Guide toPractice.” Journal of Econometrics 142(2):615–35.
Lee, David S. 2009. “Training, Wages, and Sample Selection: Estimating Sharp Bounds onTreatment Effects.” Review of Economic Studies 76(3):1071–102.
Lee, David S., and David Card. 2008. “Regression Discontinuity Inference with SpecificationError.” Journal of Econometrics 142(2):655–74.
Lindo, Jason M., Nicholas J. Sanders, and Philip Oreopoulos. 2010. “Ability, Gender, and Per-formance Standards: Evidence from Academic Probation.” American Economic Journal:Applied Economics 2(2):95–117.
Lindo, JasonM., Isaac D. Swensen, and Glen R.Waddell. 2012. “Are Big-Time Sports a Threat toStudent Achievement?” American Economic Journal: Applied Economics 4(4):254–74.
———. 2013. “Alcohol and Student Performance: Estimating the Effect of Legal Access.”Journal of Health Economics 32(1):22–32.
Maloney, Michael T., and Robert E. McCormick. 1993. “An Examination of the Role that Inter-collegiate Athletic Participation Plays in Academic Achievement: Athletes’ Feats in theClassroom.” Journal of Human Resources 28:555–70.
Mara, Jack, Lewis Davis, and Stephen Schmidt. 2018. “Social Animal House: The Economicand Academic Consequences of Fraternity Membership.” Contemporary Economic Policy36(2):263–76.
Marmaros, David, and Bruce Sacerdote. 2006. “How Do Friendships Form?” Quarterly Journalof Economics 121(1):79–119.
Even and Smith 1031
McCrary, Justin. 2008. “Manipulation of the Running Variable in the Regression DiscontinuityDesign: A Density Test.” Journal of Econometrics 142(2):698–714.
Routon, P. Wesley, and Jay K. Walker. 2014. “The Impact of Greek Organization Membershipon Collegiate Outcomes: Evidence from a National Survey.” Journal of Behavioral andExperimental Economics 49:63–70.
Routon, P. Wesley, and Jay K. Walker. 2018. “Beyond the Ivory Tower: Labor Market Returnsto Greek Membership.” LABOUR 32(2):182–204.
Sacerdote, Bruce. 2001. “Peer Effects with Random Assignment: Results for DartmouthRoommates.” Quarterly Journal of Economics 116(2):681–704.
U.S. Department of Education. 2018. “Digest of Education Statistics: 2017.” Washington, DC:National Center for Education Statistics, Institute of Education Sciences.
Walker, Jay K., Nathan D. Martin, and Andrew Hussey. 2015. “Greek Organization Membershipand Collegiate Outcomes at an Elite, Private University.” Research in Higher Education56(3):203–27.
Zhu, Maria. 2019. “Job Networks through College Classmates: Effects of Referrals for Menand Women.” Unpublished.
1032 The Journal of Human Resources
The Relationship Between Greek Exportsand Foreign Income
By Konstantinos Chisiridis* and Theodore Panagiotidis **
Abstract
This paper assesses the effect of foreign economic activity on Greek exports. We employdata from 1995:I to 2016:IV and quantify the long-run foreign income elasticity of Greekexports. We establish a cointegration relationship and employ Dynamic OLS estimations. Wefind that the aggregate foreign income elasticity of Greek exports is positive and significant.When foreign income is decomposed to the main trading partners, we reveal that economicgrowth in Turkey and in emerging markets such as the Balkans, North Africa and the MiddleEast have the greatest impact on Greek exports. The impact of the traditional European tradingpartners of Greece (Germany and Italy) are found to be positive but insignificant. Finally, thedynamic analysis shows a positive interaction between real income growth in Turkey andGreek export growth at the short-run horizon.
JEL Classification: C22, F43
Keywords: Greece, exports, foreign income, cointegration
“It is your concern when your neighbor’s wall is on fire.”
Horace
1. Introduction
The cumulative deficits of the Greek current account have been cited as one ofthe factors that contributed to the recent debt crisis (see Gros, 2011 for instance).After joining the Eurozone, the Greek economy could not anymore use currency de-valuation as a tool of consolidation of the current account deficits. Despite the slug-gish adjustment, the Greek current account recorded surplus during the third quarterof 2012 for the first time in 18 years. A critical question is whether exports or im-ports of goods and services were the driving force behind this adjustment. A reportof the National Bank of Greece in July 2013 (National Bank of Greece, 2013), ar-
Applied Economics Quarterly Vol. 64. No 1 (2018)Duncker & Humblot GmbH, 12165 Berlin
Applied Economics Quarterly 64 (2018) 1
* Department of Economics, University of Macedonia, email: k.chisiridis@gmail.com.
** Corresponding author: Theodore Panagiotidis, Department of Economics, University ofMacedonia, email: tpanag@uom.edu.gr.
gues that this adjustment in current account stems from imports that contracted (theperiod 2009–2012) by 33% in value terms. Exports have increased by 20%.
It has been argued that trade openness and export expansion can act as a catalystto stimulate growth recovery (Riedel, 1984 among others). Promotion of exports islinked with increased productivity and positive externalities, the so-called Export-Led-Growth-Hypothesis (ELGH; Emery, 1967, Balassa, 1978, Sharma & Panagioti-dis 2005). Also, an export-oriented economy can benefit from economies of scale(Helpman & Krugman 1985), increased aggregate demand and better allocation oftotal investment. Export growth ensures a strong balance of payments which pro-vides support to imports of intermediate and capital goods.
Another strand of the literature focuses on the impact of trade openness and im-ports on economic growth. Grossman & Helpman, 1991 point out that the adoptionof trade expansion policies can strengthen knowledge and innovation spillovers in asmall open economy framework. Mcnab & Moore, 1998 reiterate the latter for de-veloping countries as they relate strongly outward oriented policies with a 1.5% an-nual increase in GDP growth. In this context, imports of intermediate goods exhibitgrowth inducing effects as well, known as the import-led-growth hypothesis(ILGH). The studies of Awokuse, 2007, Awokuse 2008 and Thangavelu & Raja-guru 2004 provide evidence for the importance of imports in the growth process.Chortareas et al. 2015 show that openness is more important for developing coun-tries.
Empirical modeling of the mechanisms that define international trade relationsamong countries helps to understand the evolution of trade deficits (Crane et al.2007). Economic theory points to three factors that can determine the foreign de-mand for domestic goods and services: (i) foreign income, (ii) prices of domesticexport goods and services and (iii) prices of goods and services that compete withthe domestic ones in the global markets.
It is important though, to grasp how changes in foreign demand can affect exportgrowth. For this purpose, part of the literature focuses on the computation of the for-eign income and price elasticity of exports for developed and developing countries(Marquez & McNeilly, 1988, Marquez, 1990, Senhadji & Montenegro, 1999).These foreign income elasticities of exports, quantify the simultaneous relationshipbetween foreign economic growth and domestic exports. Higher GDP growth in thetrade partners of Greece may lead to higher demand for Greek exports of goods andservices. Thus, economic prosperity in foreign regions can be associated with high-er Greek export growth.
This paper estimates the long-run aggregate foreign income elasticity of Greekexports of goods and the long-run disaggregated income elasticity of Greek exportsof goods per trading partner along with the price elasticity. The adopted approachallows us to quantify the impact of a change in the foreign real income on Greek ex-ports separately for each trading partner. Nie & Taylor, 2013, follow the same pro-cedure to study the region-specific income effects for the U.S. exports. We employ
100 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
a Vector Autoregressive (VAR) model to gauge the dynamic inter-linkages betweenGreek exports, regional foreign growth and the real effective exchange rate ofGreece. The main questions that we address are:
1. What was the evolution of shares and destinations of Greek export goodsper region over the last twenty years (1996–2016)?
2. How changes in real GDP per region affect Greek exports?
3. Which is the dynamic response of Greek exports in positive shocks on for-eign disaggregated income and the exchange rate in the short-run and long-run?
The model includes real income activity in the main trading partners of Greeceover the last five years; Germany, Italy, Turkey and the rest of Europe. These re-gions account, on average, for the 75% of the Greek export goods. We employquarterly data from 1995:I–2016:IV. To investigate the long-run relationshipamong our variables, we conduct a cointegration test (we refer to this as the staticanalysis). The estimations of the foreign real income and price elasticities of Greekexports are based on Dynamic OLS (DOLS) procedure as proposed by Saikkonen,1992 and Stock & Watson, 1993. We reveal that economic growth in Turkey andemerging markets such as the Balkans and the Middle East and North Africa ac-count for the majority of the growth of Greek exports. The price elasticity of Greekexports is found to be negative (in line with expectations) but insignificant.
In the dynamic analysis (VAR), we examine the effect of a positive shock in theforeign real income growth on Greek export growth. We employ both generalizedimpulse response functions and local projections. The results indicate again the im-portance of the Turkish market for Greek exports. Also, a real depreciation of theGreek economy can boost export growth in the short-run (for 4 quarters).
The outline of the paper is structured as follows: Section 2 provides an overview ofthe trade partners of Greece for the last twenty years and Section 3 discusses the meth-odology. The empirical evidence provided in Section 4 and the last one concludes.
2. The Composition of Greek Export Goods per Region
An open economy benefits from trade with both developed and fast-growingeconomies (Arora & Vamvakidis, 2005). The Greek economy is influenced by eco-nomic conditions abroad with global shocks transmitted into the domestic economythrough the trading partners’ channel. An investigation of the recent evolution ofthe Greek trade partners can offer useful insights for economic policy analysis. TheGreek economy is one of the most closed economies in the EU in terms of tradeopenness. Böwer et al., 2014 attribute the latter to weak institutional quality. Inter-nal demand (rather than exports) and the housing market (rather than manufactur-ing) are more influential in the Greek case (Panagiotidis & Printzis, 2016).
The Relationship Between Greek Exports and Foreign Income 101
Applied Economics Quarterly 64 (2018) 1
Over the last twenty years, 75% of the Greek exports was absorbed by Europeanmarkets. The Middle East and North Africa (MENA) hold, an average, 8% of theGreek export goods followed by the USA with 5% and East Asia with 4%. Growthin these regions affects the growth rate of Greek exports of goods and services. Fig-ure 1 presents the region and country destination of Greek export goods from1996:II till 2016:II. The most important trade partners for Greece were (i) Germanywith 11% of Greek export goods and (ii) Italy with 10%; UK, Turkey, Bulgaria andthe USA follow.
For the period 2012–2016 Germany, Italy and Turkey emerge as the most impor-tant markets for Greece. These countries absorbed, on average, 26% of Greek ex-port goods.
Notes: Rest of Europe includes Austria, Belarus, Belgium, Czech Republic, Denmark, Estonia, Finland,Hungary, Iceland, Ireland, Latvia, Lithuania, Malta, Netherlands, Norway, Poland, Portugal, Russia, SlovakRepublic, Spain, Sweden, Switzerland, Ukraine. Balkans includes Albania, Cyprus, FYROM, Romania,Serbia, Slovenia. MENA includes Algeria, Kingdom of Bahrain, Egypt, Islamic Republic of Iran, Iraq,Israel, Jordan, Kuwait, Lebanon, Libya, Morocco, Saudi Arabia, Tunisia, United Arab Emirates. East Asiaincludes China, Hong Kong, Indonesia, Japan, Korea, Singapore. Rest of the world includes Argentina,Australia, Brazil, Canada, Colombia, Mexico, South Africa.
Source: IMF Direction of Trade Statistics and authors’ calculations.
Figure 1: Greek Export Goods Share per Region
The ranking of Greece’s trade partners has changed during the last seven years ofeconomic crisis. Important destinations such as Europe and the Balkans lost part oftheir share and more distant areas such as the USA and East Asia have emerged.Moreover, we note a notable rise of the share of the region of the Middle East andNorth Africa that nearly doubled its share to 14% in the period 2012 –2016. Figure 2summarizes the evolution of the geographical destination of Greek export goods.
102 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
Notes: Europe includes Austria, Belarus, Belgium, Czech Republic, Denmark, Estonia, Finland, France,Germany, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Malta, Netherlands, Norway, Poland, Portugal,Russia, Slovak Republic, Spain, Sweden, Switzerland, UK, Ukraine. Southeastern Europe includes Albania,Bosnia and Herzegovina, Bulgaria, Croatia, Cyprus, FYROM, Republic of Kosovo, Romania, Serbia, Slo-venia, Turkey. MENA includes Algeria, Kingdom of Bahrain, Egypt, Islamic Republic of Iran, Iraq, Israel,Jordan, Kuwait, Lebanon, Libya, Morocco, Saudi Arabia, Tunisia, United Arab Emirates. East Asia in-cludes China, Hong Kong, Indonesia, Japan, Korea, Singapore. Rest of the World includes Argentina, Aus-tralia, Brazil, Canada, Colombia, Mexico, South Africa.
Source: IMF Direction of Trade Statistics and authors’ calculations.
Figure 2: Time-Varying Share of Greek Export Goods per Region1996:II–2016:II
This geographical reorientation of the Greek exports can be attributed to the weakeconomic growth of the western and the southeastern Europe during the last sixyears. From 2008 to 2016, the Euro area had an anemic GDP growth rate of 0.6%as a result of the global financial crisis. Southeastern Europe failed to recover to thepre-crisis level. This has direct implications for the structure of Greek exports ofgoods. The weak demand from these countries forced Greek exporters to find othermarkets. Turkey showed a GDP growth rate of 3.4%, the USA grew on average1.5% and East Asia recorded a high GDP growth rate of 8%. Trade with Italy andGermany has been traditionally the most important for Greece and this has been re-mained unchanged although, exports to the German market did not increase.
3. Methodology
This section discusses the methodology employed in order to uncover the link be-tween economic growth in foreign economies and Greek exports. We provide aframework to gauge both the short run and long-run relationship between foreign
The Relationship Between Greek Exports and Foreign Income 103
Applied Economics Quarterly 64 (2018) 1
real GDP and real Greek exports (see Goldstein & Khan, 1985 for a detailed de-scription on specification issues of trade equations).
The first model examines the contemporaneous, long-run relationship betweenthe real Greek exports, real foreign income, and real effective exchange rate. As-suming that domestic and foreign tradable goods are imperfect substitutes, this spe-cification is close to the standard export demand function and can be written as:
logðextÞ ¼ a0 þ �1t þ �logðytÞ þ logðreertÞ þ a2dt þ utð1Þ
where ext is real merchandise exports of Greece; yt is the real foreign incomevariable approximated by the difference between real GDP and real exports ofgoods and services of the main Greek trading partners; reert is the real effectiveexchange rate of Greece and represents the relative price between exported Greekgoods and those that compete with them in the global markets. Parameter � can beinterpreted as the foreign income elasticity of Greek exports and is a proxy ofcompetitiveness of the Greek economy. Senhadji & Montenegro, 1999, employ thesame specification with yt to account for proxying for trading partners’ income.Also, we incorporate in (1) a dummy variable dt to capture shifts in the constantterm due to structural breaks.1 Previous studies have employed cointegrationanalysis for (1) (see for instance Caporale & Chui, 1999, Hooper et al., 2000).
The linkage between foreign income and Greek exports through the aggregate in-come elasticity is of importance. Nevertheless, it would also be interesting to exam-ine how Greek exports react to income changes in specific regions. Therefore, weestimate the region disaggregated foreign income elasticities for the Greek exports.We include in our analysis the three most important trade partners of Greece overthe last five years as defined in section 2; Germany, Italy, and Turkey. We quantifyincome in the remaining trading partners of Greece through the variable yRoW
t . Thus,the augmented model can be written as:
logðextÞ ¼ a0 þ �1t þ �1ilogðyi; tÞ þ �1yRoWt þ logðreertÞ þ a1dt þ utð2Þ
where i ¼ 1; 2; 3 for the case of the real income activity (real GDP minus realexports) in Germany, Italy, and Turkey respectively. Parameters �1i represent theregion-specific real income elasticities of the Greek exports while �1 is the elasticityof the Greek exports with respect to income in the rest of the world. This approachgauges the effect of a change in the foreign income on Greek goods separately foreach region. If all variables of equation (1) and equation (2) are I(1) and there is acointegration relationship among them, estimation can be conducted using DynamicOLS (DOLS) with the fixed specification of leads and lags (leads = 1, lags = 1 given
104 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
1 Dummy variable dt takes the value 0 before 2002Q1 and 1 afterward. The break datestems from the Bai & Perron, 1998 breakpoints test.
that the dataset is limited). Table 2 summarizes the estimation results of equation (1)and equation (2) based on the underlying method.
3.1 The Vector Autoregressive Model
A VAR model can capture linear interdependencies between real Greek exportgoods growth, Greek trade partners real income activity growth and exchange rategrowth. To quantify short-run dynamics, we utilize a reduced form VAR(p) modelrepresented by:
xt ¼ v þ A1xt�1 þ . . .þ Apxt�p þ ut; t ¼ 1; 2; . . . ; T
where xt ¼ ðx1t; . . . ; xktÞ0
is a ðk � 1Þ vector of the variables, Ai is a k � kcoefficient matrices for i ¼ ð1; . . . ; pÞ, v ¼ ðv1; . . . ; vkÞ
0is an intercept term vector
and ut ¼ ðu1t; . . . ; uktÞ0
is a k-dimensional zero-mean white noise process withcovariance matrix EðututÞ
0 ¼ �u. The underlying model consists of six endogenousvariables with xt being a 6 � 1 vector, defined as:
xt ¼
Greek export goods growtht
German real income growtht
Italian real income growtht
Turkish real income growtht
RoW real income growtht
reer growtht
0BBBBBB@
1CCCCCCA
We will investigate the dynamic properties of the VAR(p) model using the Gener-alized Impulse Response Functions (GIRFs) developed by Pesaran & Shin, 1998.GIRFs are invariant of the ordering of the variables. This will allow us to trace thetotal effect on Greek exports from a one-standard-deviation positive shock to for-eign real GDP. The Local Projections (LPs) (different to the GIRFs) were proposedby [Jordà, 2005, Jordà, 2009] and are also produced. The main advantage of thisprocedure is that LPs rely on their own IRF regression instead of previous iterationsof the model leading them to be less vulnerable to misspecification. Also, LPs bene-fit from the simplicity of their estimation (they rely on OLS with robust standard er-rors).2
The Relationship Between Greek Exports and Foreign Income 105
Applied Economics Quarterly 64 (2018) 1
2 A cautioning note should be raised here as the two approaches are not alternative; GIRFscalculate the response of a typical shock whereas the central idea of LPs consists of estimatinglocal projections at each period of interest instead of extrapolation into increasing horizons(VAR).
4. Empirical Results
4.1 Data and Unit Root Tests
We employ quarterly seasonally adjusted data for the period 1995:I –2016:IV(88 observations). The variable yt is created by the weighted average of the thirtymajor trading partners of Greece during the last twenty years. Also, real income ac-tivity for Germany, Italy, and Turkey and the rest of the world (RoW) was createdusing data for real GDP and real exports of goods and services. Region RoW con-sists of the remaining trading partners of Greece; exceptions include Germany, Italy,and Turkey whose growth impact is already incorporated into the analysis. Greekmerchandise exports were deflated using the unit value of Greek export goods.
The Augmented Dickey-Fuller (ADF) test and the modified DF test proposed byElliott et al., 1992 (DF-GLS test) are used for all the variables of equation (1) andequation (2). The unit root tests were conducted both in the levels and first differ-ences of the variables and the lag length of the test was chosen based on AIC. Theresults indicate that all variables in the analysis are I(1). Table 3 in the Appendixsummarizes the results.
4.2 Cointegration Analysis
To investigate whether there is a cointegrating relationship between the variablesin equation (1) and equation (2), we adopt the single-equation cointegration test pro-posed by Phillips & Ouliaris, 1990. This procedure implies a test for the presence ofa unit root in the residuals of the cointegrating equation. To account for a structuralbreak of unknown time in the cointegrating relation of the equation (1) and equation(2) we employ the residual based cointegration test of Gregory & Hansen, 1996a,Gregory & Hansen, 1996b.
A cointegration relationship suggests that there is a long-run equilibrium amongreal Greek exports (extÞ, real foreign income (total yt and region disaggregated yi; t,yRoW
t ) and the proxy of the relative price of Greek exports (reert). Table 1 sum-marizes the cointegration tests statistics for equation (1) and equation (2).
The results from both tests indicate the presence of a cointegrating relationshipamong the variables of equation (1) and equation (2) as the null hypothesis of no co-integration is rejected in all cases at the 5% level of significance. Hence, the long-run foreign income elasticities of the Greek exports can be quantified using DOLS.
4.3 Income and Price Elasticities for Greek Exports
Estimation of equation (1) and equation (2) quantify the foreign income (aggre-gated and disaggregated) and the price elasticity of Greek exports. If the link be-tween the growth of Greek exports and economic growth in its trade partners is sig-
106 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
nificant, this will be depicted by a high level of foreign income elasticity. The ex-pected relationship between foreign income and Greek exports is positive. Also, areal depreciation of the Greek economy can enhance the domestic productivity lead-ing to an increased market share of Greek export goods in the global market. There-fore, the literature suggests that the price elasticity of Greek exports should be nega-tive (see Stern, 1976 for a thorough analysis of price elasticities of exports).
Table 1
Cointegration Analysis of Greek Exportsand Foreign Regional Income
Phillips–Ouliaris Cointegration Test
tau-statistic z-statistic p-value
Equation 1 –4.96 –38.36 0.00
Equation 2 –7.43 –67.98 0.00
Note: p-values from MacKinnon, 1996.
Gregory-Hansen Cointegration Test
5% Critical Value Break Date
Equation 1
C
ADF t-stat. –5.69 –4.92 2001Q3
Za-stat. –47.69 –46.98 2001Q3
Zt-stat. –5.72 –4.92 2001Q3
C/T
ADF t-stat. –6.45 –5.29 2001Q3
Za-stat. –57.51 –53.92 2001Q3
Zt-stat. –6.49 –5.29 2001Q3
Equation 2
C
ADF t–stat. –6.94 –4.98* 2001Q4
Za-stat. –67.73 2001Q3
Zt-stat. –7.37 2001Q3
C/T
ADF t-stat. –6.91 –5.23* 2001Q4
Za-stat. –70.87 2007Q1
Zt-stat. –7.70 2007Q1
Note: Critical values from Gregory & Hansen, 1996b. (*) denotes ADF t-statistic critical values basedon Mackinnon, 2010 due to the fact that, in the Gregory & Hansen, 1996b paper only critical values of upto 4 variables are available.
The DOLS estimation results of equation (1) (Table 2, Column 1) indicate thatthe long-run aggregate foreign income elasticity of Greek exports is above unity(2.59) and statistically significant at the 1% level. This reveals that Greek exportsare very sensitive to changes in foreign income conditions (Tourism could be oneexplanation. Another is the special role of oil in the Greek trade balance as we willexplain later). A 1% increase in the real income of the Greek trading partners is as-sociated with an approximately 2.5% simultaneous increase in the Greek exports.
The Relationship Between Greek Exports and Foreign Income 107
Applied Economics Quarterly 64 (2018) 1
The long-run price elasticity of Greek exports is found positive (0.5) and statisti-cally significant at the 1% level of significance.3 A possible explanation for thisstems from the special role of oil. The latter is a dominant commodity for bothGreek imports and exports. In 2016, refined oil accounted for 21% of the Greek ex-ports of goods while imports of goods were mainly consisted by crude oil (13%)and refined oil (5%).4
Table 2
The Long-Run Elasticities Between Greek Exportsand Foreign Regional Real Income
Dependent Variable:real merchandise Greek exports ext Equation (1) Equation (2)
DOLS DOLS
yt 2.59���
[5.09]
reert 0.50��� –0.01
[2.76] [–0.02]
C –30.68��� –30.77
[–4.14] [–1.31]
trend 0.01��� –0.01
[3.03] [–0.74]
German real incomet 0.37
[0.37]
Italian real incomet 0.59
[1.57]
Turkish real incomet 0.87���
[3.20]
RoW real incomet 1.00��
[2.22]
dt –0.18��� –0.14��
[–5.86] [–2.31]
Observations: 88 88
R-squared: 0.94 0.96
Time Period 1995:Q1–2016:Q4 1995:Q1–2016:Q4
Note: * Significant at the 10 % level. ** Significant at the 5 % level. *** Significant at the 1 % level.
108 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
3 When the foreign income is decomposed the positive coefficient becomes negative andinsignificant.
4 Data source: UN COMTRADE.
The disaggregated approach in equation (2) (Table 2, Column 2) highlights thatdifferent regions affect distinctly Greek exports. Economic growth in all four re-gions under consideration (rest of the world, Germany, Italy, and Turkey) impactspositively on Greek exports. The rest of the world has the greatest impact on Greekexports as a 1% increase in the region’s real income is associated with a 1% in-crease of real Greek exports which is also statistically significant. In addition,growth in the Turkish economy is accompanied with a positive (0.87) and statisti-cally significant reaction of the Greek exports. Thus, the strong linkage of Greek ex-porting firms and the Turkish market of goods can be confirmed, especially duringthe years of the financial crisis.
4.4 Dynamic Analysis
We employ a VAR(4) model to study the dynamic relationship between Greekexports and real income growth in the trading partners. The number of lags structurewas decided based on the AIC and the likelihood ratio test. Figure 3 summarizesthe GIRFs of a positive one standard deviation shock on regional real income andexchange rate to Greek export goods growth. The results indicate a positive re-sponse of Greek exports to real income shocks emanating from Turkey and the restof the world. Evidence from the local projections of GIRFs (Figure 4) shows thatGreek exports are affected positively by Turkish growth which is statistically sig-nificant only for one quarter after the shock. In addition, a depreciation of the Greekreal exchange rate, corresponding to a positive shock on the real exchange rate, isassociated with a decrease in Greek real exports.
5. Conclusion
The relationship between foreign income and exports growth is well establishedin the literature. This paper examines the relation between Greek exports and thereal income in the major trading partners of Greece. We quantify the static long-runand dynamic relationships between Greek exports and the real income in various re-gions of interest.The evolution of the Greek trade partners as a share of the totalGreek exports goods was examined.
Greek exports have shifted over the last seven years from the European marketsto emerging markets such as the Balkans, Middle East, and East Asia. Regardingthe results from the region disaggregated foreign income elasticity of the Greek ex-ports, we find that economic growth in these regions and Turkey affects the mostdemand for Greek export goods. This is depicted in the strong linkage between thedemand for Greek goods in these regions and the growth of Greek exports duringthe years of financial crisis as Greek exports abandoned traditional markets such asGermany and Italy. These regions account for the bulk of the relationship betweenGreek real exports and foreign real income.
The Relationship Between Greek Exports and Foreign Income 109
Applied Economics Quarterly 64 (2018) 1
110 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
(a)
Gen
eral
ized
res
pons
e of
Gre
ek e
xpor
ts
to a
one
s.d
. sho
ck o
n G
erm
an r
eal i
ncom
e gr
owth
.(c
) G
ener
aliz
ed r
espo
nse
of G
reek
exp
orts
to a
on
e s.
d. s
hock
on
Tur
kish
rea
l inc
ome
grow
th.
(b)
Gen
eral
ized
res
pons
e of
Gre
ek e
xpor
ts
to a
one
s.d
. sho
ck o
n It
alia
n re
al in
com
e gr
owth
.
(d)
Gen
eral
ized
res
pons
e of
Gre
ek e
xpor
ts to
a
one
s.d.
sho
ck o
n th
e R
oE r
eal i
ncom
e gr
owth
.(e
) G
ener
aliz
ed r
espo
nse
of G
reek
exp
orts
to a
on
e s.
d. s
hock
on
reer
gro
wth
.
Not
e:S
hade
dar
eas
repr
esen
tthe
�2
s.e.
conf
iden
ceba
nds.
Fig
ure
3:G
ener
aliz
edR
espo
nse
ofG
reek
Exp
ortG
row
thto
aon
es.
d.S
hock
onF
orei
gnR
ealI
ncom
eG
row
th
The Relationship Between Greek Exports and Foreign Income 111
Applied Economics Quarterly 64 (2018) 1
(a)
LP
of th
e ge
nera
lize
d re
spon
se o
f G
reek
exp
orts
to
a o
ne s
.d. s
hock
on
Ger
man
rea
l inc
ome
grow
th.
(b)
LP
of th
e ge
nera
lize
d re
spon
se o
f G
reek
exp
orts
to
a o
ne s
.d. s
hock
on
Ital
ian
real
inco
me
grow
th.
(c)
LP
of th
e ge
nera
lize
d re
spon
se o
f G
reek
exp
orts
to a
on
e s.
d. s
hock
on
Tur
kish
rea
l inc
ome
grow
th
(d)
LP
of th
e ge
nera
lize
d re
spon
se o
f G
reek
exp
orts
to
a o
ne s
.d. s
hock
on
the
RoE
rea
l inc
ome
grow
th.
(e)
LP
of th
e ge
nera
lize
d re
spon
se o
f G
reek
exp
orts
to
a o
ne s
.d. s
hock
on
reer
gro
wth
.
Not
e:S
hade
dar
eas
repr
esen
tthe
95%
conf
iden
ceba
nds.
Fig
ure
4:L
ocal
Pro
ject
ions
ofth
eG
ener
aliz
edR
espo
nse
ofG
reek
Exp
ortG
row
thto
aon
es.
d.S
hock
onF
orei
gnR
ealI
ncom
eG
row
th
Furthermore, results from a dynamic analysis tend to confirm the importance ofeconomic growth in these regions for the Greek exports. Moreover, the high elas-ticity of the foreign real income of the Greek exports can signal that an export-led-growth policy can be beneficial for the recovery of the country.
References
Arora, V. / Vamvakidis, A. (2005): “How much do trading partners matter for economicgrowth?” IMF Staff Papers, 52 (1), 24–40.
Awokuse, T. O. (2007): “Causality between exports, imports, and economic growth: Evidencefrom transition economies.” Economics Letters, 94 (3), 389–395.
– (2008): “Trade openness and economic growth: is growth export-led or import-led?” Ap-plied Economics, 40 (2), 161–173.
Bai, J. /Perron, P. (1998): “Estimating and testing linear models with multiple structuralchanges.” Econometrica, (pp. 47–78).
Balassa, B. (1978): “Exports and economic growth: Further evidence.” Journal of Develop-ment Economics, 5 (2), 181–189.
Böwer, U. / Michou, V. / Ungerer, C., et al. (2014): “The puzzle of the missing greek exports.”Directorate General Economic and Financial Affairs (DG ECFIN), European Commission.
Caporale, G. M. / Chui, M. K. (1999): “Estimating income and price elasticities of trade in acointegration framework.” Review of International Economics, 7 (2), 254–264.
Chortareas, G. / Magkonis, G. / Moschos, D. / Panagiotidis, T. (2015): “Financial developmentand economic activity in advanced and developing open economies: Evidence from panelcointegration.” Review of Development Economics, 19 (1), 163–177.
Crane, L. / Crowley, M. /Quayyum, S. (2007): “Understanding the evolution of trade deficits:Trade elasticities of industrialized countries.” Economic Perspectives-Federal Reserve Bankof Chicago.
Elliott, G. / Rothenberg, T. J. / Stock, J. H. (1992): “Efficient tests for an autoregressive unitroot.” National Bureau of Economic Research.
Emery, R. F. (1967): “The relation of exports and economic growth.” Kyklos, 20 (4), 470–486.
Goldstein, M. / Khan, M. S. (1985): “Income and price effects in foreign trade.” In: Trade Cur-rencies and Finance (pp. 3–81). World Scientiffic.
Gregory, A. W. / Hansen, B. E. (1996a): “Practitioners corner: tests for cointegration in modelswith regime and trend shifts.” Oxford bulletin of Economics and Statistics, 58 (3), 555 –560.
– (1996b): “Residual-based tests for cointegration in models with regime shifts.” Journal ofEconometrics, 70 (1), 99–126.
Gros, D. (2011): “External versus domestic debt in the euro crisis.” CEPS Policy Brief,(no 234).
Grossman, G. M. / Helpman, E. (1991): “Trade, Knowledge Spillovers, and Growth.” Euro-pean Economic Review, 35 (2–3), 517–526.
112 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
Helpman, E. / Krugman, P. R. (1985): “Market structure and foreign trade: Increasing returns,imperfect competition, and the international economy.” MIT Press.
Hooper, P. /Johnson, K. / Marquez, J. R. (2000): “Trade elasticities for the g-7 countries.” Prin-ceton Studies in International Economics, (87).
Jordà, Ò. (2005): “Estimation and inference of impulse responses by local projections.”American Economic Review, 95 (1), 161–182.
– (2009): “Simultaneous condence regions for impulse responses.” The Review of Economicsand Statistics, 91 (3), 629–647.
MacKinnon, J. G. (1996): “Numerical distribution functions for unit root and cointegrationtests.” Journal of Applied Econometrics, (pp. 601–618).
– (2010). “Critical values for cointegration tests.” Technical report, Queen’s Economics De-partment Working Paper.
Marquez, J. (1990): “Bilateral trade elasticities.” The Review of Economics and Statistics,(pp. 70–77).
Marquez, J. / McNeilly, C. (1988): “Income and price elasticities for exports of developingcountries.” The Review of Economics and Statistics, (pp. 306–314).
McNab, R. M. / Moore, R. E. (1998): “Trade policy, export expansion, human capital andgrowth.” Journal of International Trade & Economic Development, 7 (2), 237–256.
National Bank of Greece (2013): “Economic & Social Market Ananlysis” (July).
Nie, J. / Taylor, L. (2013): “Economic growth in foreign regions and US export growth.” Eco-nomic Review-Federal Reserve Bank of Kansas City, (pp. 31).
Panagiotidis, T. / Printzis, P. (2016): “On the macroeconomic determinants of the housing mar-ket in greece: A VECM approach.” International Economics and Economic Policy, 13 (3),387–409.
Pesaran, H. H. / Shin, Y. (1998): “Generalized impulse response analysis in linear multivariatemodels.” Economics Letters, 58 (1), 17–29.
Phillips, P. C. / Ouliaris, S. (1990): “Asymptotic properties of residual based tests for cointe-gration.” Econometrica, (pp. 165–193).
Riedel, J. (1984): “Trade as the engine of growth in developing countries, revisited.” The Eco-nomic Journal, 94 (373), 56–73.
Saikkonen, P. (1992): “Estimation and testing of cointegrated systems by an autoregressiveapproximation.” Econometric Theory, 8 (1), 1–27.
Senhadji, A. S. / Montenegro, C. E. (1999): “Time series analysis of export demand equations:a cross-country analysis.” IMF Staff Papers, 46 (3), 259–273.
Sharma, A. / Panagiotidis, T. (2005): “An analysis of exports and growth in India: Cointegrationand causality evidence (1971–2001).” Review of Development Economics, 9 (2), 232–248.
Stern, R. M. (1976): “Price elasticities in international trade: An annotated bibliography.”Springer.
Stock, J. H. / Watson, M. W. (1993): “A simple estimator of cointegrating vectors in higher orderintegrated systems.” Econometrica: Journal of the Econometric Society, (pp. 783 –820).
The Relationship Between Greek Exports and Foreign Income 113
Applied Economics Quarterly 64 (2018) 1
Thangavelu, S. M. / Rajaguru, G. (2004): “Is there an export or import-led productivity growthin rapidly developing asian countries? a multivariate VAR analysis.” Applied Economics,36 (10), 1083–1093.
Appendix
Table 3
Unit Root Tests
ADF test statistics
constant constant &trend
Variable level first difference level first difference
ex 0.00 –7.09 –2.38 –7.11
y –0.04 –9.24 –2.84 –9.25
reer –1.43 –7.21 –1.08 –7.28
Germany –0.80 –11.07 –2.13 –11.02
Italy –0.45 –3.56 –1.03 –4.21
Turkey 0.49 –8.76 –1.70 –8.82
RoW –0.46 –9.11 –2.21 –9.06
Notes: Variable ex refers to real Greek merchandise exports. Variable y refers tothe weighted foreign real income of the Greek trading partners. Variables Germany,Italy, Turkey and RoW refer to the real income in the respective regions and variablereer is the real effective exchange rate of Greece. All the above variables are in loga-rithms of the level values. The lag length of the test was based on AIC. Values in boldindicate rejection of the null hypothesis. Critical values based on Mackinnon, 1996.
DF-GLS test statistics
constant constant & trend
Variable level first difference level first difference
ex 1.30 –3.82 –2.36 –6.85
y 1.25 –9.02 –2.57 –9.18
reer –1.26 –3.22 –1.29 –7.08
Germany 0.61 –1.73 –1.57 –3.96
Italy –0.74 –2.48 –0.56 –3.14
Turkey 2.63 –8.09 –1.59 –8.53
RoW 1.83 –9.17 –2.23 –9.17
Notes: Variable ex refers to real Greek merchandise exports. Variable y refers to theweighted foreign real income of the Greek trading partners. Variables Germany, Italy,Turkey and RoW refer to the real income in the respective regions and variable reer isthe real effective exchange rate of Greece. All the above variables are in logarithms ofthe level values. The lag length of the test was based on AIC. Values in bold indicaterejection of the null hypothesis. Critical values based on Elliott et al. 1992 and Mack-innon, 1996.
114 Konstantinos Chisiridis and Theodore Panagiotidis
Applied Economics Quarterly 64 (2018) 1
Reproduced with permission of copyright owner.Further reproduction prohibited without permission.
Konstantinos Chisiridis / Theodore Panagiotidis: The Relationship Between Greek Exports and Foreign Income
Abstract
1. Introduction
2. The Composition of Greek Export Goods per Region
3. Methodology
3.1 The Vector Autoregressive Model
4. Empirical Results
4.1 Data and Unit Root Tests
4.2 Cointegration Analysis
4.3 Income and Price Elasticities for Greek Exports
4.4 Dynamic Analysis
5. Conclusion
References
Appendix
Organizational structure, innovationperformance and customer relationship valuein the Greek advertising and media industry
Paraskevi DekoulouCentre for Innovation, Entrepreneurship and New Technologies, University of Nicosia, Nicosia, Cyprus, and
Panagiotis TrivellasDepartment of Logistics Management, Technological Educational Institute of Central Greece, Thiva, Greece
AbstractPurpose – This paper aims to explore the impact of organizational structure dimensions on innovation performance as well as its implications onbusiness customers’ relationship value and financial performance in the business-to-business (B2B) market of the Greek advertising and mediaindustry.Design/methodology/approach – Based on a sample of 180 executives, who are at the helm of 163 Greek advertising and media organizations,the authors apply the partial least square method to test the association of organizational structure with innovation performance, businesscustomers’ relationship value and financial outcomes.Findings – Findings have brought to light that training boosts organization’s capacity to innovate, whereas direct supervision as a coordinationmechanism significantly restricts this capacity. Innovation performance in the advertising B2B market fosters business customers’ relationship valueand financial performance, while financial outcomes are also beneficially affected by profitable relationships with customer relationship value.Practical implications – Because of the dramatic decline in their profitability caused by the economic crisis in the past five years, Greek advertisingand media companies are threatened with extinction; thus, they are required to enhance their effectiveness through the adoption of a moreinnovation-oriented structure. Thus, managers should facilitate structures supporting training and delimiting direct supervision to foster thedevelopment of a competitive advantage built on innovation, creativity and business clients’ relationship.Originality/value – This study contributes to the existing relationship marketing literature because it introduced Mintzberg’s typology to measureorganizational structure and led to the diagnosis of the associations between different dimensions of organizational structure and various aspectsof performance in the media and advertising industry, revealing the partial mediating role of customer relationship value between innovation andfinancial performance in the B2B market.
Keywords Performance, Innovation, Organizational structure, Buyer–seller relationship, Advertising, Greece
Paper type Research paper
1. IntroductionIn today’s highly turbulent and fiercely competitive knowledgeeconomy, organization’s capacity to innovate has beenhighlighted as a crucial strategic asset and a major source ofcompetitive advantage (Jiménez-Jiménez and Sanz-Valle,2008; Salim and Sulaiman, 2011). Innovation-basedorganizational action has a significant beneficial impact on allaspects of organizational performance, including customerrelationships and financial outcomes (Khan et al., 2010;Muiruri and Ngari, 2014). Therefore, one of the mostimportant challenges for contemporary managers consists inrendering both their organization’s staff members andprocesses as innovation-oriented as possible. Organizationalstructure is regarded as an internal factor considerablyinfluencing a company’s capacity to innovate and base its daily
operation and competitiveness on regular generation andimplementation of novel ideas (Birkinshaw et al., 2008; Coshet al., 2012).
Over the past decade, both academics’ and practitioners’research interest in the development and management ofsupplier-customer relationships has significantly increased(Chen and Fung, 2013; Grönroos, 2004; Mitręga, 2012;Palmatier et al., 2006; Rajagopal and Sanchez, 2005;Valtakoski, 2015). Drawing upon the relationship marketingliterature, supplier–customer relationships in businessmarkets provide a solid foundation for organizations toacquire new competencies, maintain resources, to build asustainable competitive advantage and achieve superiorperformance (Fink et al., 2008; Hutt et al., 2000; Schellhaseet al., 2000; Walter and Gemünden, 2000; Walter and Ritter,2003). In particular, Storbacka and Nenonen (2009, p. 362)defined business-to-business customer relationships “aslongitudinal social and economic processes for the co-creationof value”, and relationship performance “as the total value
The current issue and full text archive of this journal is available onEmerald Insight at: www.emeraldinsight.com/0885-8624.htm
Journal of Business & Industrial Marketing32/3 (2017) 385–397© Emerald Publishing Limited [ISSN 0885-8624][DOI 10.1108/JBIM-07-2015-0135]
Received 16 July 2015Revised 16 July 2015Accepted 10 August 2015
385
formed during the interaction between firm and customer overtime”. Previous research has attempted to shed light on theway that buyer–seller relationships in business-to-business(B2B) contexts can be more productive and long-lasting(Cannon and Perreault, 1999); however, metrics for valuecreation in these relationships are still in their infancy (Ulaga,2003). In this way, customer relationship value is consideredas the result of successful strategic relationship marketingbetween media firms and advertising agencies with theirbusiness clients. It is assumed that media and advertising firmsare experts in communication and usage of best practices inpromotional activities, as their main service consists ofcommunicating content to the marketplace on behalf of theirclients (Waller et al., 2001).
The emergence of the so-called knowledge economy hasbrought about the imperative need for knowledge-intensivefirms (KIFs), where knowledge, creativity and innovationconstitute the most important inputs, whereas intellectualmaterial –ideas, information and experience – conveyed byvigilant employees are regarded as the key production resource(Alvesson, 2000; Robertson and Swan, 2003; Zabala et al.,2005). Advertising and media companies are typical examplesof KIFs (Swart and Kinnie, 2003) as well as a basing link inthe B2B market, because they connect commercial companieswith end-consumers. Therefore, advertising and mediacompanies can be considered as a key player for supply chainmanagement, because they favor the creation of strategicnetworks and the consequent reinforcement of social capital;strategic networks are a powerful tactic to accelerate andsustain corporate success among small businesses (Milleret al., 2007).
This is rather crucial for Greek advertising and mediacompanies for which the current financial crisis has broughtabout a sharp decline in their profitability. This can beattributed to the decline in advertising expenditure for bothpublic and private sectors because of regulation, austeritymeasures and overtaxation imposed (Deputy Minister ofCulture in Advertising.gr, 2010; IME GSEVEE, 2013;TFGR, 2012). Economic recession and reduced turnovershave resulted in a significant number of both public andprivate organizations, as well as key customers of advertisingand media companies, being closed down or merged, whereasexisting organizations have restricted significantly theiradvertising budget in an attempt to cut down their operatingcosts (IMERISIA, 2012).
This study aims to shed light upon the interplay betweendifferent organizational structure dimensions, innovationperformance and the role of business customers’ relationshipvalue on the link between innovation and financial outcomesin the context of the Greek advertising and media industry.Thus, our research contributes to the B2B relationshipmarketing literature multi-fold. Although several scholarsunderline the significant impact of organizational structure oncompany’s overall performance (Green et al., 2005; Meijaardet al., 2005), few have examined different organizationalstructure dimensions in an holistic approach and variousaspects of organizational performance (Zheng et al., 2010).Furthermore, even though a growing quantity of researchwork has been devoted to organizational structure, to the bestof the authors’ knowledge, the latter has not been minutely
studied in the advertising and media sector and B2Binterrelationships yet. In this attempt, multi-dimensionalMintzberg’s typology has been adopted, synthesizing severalaspects recommended in empirical research. Moreover,Mitręga’s (2012) findings that customer relationship qualitymediates the impact of internal communication andrelationships crossing organizational structure on companyperformance in B2B markets set the ground for this study toinvestigate the crucial role of business customers’ relationshipvalue on the linkage among structure, innovation and financialperformance.
After this introductory section, the next section presentsconcisely the Mintzberg’s typology and the development ofresearch hypotheses. This is followed by the researchmethodology and statistical analysis. Afterwards, discussion offindings, conclusions and management implications areformed. Finally, research limitations are denoted.
2. Theoretical background
2.1 Organizational structureOrganizational structure can be described as the set of ways inwhich organizational work is divided into separate tasks,delegated and coordinated towards the achievement ofcorporate goals (Mintzberg, 1983; Stacey, 2003). Structureconfigures the context within which power and control areexerted, duties are fulfilled, strategic options are formulatedand enables the implementation of these options (Hunter,2002; Spanos et al., 2001). It influences resource allocation,favors internal and external communication and strengthensorganizational ability to respond to changes in businessenvironment, to learn and innovate (Chen et al. 2010;Martínez-León and Martínez-Garcia, 2011).
Mintzberg’s theory encompassing structural aspects is oneof the most widely influential works on strategic management;although not sufficiently used by academic researchers, itoffers a theoretical synthesis of a wide range of managementand organizational research (Lunenburg, 2012). Building onMintzberg’s work (1983), the following organizationalstructure dimensions emerge:● Formalization refers to the extent to which decision-making,
working relationships and operational routines are governedby specific standard rules, regulations, policies andprocedures (Spanos et al., 2001). It coordinates and controlswork processes and behavior through codification anddocumentation, and sets explicit standards for work attitudeand outcomes (Andrews and Kacmar, 2001; Bodewes,2002).
● Decentralization refers to the extent to which the right andauthority to make decisions and evaluate activities within anorganization is distributed among different organizationallevels (Martínez-León and Martínez-Garcia, 2011).Decentralization defines whether power and control areshared among various structural components as well aswhether all staff members are involved in decision andstrategy making, policy shaping and resource allocation(Wally and Baum, 1994).
● Specialization refers to the extent to which each employee orworking group accomplishes a limited and predefined varietyof duties (Lloria, 2007). It requires the division ofcomplicated organizational processes, operations and
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
386
activities into simpler fundamental tasks (Schilling et al.,2003).
● Training can be described as all organizational learningactivities aiming at enhancing individual performancethrough the acquisition of new knowledge and skills andthrough the implementation of novel work behavior patterns.
● Mutual adjustment is related to the extent to whichcoordination of organizational work and fulfillment oforganizational tasks are mainly based on tight cooperationand informal communication not only between members ofthe same working group but also between different workinggroups (Mintzberg, 1983). Staff members organize andmanage their own tasks and are responsible for their ownoutput, whereas internal communication and collaboration isfacilitated by coordination managers or committees.
● Direct supervision refers to the extent to which coordinationand execution of organizational work is based on theprovision of restrictive instructions and tight supervision ofsubordinates. Staff members are restricted to just executetasks that are strictly defined and delegated by a supervisor,who is fully responsible for his subordinates’ work output.
In fact, Mintzberg’s typology may amalgamate the mostcommonly used structural dimensions in empirical research,such as formalization (Meirovich et al., 2007; Pertusa-Ortegaet al., 2010), centralization (Meirovich et al., 2007;Pertusa-Ortega et al., 2010; Olson et al., 2005) andspecialization (Olson et al., 2005). Some authors embracestructural complexity as a different structural dimension;however, it may be reflected on specialization, coordinationmechanisms and hierarchical levels, depending on themeasures adopted (Ganesh and Joseph, 2011). Mutualadjustment and direct supervision are two structuralcomponents that other researchers have not taken intoconsideration as discrete dimensions.
2.2 Organizational structure, innovation performanceand customer relationship valueOrganizational structure has been advocated as one of theorganizational resources that can decisively contribute to thereinforcement of organization’s capacity to innovate(Andrews, 2010; Meijaard et al., 2005). The way in whichorganizational work is divided, delegated and coordinatedaffects cooperation and internal communication, impacts freeflow of information and exchange of ideas and favors orhinders experimentation, generation and dissemination of newknowledge.
Controversy surrounds the impact of formalization oninnovation performance. An organization’s reliance onstandardized rules and procedures reduces creativity andhinders experimentation and innovative problem-solving.When staff members are not allowed to deviate from existingknowledge and predefined behavior patterns, their ability togenerate novel ideas and develop innovative and moreeffective working practices is considerably limited (Hartlineet al., 2000; Jansen et al., 2006). On the other side of thespectrum, there are those researchers who argue thatstructural formalization fosters the generation of ideas andsuggestions favoring the enhancement of organizationalroutines (Zollo and Winter, 2002); improved routines are thenstandardized and become part of an organization’s everyday
operation (Cosh et al., 2012; Benner and Tushman, 2003).Palmer and Dunford (2002) highlight formalization’s positiveimpact on the development of novel working practices,whereas Auh and Menguc (2007), Beckmann et al. (2007),Green et al. (2005) and Froehle et al. (2000) underline thatformalization significantly facilitates the diffusion of newknowledge, its utilization and transformation into innovativeand competitive products/services.
Thus, in alignment with the majority of the relevantempirical research, the following hypothesis is proposed:
H1.1. The higher the formalization, the higher the innovationperformance.
Contrary to formalization, decentralization has a unanimouslypositive impact on innovation performance. Bandura (2000),Cosh et al. (2012) and Damanpour (1991) reveal that adecentralized structure facilitates the development andinitiation of novel products and services as well as theintroduction of pioneering organizational practices.Centralization reduces staff members’ ability to seekinnovative solutions to problems and limits the sense ofcontrol over individual work (Atuahene-Gima, 2005; Jansenet al., 2006). It hinders both internal and externalcommunication and decreases not only the quantity but alsothe quality of ideas, information and knowledge acquired,generated and diffused within the organization (Cardinal,2001; Sheremata, 2000). Hence, we propose the followinghypothesis:
H1.2. The higher the decentralization, the higher theinnovation performance.
Specialization constitutes another structural dimension thathas been found to be favorably associated with innovationperformance (Garicano, 2000; Ho and Wong, 2009). Kogutand Zander (1996), Lee et al. (2008), Nielsen (2005) andNorman (2004) underline the positive correlation betweenhorizontal specialization and an organization’s ability to useand apply new knowledge, whereas Covin and Slevin (1988)combine increased specialization with fostering inquiry,experimentation and innovation. Interaction betweenindividuals with specialized knowledge facilitates the creationof an enriched pluralistic organizational knowledge base, whileit stimulates the materialization of novel collective ideas.Therefore, specialization is expected to favor innovationperformance. Therefore, the following hypothesis has beensuggested:
H1.3. The higher the specialization, the higher the innovationperformance.
Similar to formalization, inconclusiveness surrounds the impactof training on innovation performance. There are certainresearchers who reveal an absence of a statistically significantrelationship between training and overall organizationalperformance (Purcell et al., 2003). Nevertheless, otherresearchers highlight the beneficial influence of training on allaspects of organizational performance, including innovationperformance (Shimpton et al., 2005, 2006; Zanko et al., 2008).Training favorably influences employee ability to transform andadapt to their volatile organizational environment (Susan et al.,
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
387
2006). It fosters the development and application of innovativeideas as well as the integration of novel working practices andprocedures into organization’s operation (Jiménez-Jiménez andSanz-Valle, 2008; Kuo, 2011). Thus, we advance the followinghypothesis:
H1.4. The higher the training, the higher the innovationperformance.
Ho and Wong (2009) stress that because of the growinglyknowledge-focused nature of contemporary work, mutualadjustment between staff members appears rather crucial fororganization’s overall performance and well-being. Mutualadjustment enhances internal processes and cooperation,promotes regular and unhindered both internal and externalcommunication and facilitates exchange of knowledge, ideasand information (Cleveland, 2000; Perry-Smith and Shalley,2003). It favors the development and materialization of novelideas as well as the initiation of innovative products andservices. Mintzberg (1983) underlines that advertising andmedia companies belong to this type of organizations which donot offer their customers standardized products, but theiroutput demands creativity, close collaboration and systematicinteraction not only between members of a working group butalso between different working groups. So, we promote thefollowing hypothesis:
H1.5. The higher the mutual adjustment, the higher theinnovation performance.
On the other hand, direct supervision does not favor fruitfulcooperation, regular and creative communication andinteraction among employees (Lee and Choi, 2003; Lópezet al., 2006). Organizational members are deprived of the rightto take initiative, experiment, suggest novel ideas and beinventive. In addition, they are less motivated to contribute tothe achievement of organizational objectives and therealization of organizational vision through fostering theirfirm’s innovation performance. Amitay et al. (2005), Atwoodet al. (2010), Lee (2012) and Mirkamali et al. (2011)underline that participative leadership is a crucial preconditionfor the creation, the diffusion and the application of newknowledge and ideas, whereas Yiing and Ahmad (2009) stressthat direct supervision reduces employee satisfaction,productivity and individual performance. Hence, directsupervision is expected to detrimentally affect innovationperformance:
H1.6. The higher the direct supervision, the lower theinnovation performance.
A significant stream of research confirms the benefits ofinnovation performance (Cosh et al., 2012). In particular, anorganization’s capability to systematically exhibit innovativebehavior enhances customer relationships and increasescustomer relationship value (Huang, 2008; Khan et al., 2010),mainly by improving internal process effectiveness (Bankeret al., 2004; Wu and Hung, 2008). Customer relationshipvalue refers to the value emanating from the businessrelationship between customers and suppliers, which can becreated by offering benefits or by reducing costs for bothpartners. Indeed, innovation capacity is widely regarded as a
principal source of organization’s competitiveness not onlybecause of its beneficial impact on customer relationship valuebut also because of its favorable influence on financialperformance. Innovation can be a critical element inimproving financial outcomes (Battor and Battour, 2010),because it can enable an organization to significantlydifferentiate its products and services from the competitiveones (Klomp and Van Leeuwen, 2001) as well as it tends toreinforce revenue and profitability growth (Eggert et al.,2014). Drawing from the highly competitive banking sector,Batiz-Lazo and Woldesenbet (2006), Muiruri and Ngari(2014) and Roberts and Amit (2003) highlight the undeniableimportance of innovation for product and processimprovement as well as for achieving strong competitivenessand superior financial performance. Increased customerrelationship value inevitably supports an organization toenhance its financial outcomes (Battor and Battour, 2010;Liang et al., 2009).
In the light of the above discussion, the followinghypotheses have been promoted:
H2. The higher the innovation performance, the higher thecustomer relationship value.
H3. The higher the innovation performance, the higher thefinancial performance.
H4. The higher the customer relationship value, the higherthe financial performance.
The conceptual framework describing research hypotheses isillustrated in Figure 1.
3. Research methodology
3.1 Sample and questionnaire designThe research population is composed of all advertising andmedia companies (television and radio stations, newspaperand magazine titles) that are based in Greece. This list ofcompanies is the result of a synthesis of the enlisted advertisingand media firms at the Greek Financial Directory (ICAP) andthe members of Greek Advertising and CommunicationCompanies (EDEE). More specifically, concerningadvertising agencies, the research population included 49advertising agencies, members of EDEE and the Union ofGreek Advertising and Communication Companies.Advertising agencies that are EDEE members represent 95 percent of the total sector’s turnover. Moving on to mediacompanies, with the use of the purposive sampling method,the sample was composed of the 11 national televisionstations, the 40 radio stations achieving the highest ratings, the40 national newspapers and the 40 national magazinesachieving the highest circulation rates. Therefore, the presentfield research was conducted in 49 advertising agencies(census), 40 newspaper and 40 magazine titles and 40 radio
Figure 1 Conceptual framework
H1OrganisationalStructure
Innovation Performance
H2 Customer Relationship
Value
H4 Financial Performance
H3
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
388
and 11 television stations situated in Greece. Structuredquestionnaires were delivered to 180 senior level employees,one from each company, and 163 valid questionnaires werereturned. The response rate achieved was 90.5 per cent. Themajority of respondents were male (63.2 per cent) and theirage ranged from 41 to 50 years (47.9 per cent). Most of themhold a university degree (91.5 per cent), and their total workexperience exceeds 16 years (68 per cent). Concerningcompanies, most of them employ from 11 to 50 individuals(39.4 per cent), whereas their years of operation range from 11to 30 (53.6 per cent).
In this study, organizational structure was measured withthe use of a scale based on Mintzberg’s typology oforganizational structure. It contains 21 items grouped in sixsubscales as follows: formalization (sample items: “Workaccomplishment and coordination is based on standardizedrules and formal processes” and “work attitude is guided byformal rules”), decentralization (sample items: “Decisionsare made by employees of all levels and not only by topmanagement” and “all departments take part in decisionmaking”), specialization (sample items: “Employee dutiesare clearly and strictly separated” and “employeescoordinate the accomplishment of their duties in the waythey judge as most effective”), training [sample items: “Inorder for employees to acquire necessary job-related skillsand knowledge, they are offered opportunities for (i)on-the-job training (job rotation, mentoring, coaching),and (ii) off-the-job training (lectures, seminars,distance-learning)”], mutual adjustment (sample items:“Work accomplishment is strongly based on: (i) employeemutual adjustment, and (ii) mutual adjustment amongdifferent work groups”) and direct supervision (sampleitems: “Work accomplishment and coordination is based onrestrictive orders by supervisor” and “Subordinates arerequired to simply accomplish the tasks strictly defined anddelegated by supervisor”). Mintzberg’s theory contributionon the strategic management literature regardingorganizational structure processes is widely acknowledged(Lunenburg, 2012).
Customer relationship value has its roots in the relationshipmarketing literature (Ulaga and Eggert, 2006; Walter et al.,2003), and it is considered as the value emanating from therelationship between customers and suppliers, which can becreated by offering benefits or by reducing costs for bothpartners. Hence, value is also an outcome of the process ofutilizing products/services and the interaction and sharingactivities between suppliers and buyers. Ryssel et al. (2004)recognize two kinds of benefits, the direct (e.g. profit andvolume) and the indirect value elements (e.g. innovation,market and access). Following this rationale, Kohtamaki et al.(2012) suggested relationship performance improvement insubcontractor–customer relationships because of the positiveimpact of this relationship on productivity, product quality,delivery accuracy and use of capital. However, our approachof customer relationship value exceeds the short-termperformance to include long-term and intangible outcomes ofthis relationship.
Although majority of the empirical studies were focused onconsumer markets by adopting relevant constructs such ascustomer relationship quality, a driver of relationship value
(Brashear-Alejandro et al., 2011), there is clear evidence that thisscale can be effectively used in the context of business markets(Han and Sung, 2008; Mitręga, 2012). Customer relationshipvalue scale consisted of seven items, and it was based on theprevious studies of Ulaga and Eggert (2006), Ritter, (2007),Kohtamaki et al. (2012) and Brashear-Alejandro et al. (2011).Sample items include the following:● Taking into account all the costs and benefits associated to
our relationship with our business customers, thisrelationship is valuable;
● The activities proposed by our business customers(information, processes, technology, customer service,etc). are of great value;
● The financial costs, the time and resources spent whendealing with our business customers are rewarded by thebenefits obtained from our relationship;
● Compared to other suppliers/competitors, our businesscustomers accrue more advantages by our relationship;and
● Business customers are more satisfied with us than withour competitors.
The six-item innovation performance scale was based onthe core characteristics of process and product innovationwhich has been suggested in the relevant literature(Avlonitis et al., 1994; Deshpande et al., 1993; Prajogoet al., 2004; Subramanian and Nilakanta, 1996), includingthe number of new products, the novelty of the newproducts, the speed of innovation, the level ofinnovativeness and the adoption of novel technological andorganizational processes.
Firm’s financial performance was measured as profitability,sales volume, profit margin and return on investment over thepast three years as compared to industry’s average. Severalscholars have adopted similar items (Dess and Davis, 1984) inself-reported constructs, as this comparative approach has beenconsidered as more effective at eliciting responses than theprovision of exact values by respondents (Tomaskovic-Deveyet al., 1994). Although self-reported bias may be regarded as apotential danger, research has found that perceptual measurescan be a reasonable substitute for objective indices ofperformance (Dess and Robinson, 1984), and they exhibit asignificant correlation with objective measures (Tomaskovic-Devey et al., 1994).
Responses were assessed using five-point Likert scaleitems. The questionnaire was pilot-tested via personalinterviews with ten executives with long experience in theGreek advertising and media industry. FollowingRogelberg’s and Stanton’s (2007) suggestions, t-testanalysis was performed between early- and late-waverespondents to confirm that non-response bias does notoccur in the present survey.
Harman’s single-factor test (Podsakoff et al., 2003)confirmed that no evidence of common method variance ispresent, because 11 factors emerged by conductingexploratory factor analysis. In addition, Table I shows that thehighest correlation among the latent variables is 0.49, far lowerthan the threshold value of 0.90 for the existence of severecommon methods bias (Bagozzi et al., 1991).
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
389
3.2 Reliability and validity of the scalesData were analyzed through path modeling using the partial leastsquares (PLS) approach and the SmartPLS software (Ringleet al., 2005). A strong advantage of the PLS approach comparedto the covariance-based structural equation modeling is its abilityto deal with situations where knowledge about distribution of thelatent variables is restricted, requirements about the closenessbetween estimates and the data should be met and sample size istoo small (Fornell and Cha, 1994). In our research model, allfirst-order factors are constructs specified with reflectiveindicators, as depicted in Figure 2.
First, the validity and reliability of the constructs in themeasurement model were evaluated. All constructs exhibitedCR and Cronbach’s alpha greater than the minimumacceptable level of 0.70 (Fornell and Larcker, 1981), asillustrated in Table I.
Fornell’s and Larcker’s (1981) average variance extracted(AVE) criterion is adopted for the estimation of scales’convergent validity. As shown in Table I, all scales met thiscriterion, because the AVE value of each latent variable is higherthan 0.50 (Henseler et al., 2009). In addition, factor loadings of
all items on their respective constructs are greater than 0.731(p � 0.01), surpassing the 0.70 cut-off value (Barclay et al.,1995), whereas their loadings on unrelated constructs are lessthan 0.4, designating adequate convergent validity.
Discriminant validity was examined by Fornell’s andLarcker’s (1981) AVE test and correlations criterion. Asshown in Table I, all constructs in our research modeldemonstrated adequate discriminant validity, because thediagonal elements which are the square roots of the AVEs aregreater than the off-diagonal elements (correlations) in thecorresponding rows and columns.
3.3 PLS analysisThe path relationships (standardized coefficients) of themodel were estimated by performing PLS (smartPLS, Ringleet al., 2005). The bootstrapping procedure (163 cases, 5,000samples and no sign changes option) was adopted to evaluatethe structural model and, particularly, the statisticalsignificance of all parameter estimates (Chin, 1998). The PLSstructural model is mainly evaluated by R2, goodness-of-fit(GoF) index and by using the Stone–Geiser Q2 test forpredictive relevance (Chin, 1998; Hair et al., 2014; Sarstedtet al., 2014; Tenenhaus et al., 2005). Chin (1998)characterized R2 values of 0.67, 0.33 and 0.19 as substantial,moderate and weak, respectively. Results confirmed moderateR2 values in our study for all organizational aspects, but thefinancial dimension as shown in Table II. The GoF index is0.41, meaning that the proposed model takes into account 41per cent of the achievable fit, which is considered adequategiven the complexity of the model. As shown in Table II, allQ2 statistics indicators [cross-validated communality (H2j)and cross-validated redundancy (F2j)] have larger than zerovalues, verifying the predictive relevance of the model (Fornelland Cha, 1994).
The results of the parameter estimation (path coefficientsand level of significance) are shown in Table III and illustratedin Figure 2. No serious problems of multicollinearity existbetween the independent variables as variance inflation factors(VIFs) are far below the three-points limit suggested in thesocial sciences literature.
Results indicate that among organizational structuredimensions, only direct supervision (standard beta � �0.212,p � 0.01) and training (standard beta � 0.477, p � 0.001) are
Table I Results of reliability, convergent and discriminant validity analysis of all scales
Mean SD CR Cronbach alpha AVE 1 2 3 4 5 6 7 8 9
1. MUTADJ 3.64 0.797 0.882 0.758 0.789 0.8882. TRAIN 2.16 0.825 0.820 0.741 0.603 0.390�� 0.7763. FORM 3.37 0.787 0.722 0.768 0.597 0.164� 0.177� 0.7734. SPEC 3.67 0.709 0.778 0.791 0.672 0.355�� 0.237�� 0.240�� 0.8205. DCENTR 2.75 0.787 0.798 0.740 0.540 0.362�� 0.342�� �0.020 0.188� 0.7356. DSUPR 3.17 0.555 0.781 0.714 0.648 �0.205�� �0.108 0.204�� �0.043 �0.169� 0.8057. CUSTREL 3.92 0.590 0.833 0.748 0.623 0.393�� 0.361�� 0.033 0.143 0.295�� �0.262�� 0.7898. FINPRF 2.48 0.703 0.902 0.865 0.647 0.244�� 0.269�� 0.038 �0.002 0.236�� �0.214�� 0.324�� 0.8049. INNOVPRF 2.71 0.853 0.832 0.737 0.729 0.198� 0.494�� 0.097 0.025 0.304�� �0.158� 0.428�� 0.296�� 0.854
Notes: MUTADJ: Mutual adjustment; TRAIN: Training; FORM: Formalization; SPEC: Specialization; DCENTR: Decentralization; DSUPR: Direct supervision;CUSTREL: Customer relationship value; FINPRF: Financial performance and INNOVPRF: Innovation performance; Bold diagonal elements are the squareroots of the AVEs and off-diagonal elements are correlation coefficients among variables; correlations are significant at � p � 0.05 and �� p � 0.01
Figure 2 Path model and PLS estimates
INNOVPRF (R2 = 0.390)
DSUPR CUSTREL (R2 = 0.270)
FINPRF (R2 = 0.135)
0.401***
0.183*
0.477***
TRAIN
FORM
SPEC
MUTADJ
DCENTR 0.251**
–0.212**
Notes: Dotted lines designate non-significant paths; MUTADJ:Mutual adjustment; TRAIN: Training; FORM: Formalization,SPEC: Specialization; DECENTR: Decentralization; DSUPR:Direct supervision; CUSTREL: Customer relationship value;FINPRF: Financial performance; INNOVPRF: Innovationperformance; Standardized estimated values are significant at*p < 0.05; **p < 0.01; ***p INNOVPRF 0.102 1.02 n.s H1.1 Not supportedDCENTR -> INNOVPRF 0.111 1.34 n.s H1.2 Not supportedSPEC -> INNOVPRF �0.02 0.29 n.s H1.3 Not supportedTRAIN -> INNOVPRF 0.477 5.73 p � 0.001 H1.4 SupportedMUTADJ -> INNOVPRF �0.120 1.59 n.s H1.5 Not supportedDSUPR -> INNOVPRF �0.212 2.93 p � 0.01 H1.6 SupportedINNOVPRF -> CUSTREL 0.401 6.310 p � 0.001 H2 SupportedINNOVPRF -> FINPRF 0.183 1.98 p � 0.05 H3 SupportedCUSTREL -> FINPRF 0.251 3.04 p � 0.01 H4 Supported
Notes: MUTADJ: Mutual adjustment; TRAIN: Training; FORM: Formalization; SPEC: Specialization; DCENTR: Decentralization; DSUPR: Direct supervision;CUSTREL: Customer relationship value; FINPRF: Financial performance; INNOVPRF: Innovation performance
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
391
for organization’s capacity to learn and innovate. Moreover,this result is consistent with Edmondson and Nembhard(2009), Jones (2010), Mink et al. (1993) and Pawar andEastman (1997) stating that direct supervision does not allowa manager to be a learning leader to energize employees andpromote organizational learning and innovation. The negativeimpact of direct supervision on an organization’s innovationperformance is rather critical in case of media and advertisingcompanies, typical examples of knowledge-intensive firms,whose operation and competitiveness require followingtechnological trends as well as regular acquisition of newknowledge and skills from the part of staff members.
Considering the last three hypotheses (H2, H3 and H4),findings reveal the partial mediating role of customerrelationship value on the association between innovationperformance and financial outcomes in the B2B Greekadvertising and media market.
Indeed, organization’s capacity to innovate has a significantbeneficial impact on its capacity to create customer relationshipvalue, which can be considered as a strategic weapon for thecontemporary business world (Mathuramaytha andUssahawanitchakit, 2008). Therefore, H2 has been confirmed.In the contemporary knowledge economy suffering fromeconomic crisis, innovation is a key element for the creation ofvalue in products and services (Hsu and Fang, 2009). Turningsuccessfully novel and creative ideas into innovative products andservices that reflect current consumer trends and satisfy customerneeds and wants inevitably results in the creation of customervalue (Kandampully, 2002; Kanten and Yaslioglu, 2012) andcontributes to organization’s long-term success and increasedperformance.
In exchange for being provided with high-value productsand services, satisfied customers reward the organization withtheir preference against competitors. Elevated customersatisfaction is a critical building block of high customerrelationship quality (De Wulf et al., 2001), which in turn leadsto a strong brand name, increased market share and elevatedfinancial outcomes. This finding is in line with previousstudies reporting that customer relationship value inevitablyinfluences an organization’s financial performance (Battor andBattour, 2010; Devie and Widjaja, 2012; Gupta and Zeithaml,2006; Liang et al., 2009). Building and maintaining effectiverelationships with customers especially in B2B marketsfavorably influences sales, profitability and other financialindicators. Moreover, this result is in accordance with thoseresearchers who stress that customer satisfaction brings aboutcustomer willingness to share openly their needs andpreferences, and thus, to enable organizations indicate marketopportunities, initiate profitable innovations and offercompetitive products/services that meet market demands(Doyle and Wong, 1997; Sheth, 2001). Finally, it is alsoconsistent with previous inquiries stating that by generatingmarket intelligence, effective customer relationshipsstrengthen financial outcomes (Anderson et al., 1997; Devieand Widjaja, 2012; Gupta and Zeithaml, 2006; Hallowel,1996). Thus, H3 has been confirmed.
Financial outcomes are also beneficially impacted by anorganization’s capacity to continuously innovate(Durán-Vázquez et al., 2012; Marques et al., 2010). Thisresult is congruent with Batiz-Lazo and Woldesenbet (2006),
Eggert et al. (2014), Klomp and Van Leeuwen, (2001) andMuiruri and Ngari (2014), stressing that innovationconstitutes a principal source of competitive advantage and acritical enabler of superior financial performance forcontemporary organizations. This is vitally important for KIFswhere generating notable financial outcomes heavily dependson regular knowledge acquisition and innovation initiation.Therefore, H4 has been confirmed.
5. Conclusions and managerial implicationsBased on the present study’s findings, it can be concluded that inthe context of the Greek advertising and media industry, trainingand direct supervision are the two organizational structuredimensions that critically influence positively and detrimentallythe organization’s capacity to innovate. Increased innovationperformance fosters the creation of high business customerrelationship value as well as the achievement of optimumfinancial outcomes. Hence, Greek advertising and mediacompanies in the B2B market could enhance their overallperformance with the aid of a more innovation-orientedstructure. More systematic staff training in combination with aminimized provision of direct supervision and strict guidance arethe principal organizational structure features that couldreinforce an organization’s capacity to innovate. By adoptingsuch an organizational structure, these companies could boostthe level of innovativeness, favor and accelerate generation andimplementation of novel ideas, increase the number and thenovelty of new products/services and promote the adoption ofnovel technological and organizational processes. Moreover, bypursuing a higher innovation performance, these companiescould build and maintain more profitable and effectiverelationships with their business customers; gain their preference,satisfaction and loyalty; and enhance their profitability, salesvolume, growth rate, profit margin and return on investment.
6. LimitationsThis study has certain limitations, restraining its ability forgeneralization in different contexts. First, self-administeredquestionnaires were adopted, which may raise methodologicalissues. However, methodological robustness was verified bysuccessful tests of scales’ reliability and validity as well ascommon method bias. Second, the causality of the proposedrelationships cannot be justified, because this study iscross-sectional. In the future, a longitudinal study may revealthe underlying mechanisms among the variables investigatedand their change over time. Third, structural and culturaldifferences should also be considered, because the fieldresearch was conducted in the advertising and media industryin Greece, a country suffering from fierce financial recession inthe European Union periphery. Furthermore, the conceptualframework of this study has not taken into consideration otherantecedents of organizational performance, such as businessstrategy and organizational culture (Zheng et al., 2010).
References
Advertising.gr (2010), “Reduction in advertising expenditure,Deputy Minister of Culture comments, Advertising.gr”,available at: www.advertising.gr/Article/659/Meiosi-stin-kratiki-diafimisi (accessed 8 October 2013).
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
392
Alejandro, T.B., Souza, D.V., Boles, J.S., Ribeiro, Á.H.P.,Monteiro, P.R.R. (2011), “The outcome of company andaccount manager relationship quality on loyalty,relationship value and performance”, Industrial MarketingManagement, Vol. 40 No. 1, pp. 36-43.
Alvesson, M. (2000), “Social identity and the problem ofloyalty in knowledge-intensive companies”, Journal ofManagement Studies, Vol. 37 No. 8, pp. 1101-1123.
Amitay, M., Popper, M. and Lipshitz, R. (2005), “Leadershipstyles and organizational learning in community clinics”,The Learning Organization, Vol. 12 No. 1, pp. 57-70.
Anderson, E., Fornell, C. and Rust, R. (1997), “Customersatisfaction, productivity, and profitability: differencesbetween goods and services”, Marketing Science, Vol. 16No. 2, pp. 129-145.
Andrews, M.C. and Kacmar, K.M. (2001), “Discriminatingamong organizational politics, justice, and support”, Journalof Organizational Behavior, Vol. 22 No. 4, pp. 347-366.
Andrews, R. (2010), “Organizational social capital, structureand performance”, Human Relations, Vol. 63 No. 5,pp. 583-608.
Atuahene-Gima, K. (2005), “Resolving the capability–rigidityparadox in new product innovation”, Journal of Marketing,Vol. 69 No. 4, pp. 61-83.
Atwood, M.A., Mora, J.W. and Kaplan, A.W. (2010),“Learning to lead: evaluating leadership and organizationallearning”, Leadership & Organization Development Journal,Vol. 31 No. 7, pp. 576-595.
Auh, S. and Menguc, B. (2007), “Performance implications ofthe direct and moderating effects of centralization andformalization on customer orientation”, Industrial MarketingManagement, Vol. 36 No. 8, pp. 1022-1034.
Avlonitis, G.J., Kouremenos, A. and Tzokas, N. (1994),“Assessing the innovativeness of organizations and itsantecedents: project innovstrat”, European Journal ofMarketing Vol. 28 No. 11, pp. 5-28.
Bagozzi, R.P., Yi, Y. and Singh, S. (1991), “On the use ofstructural equation models in experimental designs: twoextensions”, Marketing, Vol. 8 No. 2, pp. 125-140.
Bandura, A. (2000), “Exercise of human agency throughcollective efficacy”, Current Directions in PsychologicalScience, Vol. 9 No. 3, pp. 75-78.
Banker, R.D., Chang, H. and Pizzini, M.J. (2004), “Thebalanced scorecard: judgmental effects of performancemeasures linked to strategy”, The Accounting Review, Vol. 79No. 1, pp. 1-23.
Barclay, D., Higgins, C. and Thompson, R. (1995), “Thepartial least squares (PLS) approach to causal modeling:personal computer adoption and use as an illustration”,Technology Studies, Vol. 2 No. 2, pp. 285-323.
Batiz-Lazo, B. and Woldesenbet, K. (2006), “The dynamicsof product and process innovation in UK banking”,International Journal of Financial Services Management,Vol. 1 No. 4, pp. 400-421.
Battor, M. and Battour, M. (2010), “The impact of customerrelationship management capability on innovation andperformance advantages: testing a mediated model”,Journal of Marketing Management, Vol. 8 No. 26,pp. 842-857.
Beckmann, C., Otto, H.U., Shaarschuch, A. and Shrödter, M.(2007), “Quality management and formalization in socialservice organizations: a survey on home-based family
intervention services”, Social Work and Society, Vol. 5 No. 1,pp. 78-93.
Benner, M.J. and Tushman, M.L. (2003), “Exploitation,exploration, and process management: the productivitydilemma revisited”, Academy of Management Review,Vol. 28 No. 2, pp. 238-256.
Berson, Y., Shamir, B., Avolio, B.J. and Popper, M. (2001),“The relationship between vision strength, leadership style,and context”, The Leadership Quarterly, Vol. 12 No. 1,pp. 53-73.
Birkinshaw, J., Hamel, G. and Mol, M.J. (2008),“Management Innovation”, Academy of ManagementReview, Vol. 39 No. 4, pp. 825-845.
Bodewes, W.E.J. (2002), “Formalization and innovationrevisited”, European Journal of Innovation Management,Vol. 5 No. 4, pp. 214-223.
Cannon, J.P. and Perreault, W. (1999), “Buyer–sellerrelationships in business markets”, Journal of MarketingResearch, Vol. 36 No. 4, pp. 439-460.
Cardinal, L.B. (2001), “Technological innovation in thepharmaceutical industry: the use of organizational control inmanaging research and development”, Organization Science,Vol. 12 No. 1, pp. 19-36.
Chang, S. and Lee, M. (2007), “A study on relationshipamong leadership, organizational culture, the operation oflearning organization and employees’ job satisfaction”, TheLearning Organization, Vol. 14 No. 2, pp. 155-185.
Chen, I. and Fung, P. (2013), “Relationship configurations inthe apparel supply chain”, Journal of Business & IndustrialMarketing, Vol. 28 No. 4, pp. 303-316.
Chin, W.W. (1998), “The partial least squares approach tostructural equation modelling”, in Marcoulides, G.A. (Ed.),Modern Methods for Business Research, Lawrence ErlbaumAssociates, Mahwah, NJ, pp. 295-358.
Cleveland, H. (2000), “The Donald C. Stone Lecture”, inThe Political Environment of Public Management, 2nd ed.,Longman, New York, NY, pp. 397-404.
Cosh, A., Fu, X. and Hughes, A. (2012), “Organizationalstructure and innovation performance in differentenvironments”, Small Business Economics, Vol. 39 No. 2,pp. 301-317.
Covin, J.G. and Slevin, D.P. (1988), “The influence oforganization structure on the utility of an entrepreneurialtop management style”, Journal of Management Studies,Vol. 22 No. 4, pp. 217-234.
Damanpour, F. (1991), “Organizational innovation: ameta-analysis of effects of determinants and moderators”,Academy of Management Journal, Vol. 34 No. 3,pp. 555-590.
De Wulf, K., Odekerken-Schroder, G.J. and Iacobucci, D.(2001), “Investments in consumer relationships: across-country and cross-industry exploration”, Journal ofMarketing, Vol. 65 No. 4, pp. 33-50.
Deshpande, R., Farley, J.U. and Webster, F.E. Jr. (1993),“Corporate culture, customer orientation, andinnovativeness in Japanese firms: a quadrad analysis”,Journal of Marketing, Vol. 57 No. 1, pp. 23-27.
Dess, G. and Davis, P. (1984), “Porter’s (1980) genericstrategies as determinants of strategic group membershipand organizational performance”, Academy of ManagementReview, Vol. 27 No. 3, pp. 467-488.
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
393
Dess, G.G. and Robinson, R.B. (1984), “Measuringorganizational performance in the absence of objectivemeasures: the case of the privately-held firm andconglomerate business unit”, Strategic Management Journal,Vol. 5 No. 3, pp. 265-273.
Devie, T.J. and Widjaja, D.C. (2012), “The relationshipbetween non-financial performance and financialperformance using balanced scorecard framework: aresearch in cafe and restaurant sector”, International Journalof Innovation, Management and Technology, Vol. 3 No. 5,pp. 614-618.
Doyle, P. and Wong, V. (1997), “Marketing and competitiveperformance: an empirical study”, European Journal ofMarketing, Vol. 32 Nos 5/6, pp. 514-535.
Durán-Vázquez, R., Lorenzo-Valdés, A. and Moreno-Quezada, G.E. (2012), “Innovation and CSR impact onfinancial performance of selected companies in Mexico”,Journal of Entrepreneurship, Management and Innovation,Vol. 8 No. 3, pp. 5-20.
Edmondson, A.C. and Nembhard, I.M. (2009), “Productdevelopment and learning in project teams: the challengesare the benefits”, Journal of Product Innovation Management,Vol. 26, pp. 123-138.
Eggert, A., Thiesbrummel, C. and Deutscher, C. (2014),“Differential effects of product and service innovations onthe financial performance of industrial firms”, Journal ofBusiness Market Management, Vol. 5 No. 1, pp. 380-405.
Fink, R.C., James, W.L., Hatten, K.J. and Bakstran, L.(2008), “Supplier strategies to increase customer purchasesover the duration of customer-supplier relationships”,Journal of Business & Industrial Marketing, Vol. 23 No. 8,pp. 529-543.
Fornell, C. and Cha, J. (1994), “Partial least squares”, inBagozzi, R.P. (Ed.), Advanced Methods of MarketingResearch, Basil Blackwell, Cambridge, pp. 52-78.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structuralequation models with unobservable variables andmeasurement error”, Journal of Marketing Research, Vol. 18No. 1, pp. 39-50.
Froehle, C.M., Roth, A.V., Chase, R.B. and Voss, C.A.(2000), “Antecedents of new service developmenteffectiveness: an exploratory examination of strategicoperations choices”, Journal of Service Research, Vol. 3No. 1, pp. 3-15.
Ganesh, S. and Joseph, J. (2011), “Exploring perceivedorganisational formalisation and performance review systemcomplexity as predictors of executive alienation inperformance review systems”, IIMB Management Review,Vol. 23 No. 4, pp. 197-207.
Garicano, L. (2000), “Hierarchies and the organization ofknowledge in production”, Journal of Political Economy,Vol. 108 ��. 5, pp. 874-904.
Glaveli, N. and Karassavidou, E. (2011), “Exploring apossible route through which training affects organizationalperformance: the case of a Greek bank”, InternationalJournal of Human Resource Management, Vol. 22 No. 14,pp. 2892-2923.
Green, K.W. Jr, Inman, R.A., Brown, G. and Willis, T.H.(2005), “Market orientation: relation to structure andperformance”, Journal of Business & Industrial Marketing,Vol. 20 No. 6, pp. 276-284.
Grönroos, C. (2004), “The relationship marketing process:communication, interaction, dialogue, value”, Journal ofBusiness & Industrial Marketing”, Vol. 19 No. 2, pp. 99-113.
Gupta, S. and Zeithaml, V. (2006), “Customer metrics andtheir impact on financial performance”, Marketing Science,Vol. 25 No. 6, pp. 718-739.
Hair, J.F. Jr, Hult, G.T.M., Ringle, C.M. and Sarstedt, M.(2014), A Primer on Partial Least Squares Structural EquationModeling (PLS-SEM), SAGE Publications, ThousandOaks, CA.
Hallowel, R. (1996), “The relationship of customersatisfaction, customer loyalty and profitability: an empiricalstudy”, International Journal of Services IndustryManagement, Vol. 7 No. 14, pp. 27-42.
Han, S.L. and Sung, H.S. (2008), “Industrial brand value andrelationship performance in business markets – a generalstructural equation model”, Industrial MarketingManagement, Vol. 37 No. 7, pp. 807-818.
Hartline, M., Maxham, J. and McKee, D. (2000), “Corridorsof influence in the dissemination of customer-orientedstrategy to customer contact service employees”, Journal ofMarketing, Vol. 64 No. 2, pp. 35-50.
Henseler, J., Ringle, C.M. and Sinkovics, R.R. (2009), “Theuse of partial least squares path modeling in internationalmarketing”, in Sinkovics, R.R. and Ghauri, P.N. (Eds),Advances in International Marketing, Emerald, Bingley,pp. 277-320.
Ho, V.T. and Wong, S.S. (2007), “Knowing who knows whatand who knows whom: expertise recognition, networkrecognition, and individual work performance”, Journal ofOccupational and Organizational Psychology, Vol. 82 No. 1,pp. 147-158.
Hsu, Y.H. and Fang, W. (2009), “Intellectual capital and newproduct development performance: the mediating role oforganizational learning capability”, Technological Forecastingand Social Change, Vol. 76 No. 5, pp. 664-677.
Huang, L. (2008), “Strategic orientation and performancemeasurement model in Taiwan’s travel agencies”, TheService Industries Journal, Vol. 8 No. 10, pp. 1357-1383.
Hunter, J. (2002), “Improving organizational performancethrough the use of effective elements of organizationalstructure”, Leadership in Health Services, Vol. 15 No. 3,pp. 12-21.
Hutt, M.D., Stafford, E.R., Walker, B.A. and Reinge, P.H.(2000), “Defining the social network of a strategic alliance”,Sloan Management Review, Vol. 41 No. 2, pp. 51-62.
IME GSEVEE (2013), “Financial climate in small business”,The Small Enterprises’ Institute of the HellenicConfederation of Professionals, Craftsmen and Merchants(IME GSEVEE), available at: www.gsevee.gr/meletes-2/258-2013 (accessed 8 October 2013).
IMERISIA (2012), “30% reduction in advertisingexpenditure for the last three years”, Newspaper IMERISIA,available at: www.imerisia.gr (accessed 8 October 2013).
Jansen, J.J.P., Van Den Bosch, F.A.J. and Volberda, H.W.(2006), “Exploratory innovation, exploitative innovation,and performance: effects of organizational antecedents andenvironmental moderators”, Management Science, Vol. 52No. 11, pp. 1661-1674.
Jiménez-Jiménez, D. and Sanz-Valle, R. (2008), “CouldHRM support organizational innovation?”, International
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
394
Journal of Human Resource Management, Vol. 19 No. 7,pp. 1208-1221.
Jones, G.R. (2010), Organizational Theory, Design and Change,6th ed., Pearson Education Inc, Upper Saddle River, NJ.
Kandampully, J. (2002), “Innovation as the core competencyof a service organization: the role of technology, knowledgeand networks”, European Journal of Innovation Management,Vol. 5 No. 1, pp. 18-26.
Kanten, S. and Yaslioglu, M. (2012), “Role of innovation increating customer value in hotel establishments: a study ofmanagers”, The Journal of Faculty of Economics andAdministrative Sciences, Vol. 17 No. 2, pp. 437-449.
Khan, �., Halabi, A. and Masud, Z. (2010), “Empirical studyof the underlying theoretical hypotheses in the balancedscorecard (bsc) model: further evidence from Bangladesh”,Asia-Pacific Management Accounting Journal, Vol. 5 No. 2,pp. 45-73.
Klomp, L. and Van Leeuwen, G. (2001), “Linking innovationand firm performance: a new approach”, InternationalJournal of the Economics of Business, Vol. 8 No. 3,pp. 343-364.
Kogut, B. and Zander, U. (1996), “What do firms do?Coordination, identity and learning”, Organization Science,Vol. 7 No. 5, pp. 502-518.
Kohtamaki, M., Vesalainen, J., Henneberg, S., Naudé, P. andVentresca, M.J. (2012), “Enabling relationship structuresand relationship performance improvement: the moderatingrole of relational capital”, Industrial Marketing Management,Vol. 41 No. 8, pp. 1298-1309.
Kumpikaite, V. and Ciamiene, R. (2008), “New trainingtechnologies developing human resources”, Economics andManagement, Vol. 9 No. 2, pp. 155-159.
Kuo, T.H. (2011), “How to improve organizationalperformance through learning and knowledge?”,International Journal of Manpower, Vol. 32 Nos 5/6,pp. 581-603.
Lee, H. and Choi, B. (2003), “Knowledge managementenablers, processes, and organizational performance: anintegrative view and empirical examination”, Journal ofManagement Information Systems, Vol. 20 No. 1,pp. 179-228.
Lee, K., Chun, K. and Lee, J. (2008), “Reforming the hospitalservice structure to improve efficiency”, Health Policy,Vol. 87 No. 5, pp. 41-49.
Lee, T.H. (2012), “Transformational leadership andjob-related learning”, Management Research Review, Vol. 35Nos 3/4, pp. 192-205.
Liang, C., Wang, W. and Farquhar, J.D. (2009), “Theinfluence of customer perceptions on financial performancein financial services”, International Journal of BankMarketing, Vol. 27 No. 2, pp. 129-149.
Lloria, �.�. (2007), “Differentiation in knowledge-creatingorganizations”, International Journal of Manpower, Vol. 28No. 8, pp. 674-693.
López, S., Peón, J.M. and Ordás, C.J. (2006), “Theorganisational context of learning: an empirical analysis”,International Journal of Technology Management, Vol. 35Nos 1/2/3/4, pp. 196-223.
Lunenburg, F.C. (2012), “Organizational structure:Mintzberg’s framework”, International Journal of Scholarly,Academic, Intellectual Diversity, Vol. 15 No. 1, pp. 1-8.
Martínez-León, I.M. and Martínez-Garcia, J.A. (2011), “Theinfluence of organizational structure on organizationallearning”, International Journal of Manpower, Vol. 32Nos 5/6, pp. 537-566.
Mathuramaytha, C. and Ussahawanitchakit, P. (2008),“What is a guide for firms? The creation of customer value”,International Journal of Business Strategy, Vol. 8 No. 2,pp. 128-136.
Meijaard, J., Brand, M.J. and Mosselman, M. (2005),“Organizational structure and performance in Dutch smallfirms”, Small Business Economics, Vol. 25 No. 2, pp. 97-115.
Meirovich, G., Brender-Ilan, Y. and Meirovich, A. (2007),“Quality of hospital service: the impact of formalization anddecentralization”, International Journal of Health CareQuality Assurance, Vol. 20 No. 3, pp. 240-252.
Miller, N.J., Besser, T. and Malshe, A. (2007), “Strategicnetworking among small businesses in small UScommunities”, International Small Business Journal, Vol. 25No. 6, pp. 631-665.
Mink, O.G., Owen, K.Q. and Mink, B.P. (1993), DevelopingHigh-Performance People: The Art of Coaching,Addison-Wesley Publishing Company, Reading, MA.
Mintzberg, H. (1983), Structure in Fives: Designing EffectiveOrganizations, Prentice Hall, Englewood Cliffs, NJ.
Mirkamali, S.M., Thani, F.N. and Alami, F. (2011),“Examining the role of transformational leadership and jobsatisfaction in the organizational learning of an automotivemanufacturing company”, Social and Behavioural Sciences,Vol. 29, pp. 138-148.
Mitręga, M. (2012), “Network partner knowledge andinternal relationships influencing customer relationshipquality and company performance”, Journal of Business &Industrial Marketing, Vol. 27 No. 6, pp. 486-496.
Muiruri, J.M. and Ngari, J.M. (2014), “effects of financialinnovations on the financial performance of commercialbanks in Kenya”, International Journal of Humanities andSocial Science, Vol. 4 No. 7, pp. 51-57.
Nielsen, B.B. (2005), “The role of knowledge embeddednessin the creation of synergies in strategic alliances”, Journal ofBusiness Research, Vol. 58 No. 9, pp. 1194-1204.
Norman, P. (2004), “Knowledge acquisition, knowledge loss,and satisfaction in high technology alliances”, Journal ofBusiness Research, Vol. 57 No. 6, pp. 610-619.
Olson, E.M., Slater, S.F. and Hult, T.M. (2005), “Theperformance implications of fit among business strategy,marketing organization structure, and strategic behavior”,Journal of Marketing, Vol. 69 No. 3, pp. 49-65.
Palmatier, R.W., Dant, R., Grewal, D. and Evans, K. (2006),“Factors influencing the effectiveness of relationshipmarketing: a meta-analysis”, Journal of Marketing, Vol. 70No. 4, pp. 136-153.
Palmer, I. and Dunford, R. (2002), “Who says change can bemanaged? Positions, perspectives and problematics”,Strategic Change, Vol. 11 No. 5, pp. 243-251.
Pawar, B.S. and Eastman, K.K. (1997), “The nature andimplications of contextual influences on transformationalleadership: a conceptual examination”, Academy ofManagement Review, Vol. 22 No. 1, pp. 80-109.
Perry-Smith, J.E. and Shalley, C.E. (2003), “The social sideof creativity: a static and dynamical social networkperspective”, Academy of Management Review, Vol. 28No. 1, pp. 89-108.
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
395
Pertusa-Ortega, E.M., Molina-Azorín, J.F. andClaver-Cortés, E. (2010), “Competitive strategy, structureand firm performance: a comparison of the resource-basedview and the contingency approach”, Management Decision,Vol. 48 No. 8, pp. 1282-1303.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y. and Podsakoff,N.P. (2003), “Common method biases in behaviouralresearch: a critical review of the literature andrecommended remedies”, Journal of Applied Psychology,Vol. 88 No. 5, pp. 879-903.
Prajogo, D., Power, D. and Sohal, A. (2004), “The role oftrading partner relationships in determining innovationperformance: an empirical examination”, European Journalof Innovation Management Vol. 7 No. 3, pp. 178-186.
Purcell, J., Hutchinson, S., Kinnie, N., Rayton, B. andBoulin, P.V. (2003), Understanding the Pay and PerformanceLink: Unlocking the Black Box, CIPD, London.
Rajagopal and Sanchez, R. (2005), “Analysis of customerportfolio and relationship management models: bridgingmanagerial dimensions”, Journal of Business & IndustrialMarketing, Vol. 20 No. 6, pp. 307-316.
Ringle, C.M., Wende, S. and Will, A. (2005), SmartPLS 2.0(M3) beta, University of Hamberg, Hamberg, available at:www.smartpls.de (accessed 11 June 2012).
Ritter, T. (2007), “A framework for analyzing relationshipgovernance”, Journal of Business & Industrial Marketing,Vol. 22 No. 3, pp. 196-201.
Roberts, P.W. and Amit, R. (2003), “The dynamics ofinnovative activity and competitive advantage: the case ofAustralian retail banking, 1981 to 1995”, OrganizationScience, Vol. 14 No. 2, pp. 107-122.
Robertson, M. and Swan, J. (2003), “Control-what control?Culture and ambiguity within a knowledge intensive firm”,Journal of Management Studies, Vol. 40 No. 4, pp. 831-858.
Rogelberg, S.G. and Stanton, J.M. (2007), “Understandingand dealing with organizational survey nonresponse”,Organizational Research Methods, Vol. 10 No. 2,pp. 195-209.
Ryssel, R., Ritte, T. and Gemunden, H.G. (2004), “Theimpact of information technology deployment on trust,commitment and value creation in business relationships”,The Journal of Business & Industrial Marketing, Vol. 19No. 3, pp. 197-207.
Salim, I.M. and Sulaiman, M. (2011), “Organizationallearning, innovation and performance: a study of Malaysiansmall and medium sized enterprises”, International Journalof Business and Management, Vol. 6 No. 12, pp. 118-125.
Sarstedt, M., Ringle, C., Smith, D., Reams, R. and Hair, J. Jr.(2014), “Partial least squares structural equation modeling(PLS-SEM): a useful tool for family business researchers”,Journal of Family Business Strategy, Vol. 5 No. 1, available at:doi.org/10.1016/j.jfbs.2014.01.002
Schellhase, R., Hardock, P. and Ohlwein, M. (2000),“Customer satisfaction in business-to-business marketing:the case of retail organizations and their suppliers”, Journalof Business & Industrial Marketing, Vol. 15 Nos 2/3,pp. 106-121.
Schilling, M.A., Vidal, P., Ployhart, R.E. and Marangoni, A.(2003), “Learning by doing something else: variation,relatedness, and the learning curve”, Management Science,Vol. 49 No. 1, pp. 39-56.
Sheremata, W.A. (2000), “Centrifugal and centripetal forcesin radical new product development under time pressure”,Academy of Management Review, Vol. 25 No. 2,pp. 389-408.
Sheth, J.N. (2001), “Competitive advantage throughcustomer satisfaction”, Bombay Management AssociationReview, Vol. 18 No. 1, pp. 13-25.
Shimpton, H., Fay, D., West, M., Patterson, M. and Birdi, K.(2005), “Managing people to promote innovation”,Creativity and Innovation Management, Vol. 14 No. 2,pp. 118-128.
Shimpton, H., West, M.A., Dawson, J., Birdi, K. andPatterson, M. (2006), “HRM as a predictor of innovation”,Human Resource Management Journal, Vol. 16 No. 1,pp. 3-27.
Spanos, Y., Prastacos, G. and Papadakis, V. (2001), “Greekfirms and EMU: contrasting SMEs and large-sizedenterprises”, European Management Journal, Vol. 19 No. 6,pp. 638-648.
Stacey, R.D. (2003), Strategic Management and OrganizationalDynamics, Prentice Hall, London.
Storbacka, K. and Nenonen, S. (2009), “Customerrelationships and the heterogeneity of firm performance”,Journal of Business & Industrial Marketing, Vol. 24 Nos 5/6,pp. 360-372.
Subramanian, A. and Nilakanta, S. (1996), “Organizationalinnovativeness: exploring the relationship betweenorganizational determinants of innovation, types ofinnovations, and measures of organizational performance”,Omega, Vol. 24 No. 6, pp. 631-647.
Susan, C., Benton, J.M., Laudal, T. and Thomas, R.J.(2006), “Measuring the value of human capitalinvestments: the SAP case”, Strategy & Leadership, Vol. 34No. 2, pp. 43-55.
Swart, J. and Kinnie, N. (2003), “Sharing knowledge inknowledge-intensive firms”, Human Resource ManagementJournal, Vol. 13 No. 2, pp. 60-75.
Tenenhaus, M., Esposito Vinzi, V., Chatelin, Y.M. andLauro, C. (2005), “PLS path modeling”, ComputationalStatistics & Data Analysis, Vol. 48 No. 1, pp. 159-205.
TFGR (2012), “The 2nd quarterly report of the EuropeanCommission’s task force for Greece, Task Force for Greece(TFGR)”, available at: www.eubusiness.com/europe/greece/task-force (accessed 5 May 2012).
Tomaskovic-Devey, D., Leiter, J. and Thompson, S. (1994),“Organizational survey nonresponses”, AdministrativeScience Quarterly, Vol. 39 No. 3, pp. 439-457.
Trautmann, K., Maher, J.K. and Motley, D.G. (2007),“Learning strategies as predictors of transformationalleadership: the case of nonprofit managers”, Leadership &Organization Development Journal, Vol. 28 No. 3,pp. 269-287.
Ulaga, W. (2003), “Capturing value creation in businessrelationships: a customer perspective”, Industrial MarketingManagement, Vol. 32 No. 8, pp. 677-693.
Ulaga, W. and Eggert, A. (2006), “Relationship value andrelationship quality”, European Journal of Marketing, Vol. 40Nos 3/4, pp. 311-327.
Valtakoski, A. (2015), “Initiation of buyer–seller relationships:the impact of intangibility, trust and mitigation strategies”,Industrial Marketing Management, Vol. 44, pp. 107-118.
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
396
Waller, D.S., Cusick, D.J., Matheson, H.D. and Miller, M.K.(2001), “Advertising agency activities used to attract newclients in Australia”, Journal of Business & IndustrialMarketing, Vol. 16 No. 2, pp. 129-140.
Wally, S. and Baum, J.R. (1994), “Personal and structuraldeterminants of the pace of strategic decision making”,Academy of Management Journal, Vol. 37 No. 4,pp. 932-956.
Walter, A. and Gemünden, H.G. (2000), “Bridging the gapbetween suppliers and customers through relationshippromoters: theoretical considerations and empiricalresults”, Journal of Business & Industrial Marketing, Vol. 15Nos 2/3, pp. 86-105.
Walter, A., Müller, T.A., Helfert, G. and Ritter, T. (2003),“Functions of industrial supplier relationships and theirimpact on relationship quality”, Industrial MarketingManagement, Vol. 32 No. 2, pp. 159-169.
Walter, A. and Ritter, T. (2003), “The influence ofadaptations, trust, and commitment on value-creatingfunctions of customer relationships”, Journal of Business &Industrial Marketing, Vol. 18 Nos 4/5, pp. 353-365.
Wu, S. and Hung, J. (2008), “A performance evaluationmodel of CRM on non-profit organizations”, Total QualityManagement, Vol. 19 No. 4, pp. 321-342.
Yiing, L.H. and Ahmad, K.Z.B. (2009), “The moderatingeffects of organizational culture on the relationshipsbetween leadership behaviour and organizationalcommitment and between organizational commitment andjob satisfaction and performance”, Leadership &Organization Development Journal, Vol. 30 No. 1, pp. 53-86.
Zabala, I., Panadero, G., Gallardo, L.M., Amate, C.M.,Sánchez-Galindo, Tena, M.I. and Villalba, I. (2005),“Corporate reputation in professional services firms:reputation management based on intellectual capitalmanagement”, Corporate Reputation Review, Vol. 8 No. 1,pp. 59-71.
Zanko, M., Badham, R., Couchman, P. and Schubert, M.(2008), “Innovation and HRM: absences and politics”, TheInternational Journal of Human Resource Management,Vol. 19 No. 4, pp. 562-581.
Zheng, W., Yang, B. and McLean, G.N. (2010), “Linkingorganizational culture, structure, strategy, andorganizational effectiveness: mediating role of knowledgemanagement”, Journal of Business Research, Vol. 63 No. 7,pp. 763-771.
Zollo, M.M. and Winter, S.G. (2002), “Deliberate learningand the evolution of dynamic capabilities”, OrganizationScience, Vol. 13 No. 3, pp. 339-351.
Further reading
Chen, C., Huang, J. and Hsiao, Y. (2010), “Knowledgemanagement and innovativeness: the role of organizationalclimate and structure”, International Journal of Manpower,Vol. 31 No. 8, pp. 848-870.
Marques, D.P., Narangajavana, Y. and Simon, F.J. (2005),“Competitive strategies and performance in Spanishhospitality firms”, International Journal of ContemporaryHospitality Management, Vol. 17 No. 2, pp. 22-38.
About the authors
Paraskevi Dekoulou holds a Phd degree in MediaManagement from the Aristotle University of Thessaloniki.She received a MSc degree in Media Management fromStirling University (Scotland) and an MBA degree fromStaffordshire University (UK). She has taught MediaManagement, Innovation and Entrepreneurship, CulturalManagement and Advertising at the Aristotle University ofThessaloniki as well as at the Alexander TechnologicalEducational Institute of Thessaloniki. She has publishedarticles in The International Journal of Media Management,the Journal of Media Business Studies and the Journal ofApplied Journalism and Media Studies. In parallel with heracademic activity, she has worked as a Marketing Officer inthe private sector. She is currently working as a StrategicDevelopment Manager at the Centre for Innovation,Entrepreneurship and Employment (KEA), part of theUniversity of Nicosia.
Panagiotis Trivellas is a Professor of Strategic HumanResources Management and Deputy Head of LogisticsManagement Department at the Technological EducationInstitute (TEI) of Central Greece. He holds a PhD degreein Management Styles from the National TechnicalUniversity of Athens (NTUA). He has been a ScientificCoordinator and Head Researcher in several researchprojects in the field of organizational behavior, leadershipand HRM. He has published more than 35 research papersin academic journals and international conferences and hewas a co-author of two books. His core research interestsfocus on the areas of OB, HRM, corporate culture, strategicmanagement and leadership, managerial skills andcompetences, TQM and quality in higher education.Panagiotis Trivellas is the corresponding author and can becontacted at: ptriv@tee.gr
For instructions on how to order reprints of this article, please visit our website:www.emeraldgrouppublishing.com/licensing/reprints.htmOr contact us for further details: permissions@emeraldinsight.com
Greek advertising and media industry
Paraskevi Dekoulou and Panagiotis Trivellas
Journal of Business & Industrial Marketing
Volume 32 · Number 3 · 2017 · 385–397
397
Organizational structure, innovation performance and customer relationship value in the Greek ad …
1. Introduction
2. Theoretical background
3. Research methodology
4. Discussion
5. Conclusions and managerial implications
6. Limitations
References
Journal for the Study of Religion 31,1 (2018) 41 – 53 41 On-line ISSN 2413-3027; DOI: http://dx.doi.org/10.17159/2413-3027/2018/v31n1a2
World Religions in the World
David Chidester
david.chidester@uct.ac.za
Abstract The classification of ‘world religions’ is highly problematic because of its
arbitrary construction, its exclusion of indigenous religions, and its easy
availability for ideological manipulation. The imperial edifice of ‘world
religions’ has been dismantled in recent scholarship in the study of religion.
Yet, the notion of ‘world religions’ has been enthusiastically embraced by
advocates of inclusive citizenship in democratic societies and by advocates of
indigenous empowerment in postcolonial societies. This brief essay reviews
the terms of engagement for critically reflecting on the various deployments of
‘world religions’ as a prelude to thinking about religion in the world.
Keywords: world religions, indigenous religions, ideology, inclusive citizen-
ship, democracy, indigenous empowerment, postcolonial societies, religion in
the world
I have been researching and teaching about religion and religions in South
Africa since 1984. I confess that when I first arrived I thought that the study of
religion in the country was underveloped, except for the work of Martin
Prozesky, who besides leading a professional association and editing a peer-
reviewed journal was developing research in explaining religion (Prozesky
1984), critically analyzing the entanglements of Christianity with apartheid
(Prozesky 1990), profiling the emergent field in the region (Prozesky 1996),
and providing textbook resources for the classroom (Prozesky & De Gruchy
1991). While his explanatory theory accounted for religions of the world as
ways of maximizing human well-being, his teaching also focused on world
David Chidester
42
religions in South Africa. In this brief essay in tribute to Martin Prozesky, I
want to reflect on the ongoing importance of the study of religion and religions
– even the study of ‘world religions’ – against the background of what I have
learned in South Africa.
Why is the study of religion important? If we were only interested in
writing advertising copy or generating propaganda for the academic study of
religion, we might advance this circular argument: Like politics, economics,
music, or literature, religion is an important and pervasive human activity.
Therefore, teaching and learning about such an activity must obviously be
important if we want to know about human beings. This argument is circular
because it assumes that the study of something is important because the thing
is important. But it tells us nothing about the importance of the study, about its
distinctive value proposition. What does this field of teaching and learning
bring to the party?
As we know, there are many answers to this question, which is
something I like about the academic study of religion. This field is resistant to
any orthodoxy. Many voices can be heard. Many positions can contend. This
multiplicity of perspectives and positions, I am convinced, is a strength rather
than a weakness of the academic study of religion.
Nevertheless, positioning myself, I have found that our key terms –
religion and religions – are not merely objects for study. They are occasions
for critical and creative reflection on problems of interpretation, explanation,
and analysis in the humanities and social sciences. The study of religion, as I
understand it, is a critical and creative enterprise. While the criticism of
religion, as Karl Marx proposed, is the beginning of all criticism, the creative
enterprise of imagining religion as a human project opens new possibilities
for understanding a diverse array of powerful discourses, practices, and
social formations that are underwritten by claims on transcendence or the
sacred.
In my teaching, I dwell in the ambiguity of the very word, ‘religion’. I
focus on boundary situations. I concentrate on situations in which the
designation has been denied to alternative religious movements in the United
States (Chidester 1988a) or to indigenous religions in southern Africa
(Chidester 1996). By contrast, I also focus on situations in which the term has
been extended to include the ultimate commitments of modern nationalisms
(Chidester 1998b) or the production, circulation, and consumption of popular
culture (Chidester 2005). Accordingly, I find that the term, ‘religion’, is an
World Religions in the World
43
enabling term, because it allows for critical and creative reflection on crucial
problems of inclusion and exclusion that have both intellectual and social
consequences.
The term ‘religions’ poses a related set of problems. How many are
there? In principle, their number might be indeterminate and innumerable, but
their classification bears traces of particular kinds of social projects. In trying
to conceptualize, contain, and perhaps even manage this diversity, European
and Euro-American scholars during the nineteenth century came up with the
notion of ‘world religions’. We live with that legacy. What do we do about it?
In teaching and learning about religions, we must critically interrogate
the historical conditions that have produced the classification of ‘world
religions’. This critical reflection, however, cannot be an end in itself. Against
this background, we still need to find ways of creatively engaging,
understanding, and explaining the discourses and forces that move and
motivate people, personally and collectively, religiously.
Although the classification of ‘world religions’, as I will suggest, is
highly problematic because of its arbitrary construction, its exclusion of
indigenous religions, and its easy availability for being manipulated by agents
of various imperial projects, we must also recognize that the notion of ‘world
religions’ has also been enthusiastically embraced by advocates of inclusive
citizenship in democratic societies and by advocates of indigenous empower-
ment in postcolonial societies.
Briefly, I review this history of the notion of ‘world religions’, not as
if recounting this history were an end in itself, but in the interest of advancing
efforts to clear ground and open space for teaching and learning about religion,
religions, and religious diversity.
In his series of lectures delivered in 1870 on the science of religion,
Friedrich Max Müller, who is often regarded as the founder of the academic
study of religion, saw his primary task as classification. Taking as his motto
the aphorism, divide et impera, which he rendered ‘classify and conquer’, Max
Müller proceeded to classify the major religions of the world into three
language groups, the Semitic (Judaism, Christianity, and Islam), the Aryan
(Hinduism, Buddhism, and Zoroastrianism), and the Turanian (Confucianism
and Taoism). Although he did not use the phrase, ‘world religions’, Max
Müller nevertheless argued that these eight religions comprised the ‘library of
the sacred books of the world’. These textual traditions, with their sacred books
and interpretive communities, could be regarded as a library, a religious
David Chidester
44
archive that could be organized, like any library, according to a general system
of classification.
Any modern library, however, whatever system of classification it
employs, whether Dewey Decimal, Library of Congress, or some other system,
must be all-inclusive. Anything and everything must fit somewhere. But
Friedrich Max Müller’s ‘library of sacred books of the world’ was organized
by a system of classification, as he quickly admitted, which left out most of the
religious life of the world. In his library of eight religions, Müller
observed,
The largest portion of mankind, – ay, and some of the most valiant
champions in the religious and intellectual struggles of the world,
would be unrepresented in our theological library (1873: 116).
F. Max Müller’s classification of religions, therefore, was based on a very
peculiar system that left out many – if not most – of the religious struggles of
the world.
Although the phrase has become conventional, the notion of ‘world
religions’ is actually a very strange construction. As deep background, it arose
out of medieval Christian reflections on the variety of religious laws or sects
(Biller 1984; Bossy 1982), medieval Christian travel accounts of strange
beliefs and customs (for example, Mandeville 1900; see Chidester 2000: 335-
340), the early modern Christian ‘wars of religion’ that made religion a highly
charged marker of political difference in Europe (Holt 1995), and the
commercial expansion into the Atlantic and Pacific worlds that made religion
a highly charged marker of human value outside of Europe (Chidester 1996;
2000:353-490). But European scholars in the late nineteenth century trans-
formed these reflections on difference and encounters with diversity into a
science of religion. Raising basic questions, which might have arisen from
intellectual curiosity, about human identity and difference, the notion of
separate and distinct religions of the world was integrated into European
political projects in forging identities based on race, language, and territory
(Masuzawa 2005). This construction of ‘world religions’ capitalized on the
ambiguity inherent in the ancient Latin term, religio, which could refer to either
personal faith or public ritual. Within the classificatory system of ‘world
religions’, personal subjectivity could be defined as symptomatic of adherence
to a religious collectivity. Accordingly, people all over the world could be
World Religions in the World
45
classified as if their identity, subjectivity, and agency were determined by their
religions.
However, as even Max Müller recognized, this system of classification
was not adequate, although its inadequacy was not merely its inherent bias
towards textual traditions. More seriously and substantially, I would argue, the
very notion of ‘world religions’ failed to account for religion, religions, and
religious diversity in the world because it was arbitrary, exclusionary, and
immediately available for ideological manipulation.
First, the framework of ‘world religions’ is completely arbitrary. How
many religions, we might ask, are there in the world? In the 1590s, when the
word, ‘religions’, first appeared in the English language, there were two,
Protestant and Catholic (Harrison 1990: 39). During the eighteenth century,
there were four, Christianity, Judaism, Islam, and Paganism (Pailin 1984).
Following Max Müller’s identification of eight religions in 1870, as the study
of religion developed in the twentieth century the list of major ‘world religions’
was altered on account of contingent historical factors to remove Zoro-
astrianism and add Shintoism. Although a recent survey has identified thirty-
three principal ‘world religions’ (Eliade et al. 2000), common usage of the
framework has generally settled on a kind of G8 of major religions in the world.
The gradual increase in the number of recognized religions in the
world might suggest an expanding scope of human recognition. But all of these
accountings have been based on arbitrary definitions of indeterminate
diversity. The arbitrary construction of ‘world religions’ is immediately
revealed by considering global demographics. Indigenous African religion, in
all of its variety, is a powerful and pervasive religious force in the world.
According to statistics compiled by Adherents.com, African traditional and
diasporic religions account for the religious affiliation of 100 million people,
ranking eighth in this website’s profile of the ‘major religions of the world’
(Adherents.com 2005). Yet African religion never appears on any conventional
list of world religions (see Baum 2005; Lewis 1990).
Second, as the example of African indigenous and diasporic religion
suggests, the framework of ‘world religions’ is exclusionary. By privileging
the religions that emerged from urban, agricultural civilizations of the Middle
East, India, and the Far East, the model of ‘world religions’ implicitly excludes
all forms of indigenous religious life. Max Müller’s library of religions, as he
recognized, could only be secured by first factoring out the ‘religious struggles’
of indigenous people all over the world. When not ignored entirely, as they
David Chidester
46
often are, indigenous religions are incorporated in introductory textbooks to
‘world religions’ as traces of origins and absences. They might no longer
register as ‘savage’ or ‘primitive’, but indigenous religions have been
classified as ‘primal religions’ (Smart 1996; Smith 1994; Richards 1997) or
‘basic religions’ (Hopfe & Woodward 2007), suggesting that they represent the
earliest, simplest point of origin transcended by major world religions, an
impression reinforced by including them in a consideration of ‘primal and
bygone religions’ (Noss 2003). A similar exclusion of indigenous religions is
suggested by classifying them as ‘religions of nature’ or ‘nature religions’
(Elwood & McGraw 2005; Kung & Kuschel 1995), which risks suggesting that
they belong in the natural rather than cultural world, or by classifying them as
‘tribal religions’ (Carmody & Brink 2006; Matthews 2003) that implicitly have
no place within a world comprising civilizations, nations, and especially
modern states. Sometimes introductory texts have defined indigenous religions
simply as an absence, as in the category, ‘non-literate religions’ (Coogan
1998), as if an entire category could be defined by what it lacks, or by a mix of
natural origin and cultural absence, as in the category, ‘nonscriptural nature
religions’ (Toropov & Buckles 1997), but the model of ‘world religions’ has
struggled with finding terminology for indigenous religions because it is
premised on their exclusion.
Although we might assume that the phrase, ‘world religions’, stands
in contrast to either non-religion or religions from other planets, it actually
seems to operate in opposition to the indigenous religions of colonized people
all over the world. In general surveys of ‘world religions’, indigenous religions
are rarely referred to as ‘indigenous’ (although see Chidester 2002; Fisher
2006; Ludwig 2006). As William Pietz has observed, that term would imply
‘the right to land, territories, and place’ associated with the kind of indigenous
national autonomy asserted by the International Covenant on the Rights of
Indigenous Nations (Pietz 1999: 7-8; Martin & Stahnke 1998: 133-37). By
rendering indigenous religions a residual category, the framework of ‘world
religions’ excludes any consideration of such claims to indigenous identity and
place in the world. Accordingly, it might be argued that the very notion of
‘world religions’ emerged as part of a larger project to exclude such indigenous
claims.
Third, the framework of ‘world religions’ is readily available for the
ideological work of asserting conceptual control over the entire world. In the
case of Friedrich Max Müller, following from his guiding aphorism, ‘classify
World Religions in the World
47
and conquer’, the division of the world into ‘world religions’ promised
conceptual control over religious diversity in the service of the British imperial
project (Chidester 2004; see Chidester 2014). Arguably, recent systems of
classification, such as Samuel Huntington’s nine ‘world civilizations’, which
can be easily mapped as ‘world religions’, continues this ideological work of
asserting global conceptual control (Huntington 1993; 1998). Organized within
the framework of ‘world religions’, clashing civilizations can be not only
understood but also managed from the imperial center.
Certainly, we can find evidence of such imperial use of the idea of
‘world religions’. In the middle of the nineteenth century, as Great Britain was
expanding its empire, the British theologian F. D. Maurice undertook a study
of world religions, which he justified on the grounds that knowledge about
religions would be useful for a nation that was ‘engaged in trading with other
countries, or in conquering them, or in keeping possession of them’ (Maurice
1847: 255; see Chidester 1996: 131-32). In the middle of the twentieth century,
as the United States was assuming an imperial role in the wake of the collapse
of European empires, American scholar of religion Huston Smith undertook a
study of world religions, which he justified in 1958, based on his experience
of lecturing to officers of the U.S. Air Force, as providing useful knowledge
for military personnel because ‘someday they were likely to be dealing with
the peoples they were studying as allies, antagonists, or subjects of military
occupation’ (Smith 1958: 7-8; see McCutcheon 1997: 180-81).
These recommendations for the study of religion suggest a remarkable
continuity from British imperialism to American neo-imperialism in justifying
the field of study as an intellectual instrument of international trade, military
conquest, and political administration of alien subjects. Such strategic
justifications for the study of religion and religions persist, as we find in the
introductory course, ‘Religious Factors in Special Operations’, offered by
Chaplain Ken Stice at the United States Army John F. Kennedy Special
Warfare Center and School. In the syllabus for this course, Chaplain Stice
identified the ‘terminal learning objective’ as enabling a Special Operations
soldiers to brief their commanders on the impact of religion and religions on a
mission and its forces. ‘Why do Special Operations soldiers need to study
religion at all?’ Chaplain Stice asked. ‘Primarily, because of the truth of
Special Operations Imperative #1: Understand the Operational Environment!’
As an adjunct to military strategy and tactics, therefore, the study of religion
and religions can be useful in gaining the cooperation or submission of
David Chidester
48
adherents of foreign, unfamiliar religions that Chaplain Stice could charac-
terize as ‘different from our own’ (Stice 1997).
But military strategy cannot provide the only rationale for the study of
religion, religions, and religious diversity. As an alternative, we can consider
the rationale provided in a popular text, The Complete Idiot’s Guide to the
World’s Religions, which argues that understanding religions is important for
dealing not with foreign aliens but with fellow citizens in a diverse society. ‘In
an earlier era’, the authors suggest, ‘unfamiliar religious systems could be
dismissed as ‘foreign’ and left for the scholars to explore’. In a rapidly
changing world of increasing local diversity and expanding global
connectivity, however, learning about religion and religions has become
necessary for everyone, ‘even if you don’t have an advanced degree in
comparative religion’, the authors of the Idiot’s Guide urge, adding the
tantalizing question: ‘Why leave all the excitement to academics?’ (Toporov
& Buckles 1997: 7).
By contrast to the imperial strategy, the Idiot’s Guide announces a
different rationale for studying religion and religions that has emerged under
conditions of increased religious, cultural, and linguistic diversity within urban
centers of the West. Increasingly, people encounter adherents of other religions
not only in international business, military operations, or foreign missions but
also at home. As the Idiot’s Guide explains, ‘At one point or another, just about
everyone has felt some form of anxiety about encountering an unfamiliar
religious tradition’ (Toropov & Buckles 1997: frontis). Therefore, everyone
needs to learn how to deal with personal feelings of anxiety about the
unfamiliar; to avoid personal embarrassment in dealing with others; and to live
knowledgably, comfortably, and confidently in a multicultural, multireligious
world. Ultimately, the Idiot’s Guide recommends the study of religion and
religions as an antidote to fear of the unknown. ‘Perhaps the most important
reason to study faiths beyond one’s own’, the authors advise, ‘is that it is a
marvelous way to replace fear with experience and insight. It’s hard to be
frightened of something you really understand’ (Toropov & Buckles 1997: 8).
The study of religion and religions, therefore, emerges as a kind of therapy for
fear. ‘The more you know about other faiths’, the authors promise, ‘the less
fear will be a factor in your dealings with people who practice those faiths’
(Toropov & Buckles 1997: 10).
By treating adherents of different religions as local citizens rather than
World Religions in the World
49
as foreign subjects, The Complete Idiot’s Guide to the World’s Religions
represents a significant alternative to the imperial study of religion, suggesting
that the very notion of ‘world religions’ can be interpreted against the grain of
its imperial origin. Recent research on religion in public education in Europe
and Africa has shown that ‘world religions’ can signify different things – an
alienating framework to be rejected, an inclusive framework to be embraced –
depending upon the aims and objectives of specific national projects.
Researchers in Britain or Germany, for example, have found the notion of
‘world religions’ to be an obstacle that has to be overcome through local
ethnography (Jackson 1997) or dialogue (Weisse 1999). By contrast,
researchers in southern Africa, as elsewhere in sub-Saharan Africa, who have
been subjected to specific regimes of Christian establishment, have found the
notion of ‘world religions’ to be an inclusive, empowering avenue for opening
the study of religion, religions, and religious diversity (Chidester 2003). The
inclusion of African indigenous religion, in particular, has been advocated as a
liberating initiative (Mndende 1998; 1999). Therefore, if the notion of ‘world
religions’ has enduring political import, its educational politics is currently
being engaged differently all over the world.
The intellectual construction of ‘world religions’ bears a complex
political history, with its origin in imperial conquest, its mobilization within
pluralistic modern states for liberal tolerance and co-existence, and its more
recent redeployment within various colonized regions of the world for the local
liberation of suppressed communities from oppressive religious discrimina-
tion. In the world, therefore, the framework of ‘world religions’ is a contested
terrain. Fortunately, the academic study of religious discourses and practices,
religious subjectivities and collectivities, religious traditions and interactions
does not depend upon any notion of ‘world religions’. However, as long as
traces of this notion arise, whether in pedagogical practice or national policy,
critical reflection on the historical emergence and various deployments of the
notion of ‘world religions’ will be useful in clearing the ground for thinking
about religion in the world.
References Adherents.com 2005. Major Religions of the World. Available at:
http://www.adherents.com/Religions_By_Adherents.html (Accessed on
20 December 2005.)
David Chidester
50
Baum, R.M. 2005. The Forgotten South: African Religions in World Religions
Textbooks. Religious Studies Review 31,1-2: 27-30.
Biller, P. 1985. Words and the Medieval Notion of ‘Religion’. Journal of
Ecclesiastical History 36: 351-369.
Bossy, J. 1982. Some Elementary Forms of Durkheim. Past and Present 95:
3-18.
Carmody, D.L. & T.L. Brink 2006. Ways to the Center: An Introduction to
World Religions. 6th Edition. Belmont, Calif.: Wadsworth.
Chidester, D. 1998a. Salvation and Suicide: An Interpretation of Jim Jones,
the Peoples Temple, and Jonestown. Bloomington: Indiana UP.
Chidester, D. 1998b. Patterns of Power: Religion and Politics in American
Culture. Englewood Cliffs, NJ: Prentice Hall.
Chidester, D. 1996. Savage Systems: Colonialism and Comparative Religion
in Southern Africa. Charlottesville: University Press of Virginia.
Chidester, D. 2000. Christianity: A Global History. London & San Francisco:
Penguin & Harper Collins.
Chidester, D. 2002. Patterns of Transcendence: Religion, Death, and Dying.
2nd Edition. Belmont, Calif. Wadsworth.
Chidester, D. 2003. Global Citizenship, Cultural Citizenship, and World
Religions in Religion Education. In Jackson, R. (ed.): International
Perspectives on Citizenship, Education, and Religious Diversity. London:
Routledge.
Chidester, D. 2004. ‘Classify and Conquer’: Friedrich Max Müller, Indigenous
Religious Traditions, and Imperial Comparative Religion. In Olupona,
J.K. (ed.): Beyond Primitivism: Indigenous Religious Traditions and
Modernity. London and New York: Routledge.
Chidester, D. 2005. Authentic Fakes: Religion and American Popular Culture.
Berkeley: University of California Press.
Chidester, D. 2014. Empire of Religion: Imperialism and Comparative
Religion. Chicago: University of Chicago Press.
Coogan, M.D. (ed.) 1998. The Illustrated Guide to World Religions. New
York: Oxford University Press.
Eliade, M., I.P. Culianu & H.S. Wiesner 2000. The Harper Collins Concise
Guide to World Religions. San Francisco: Harper Collins.
Ellwood, R.S. & B.A. McGraw 2005. Many Peoples, Many Faiths: Women
and Men in the World Religions. 8th Edition. Upper Saddle River, N.J.:
Prentice-Hall.
World Religions in the World
51
Fisher, M.P. 2006. Living Religions. 6th Edition. Upper Saddle River, N.J.:
Prentice-Hall.
Harrison, P. 1990. ‘Religion’ and the Religions in the English Enlightenment.
Cambridge: Cambridge University Press.
Holt, M.P. 1995. The French Wars of Religion 1562-1629. Cambridge:
Cambridge University Press.
Hopfe, L.M. & M.R. Woodward. 2007. Religions of the World. 10th Edition.
Upper Saddle River, N.J.: Prentice-Hall.
Huntington, S. 1993. The Clash of Civilizations? Foreign Affairs 72,3:22-49.
Huntington, S. 1998. The Clash of Civilizations. New York: Simon and
Schuster.
Jackson, R. 1997. Religious Education: An Interpretive Approach. London:
Hodder & Stoughton.
Küng, H. & K-J. Kuschel (eds.). 1995. A Global Ethic: The Declaration of the
Parliament of the World’s Religions. New York: Continuum.
Lewis, J.R. 1990. Images of Traditional African Religions in Surveys of World
Religions. Religion 20: 311-322.
Ludwig, T.M. 2006. The Sacred Paths: Understanding the Religions of the
World. 4th Edition. Upper Saddle River, N.J.: Prentice-Hall.
Mandeville, J. 1900. The Travels of Sir John Mandeville: The Version of the
Cotton Manuscript in Modern Spelling. Pollard, A.W. (ed.): London:
Macmillan.
Martin, J.P. & T. Stahnke (eds.). 1998. Religion and Human Rights: Basic
Documents. New York: Columbia University, Center for the Study of
Human Rights.
Masuzawa, T. 2005. The Invention of World Religions: Or, How European
Universalism was Preserved in the Language of Pluralism. Chicago:
University of Chicago Press.
Matthews, W. 2003. World Religions. 4th Edition. Belmont, Calif.: Wadsworth.
Maurice, F.D. 1847. The Religions of the World and their Relations to
Christianity. London: John W. Parker.
Max Müller, F. 1873. Introduction to the Science of Religion: Four Lectures
Delivered at the Royal Institution with Two Essays on False Analogies,
and the Philosophy of Mythology. London: Longmans, Green.
McCutcheon, R.T. 1997. Manufacturing Religion: The Discourse of Sui
Generis Religion and the Politics of Nostalgia. New York and Oxford:
Oxford University Press.
David Chidester
52
Mndende, N. 1998. From Underground Praxis to Recognized Religion:
Challenges Facing African Religions. Journal for the Study of Religion
11, 2: 115-24.
Mndende, N. 1999. From Racial Oppression to Religious Oppression: African
Religion in the New South Africa. In Walsh, T.G. & F. Kaufmann (eds.):
Religion and Social Transformation in Southern Africa. St. Paul, Minn.:
Paragon House.
Noss, D. S. 2003. History of the World’s Religions. 11th Edition. Upper Saddle
River, N.J.: Prentice-Hall.
Pailin, D.A. 1984. Attitudes to Other Religions: Comparative Religion in
Seventeenth- and Eighteenth-Century Britain. Manchester: Manchester
University Press.
Pietz, W. 1999. Colonial Forensics and Hybrid Rights: Reconceiving Magic,
Religion, and Knowledge to Accommodate an Imperial Market in British
and U.S. Law. Unpublished paper, Ford Foundation lecture series, ‘World
Religions as Colonial Formations’, University of North Carolina, Chapel
Hill.
Prozesky, M. 1984. Religion and Ultimate Well Being: An Explanatory
Theory. London: Macmillan; New York: St Martin’s Press.
Prozesky, M. (ed.) 1990. Christianity Amidst Apartheid. London: Macmillan.
Prozesky, M. 1996. South Africa’s Contribution to Religious Studies. In
Platvoet, J., J. Cox & J. Olupona (eds.): The Study of Religions in Africa:
Past, Present, and Prospects. Cambridge: Roots and Branches.
Prozesky, M. & J. de Gruchy (eds.): 1991. A Southern African Guide to World
Religions. Cape Town: David Philip.
Richards, C. (ed.). 1997. Illustrated Encyclopedia of World Religions.
Rockport, Mass.: Element.
Smart, N. 1996. The Religious Experience. 5th Edition. Upper Saddle River,
N.J.: Prentice-Hall.
Smith, H. 1958. The Religions of Man. New York: Harper and Row.
Smith, H. 1994. The Illustrated World’s Religions: A Guide to our Wisdom
Traditions. New York: HarperCollins.
Stice, K. 1997. Religious Factors in Special Operations. Available at:
http://www-cgsc.army.mil/chap/COURSES/W-REL/STICE/sf-relig.doc
(Accessed on 20 January 2002.)
Toropov, B. & L. Buckles 1997. The Complete Idiot’s Guide to the World’s
Religions. New York: Alpha Books.
World Religions in the World
53
Weisse, W. (ed.) 1999. Vom Monolog zum Dialog: Ansätze einer dialogischen
Religionspädagogik. 2nd. Edition. Münster and New York: Waxmann.
David Chidester
Emeritus Professor
Religious Studies
University of Cape Town
davidc@iafrica.com
Journal for the Study of Religion 31,1 (2018) 330 – 336 330 On-line ISSN 2413-3027
Contributors
Judith Brown (Ph.D., Hon. Soc. Sci.) recently retired from the Beit Chair of
Commonwealth History at Oxford. She is an emeritus Professorial Fellow of
Balliol College. Her main interests lie in modern South Asia, the South Asian
diaspora, and religious change in South Asia. Among her many books and
articles are Gandhi. Prisoner of Hope (1989), Nehru. A Political Life (2003)
Modern India. The Origins of an Asian Democracy (several eds.), Global South
Asians. Introducing the Modern Diaspora (2006), Mahatma Gandhi. The
Essential Writings (2008 in the Oxford World’s Classics), and ed. with A.
Parel, The Cambridge Companion to Gandhi (2011). Contact details: judith.
brown@history.ox.ac.uk
Denzil Chetty is a lecturer in the Department of Religious Studies and Arabic
at the University of South Africa (Unisa). His teaching and research niche
focuses on religion, civil society and technology. In addition over the past years
he has been probing the integration of technology in teaching and learning and
the advancing of religious studies within the digital humanities. In 2003 Chetty
was awarded the Abe Bailey Fellowship to the United Kingdom; in 2014 he
was nominated as the Shanghai Open University Visiting Scholar; and in 2015
he was the recipient of Unisa’s Excellence Award in Teaching and Learning.
He serves as a member of the editorial committee for Alternation –
Interdisciplinary Journal for the Study of Arts and Humanities in Southern
Africa; and a member of the editorial committee for the IAHR book series with
Equinox. Contact details: Chettd@unisa.ac.za
David Chidester is Professor of Religious Studies and Director of the Institute
for Comparative Religion in Southern Africa (ICRSA) at the University of
Cape Town in South Africa. Chidester is the author or editor of over twenty
books in North American studies, South African studies, and comparative
religion. His major publications include Salvation and Suicide: Jim Jones, the
Peoples Temple, and Jonestown (Indiana University Press, 1988; revised
edition 2003); Authentic Fakes: Religion and American Popular Culture
Contributors
331
(University of California Press, 2005); Christianity: A Global History
(Penguin; Harper Collins, 2000); Savage Systems: Colonialism and
Comparative Religion in Southern Africa (University of Virginia Press, 1996);
Wild Religion: Tracking the Sacred in South Africa (University of California
Press, 2012); and Empire of Religion: Imperialism and Comparative Religion
(University of Chicago Press, 2014). Contact details: davidc@iafrica.com
John B. Cobb, Jr. (born February 9, 1925) is an American theologian,
philosopher, and environmentalist. Gary Dorrien has described Cobb as one of
the two most important North American theologians of the twentieth century
(the other being Rosemary Radford Ruether). Cobb is often regarded as the
preeminent scholar in the field of process philosophy and process theology—
the school of thought associated with the philosophy of Alfred North
Whitehead. Cobb is the author of more than fifty books. Contact details:
cobbj@cgu.edu
John W. de Gruchy (D.Th, D.SocSc.) is Emeritus Professor of Christian
Studies at the University of Cape Town and an Extraordinary Professor in the
Faculty of Theology at Stellenbosch University. He is the author and editor of
many books and journal articles, and has lectured around the world. Amongst
his monographs are The Church Struggle in South Africa, Bonhoeffer and
South Africa, Democracy and Christianity, Christianity, Art and Transfor-
mation, Reconciliation: Restoring Justice, Icons as a Means of Grace, Being
Human: Confessions of a Christian Humanist, Led into Mystery, and most
recently I Have Come a Long Way. Contact details: john@degruchy.co.za
Basil Moore is retired from his former position as Professor of Education at
the University of South Australia. He was the lead editor of the collection of
Black theology papers, published under the title Black Theology: The South
African Voice. He was co-author with Norman Habel of the text When Religion
Goes to School, and with Norman Habel, Michael O’donoghue, Marie and
Robert Crotty, of Finding a Way, the Religious Worlds of Today, and with Alan
Reid of Issues in Australian Studies: People and Power and of Teaching for
Resistance. Contact details: basilm@comstech.com
Lloyd Geering was born in Rangiora in 1918, educated chiefly in Otago and
holds Honours degrees in Mathematics and Old Testament Studies. Ordained
Reproduced with permission of copyright owner.Further reproduction prohibited without permission.
Are consumers’ financial needs and values common acrosscultures? Evidence from six countriesLan Guo1, Dan Stone2, Stephanie Bryant3, Benson Wier4, Alex Nikitkov5, Chunyan Ren6,Edson Luiz Riccio7, Milton Shen8, Samir Trabelsi9 and LiFang Zhang10
1School of Business and Economics, Wilfrid Laurier University, Waterloo, ON, Canada2Von Allmen School of Accountancy, University of Kentucky, Lexington, KY, USA3College Of Business, Missouri State University, Springfield, MO, USA4School of Business, Virginia Commonwealth University, Richmond, VA, USA5Brock University, St. Catharines, ON, Canada6Department of Accountancy, Xiamen University, Xiamen, China7School of Economics, Business and Accountancy, University of São Paulo, São Paulo, São Paulo State, Brazil8College of Business Administration, University of Alabama – Huntsville, Huntsville, AL, USA9Brock University, St. Catharines, ON, Canada10School of Business, Nanjing University, Nanjing, China
Keywords
Consumer behaviour, consumer attitudes,financial need beliefs, financial values,multi-group confirmatory factor analysis,self-determination theory.
Correspondence
Dan N. Stone, University of Kentucky,Accountancy, 355F B&E Building, Lexington,KY, USA.E-mail: dstone@uky.edu
doi: 10.1111/ijcs.12047
AbstractAre consumers’ financial needs, and financial values, the same or different across cultures?Two studies, with student (Study 1; n = 988) and non-student (Study 2; n = 959) partici-pants, explore the extent of equivalence, across six countries (Brazil, Russia, China,Taiwan, Tunisia and US), in financial need belief, and financial value, measurementmodels. The financial need beliefs, derived from self-determination theory (SDT) princi-ples, include financial self-efficacy, financial autonomy, financial community trust andsupport; the financial values include materialism and financial altruism. Both the financialneed and financial value constructs evidence configural invariance (similar factor struc-ture), and factor invariance among student but not non-student samples. The financial needconstructs evidence full, and the financial value constructs evidence partial, metric (factorloading) invariance. Factor covariance invariance obtains for the financial need beliefsconstructs but not the financial value constructs. Finally, neither financial need nor financialvalue constructs evidence scalar (intercept) invariance. These results provide partialsupport for extending SDT’s hypothesis of universal human needs to the financial domain.In contrast, the financial value constructs of altruism and materialism are largely instableacross cultures, suggesting that consumer views of giving, and the role of wealth in socialstatus, differ between countries.
IntroductionConsumers’ financial need beliefs, i.e. their beliefs about theirability to manage money, gain financial independence, and trustothers about money, and financial attitudes, i.e. beliefs about altru-ism and materialism, are potentially important influences onchoice and behaviour. For example, believing that one is compe-tent with money, that one’s marketplace choices are volitionalchoices and that one can rely on others for financial help, likelyengenders more rational and competent consumer behaviour thando beliefs that one lacks competence, that one’s marketplacechoices are manipulated and that others are untrustworthy (cf.Moller et al., 2006). Further, the belief that the principle value ofmoney is in contributing to building relationships and communi-ties, vs. that money’s principle value is in gaining power, prestige
and enhanced attractiveness to potential mates would likely influ-ence philanthropy and spending very differently.
The theoretical foundation of our exposition of financial needbeliefs is the self-determination theory (SDT, e.g. Deci and Ryan,1985; Ryan and Deci, 2000) of motivation, which posits threeuniversal human psychological needs: competence, autonomy andrelatedness. Based on this assumption, SDT-based research pre-dicts, and generally finds, that individuals experience higher levelsof happiness when the social environment supports the fulfilmentof, and when individuals’ values and goal pursuits are congruentwith, these three core psychological needs (for a review, see Ryanand Deci, 2002). Stone et al. (2010) extend these constructs, andposit that these relations hold, in the financial domain. Specifi-cally, they propose three core financial needs, i.e. (1) financialself-efficacy (mapping to SDT’s competence), (2) financial
bs_bs_banner
International Journal of Consumer Studies ISSN 1470-6423
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
675
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
autonomy (mapping to autonomy) and (3) financial communitytrust/support (mapping to relatedness; see Table 1). In addition,they explore two financial values: (1) materialism and (2) financialaltruism, one of which, materialism, SDT predicts will negativelyinfluence happiness (cf. Kasser, 2002). The results in Stone et al.(2010), with US student participants, support the prediction thatthe proposed financial needs influence financial values, which, inturn, influence consumers’ happiness and vitality.
The current paper contributes to consumer research by testingthe extent to which the financial need beliefs and financial valuesexplored in Stone et al. (2010) are universal vs. culturally embed-ded. Finding that financial need belief and financial value con-structs are universal would allow consumer research to confidentlyuse the constructs, and their associated measures, across cultures(cf. Manrai and Manrai, 1996). Alternatively, if the meaning ofthese constructs and their associated measures is culturally embed-ded, then cross-cultural comparisons of these metrics would beinvalid. Hence, the present study provides a foundation for study-ing consumers’ financial need beliefs and financial values outsidethe US. In addition, it provides evidence on the cross-culturalvalidity of measures of financial need beliefs and financial values.
Next, we describe the measurement models followed by anexposition of the constructs that underlie the measurement models.Following this, we present the purpose, methods, participants,procedures, measures and results of two studies that test the cross-country invariance of aspects of the measurement models. We thensummarize the results and their implications and limitations.
Measurement modelsTwo measurement models, one for financial needs beliefs, and onefor financial values, are the basis for assessing cross-countryequivalence. In the financial need belief measurement model, 15items measure the three posited financial need belief constructs.Validating this measurement model across countries is importantbecause cross-country variance in the meaning, or importance, offinancial needs, may arise due to socio-economic and culturalfactors. For example, the need for financial self-efficacy may begreater, and more beneficial, among consumers with low incomes,and thus in poorer than wealthier countries (cf. Maslow, 1954). Inaddition, individualist cultures may engender stronger needs forfinancial autonomy than in collectivist cultures (cf. Kashima et al.,1995; Iyengar and Lepper, 1999), while the reverse may be true forfinancial community needs, i.e. they are stronger in collectivistthan individualist cultures (cf. Deci et al., 2001; Grouzet et al.,2005).
The financial value measurement model consists of elevenitems that measure materialism and altruism. Evidence suggeststhat materialism has multiple deleterious outcomes, including
compulsive purchasing (Rindfleisch et al., 1997), insecurity(Chang and Arkin, 2002) and social conformity (Schroeder andDugal, 1995). In contrast, altruistic financial values correlate withpositive outcomes (e.g. happiness, subjective well-being) inUS samples (Dunn et al., 2008; Stone et al., 2010). However,cross-country differences may also arise in this financial valuemeasurement model. For example, Schaefer et al. (2004) findcross-country differences in the factor structure of the materialismconstruct between samples of Chinese vs. Japanese and Americanadolescents. Similarly, Tang et al. (2008) find that the ‘love ofmoney’, a value that is similar to materialism, positively relates tohelping behaviour among Egyptians and Poles, but is unrelated tohelping behaviour among Americans and Taiwanese. Hence, themeaning of, and relations between, materialistic and altruisticfinancial values may be culturally embedded and dependent.
We conduct two studies to examine the extent of cross-countryequivalence of the above mentioned financial need belief andfinancial value measurement models in student (Study 1) andnon-student (Study 2) samples. Our hypotheses test whetherthe two measurement models demonstrate configural (H1),metric (H2), factor variance/covariance (H3 and H4) and scalar(H5) invariance across countries. Table 2 summarizes the testedhypotheses. Because theory and research do not clearly suggestcross-country variance or invariance in most cases, we state allhypotheses in null form. Multi-group confirmatory factor analyses(CFA) test for invariance.
Hypothesis tests use data are from six countries: Brazil, mainlandChina, Taiwan, Tunisia, Russia and the US, which were selectedbased on both availability (i.e. a reliable native-language resident orexpatriate to facilitate data collection) and to provide contrastingsocio-economic and cultural samples. Table 3 Panel A presentsselected socio-economic data on these countries and illustrates keysimilarities and differences. Table 3 Panel B presents survey datacomparing these countries on Hofstede’s (2001) five culturaldimensions.1 These six countries’ unique combinations of eco-nomic status and growth, language, religion, traditions and culturalvalues provide contrasting samples. Hence, sampling from thesecountries allows us to conduct a stringent equivalence test of thefinancial need belief and financial value measurement models.
This investigation also contributes by advancing theory.Although creating a functional relationship with money and mate-rial possessions is a critical consumer skill in modern culture (Belkand Pollay, 1985; Belk and Bryce, 1986), this competency isunder-explored in research (Lastovicka et al., 1999; Mitchell andMickel, 1999). Instead, consumer and psychology research hasmainly focused on identifying the determinants and consequencesof dysfunctional relationships to money and material possessions(e.g. Kasser, 2002; Kasser and Kanner, 2003; see also Molleret al., 2006). However, growing scientific attention to what is bestin human strivings and achievement (Zwierlein, 1986; Sheldonand Kasser, 2001; Csikszentmihalyi and Csikszentmihalyi, 2006;Polak and McCullough, 2006; Giacalone et al., 2008) suggests thewisdom of investigating the nature of a functional consumer rela-tionship to money and material possessions. This paper extendsthese efforts, and research on SDT, altruism and materialism, by
1While concerns exist about the validity of Hofstede’s constructs (e.g. seeBaskerville, 2003), we apply these constructs as a rough, imprecise basisfor comparing cultural differences among the studied countries.
Table 1 Mapping of self-determination theory (SDT) to financial needbelief constructs
SDT core need Financial need belief
Autonomy Financial autonomyCompetence Financial self-efficacyRelatedness Financial community – trust
Financial community – support
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
676
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
examining, across countries, financial needs and values that maycontribute to, or diminish, individual well-being.
Measured constructs
Financial self-efficacy
Within SDT, the core psychological need of competence orself-efficacy refers to a belief that one can effectively influence
meaningful outcomes. Extending this construct to the financialdomain suggests financial self-efficacy (FSE), which is the extentof one’s belief that one is capable of successfully managingthe financial events in one’s life. No research of which weare aware investigates financial self-efficacy as we define it.However, research explores similar constructs such as the abilityto cope with economic risks (Engelberg, 2007), skill at handlingmoney (Mitchell and Mickel, 1999), frugality (Lastovicka et al.,1999; Kasser, 2005) and a construct labelled ‘budgeting’ in
Table 2 Summary of hypotheses and results
Hypotheses and predictions Model results
HypothesisType of equivalence/invariance Prediction (null)
Financial need beliefs (H1a,H2a, H3a, H4a and H5a)
Financial values (H1b, H2b,H3b, H4b and H5b)
H1a and H1b Configural Invariant cross-country factor structures willobtain
Supported: Studies 1 and 2 Supported: Studies 1 and 2
H2a and H2b Metric Invariant factor loadings across countries Supported: Studies 1 and 2 Partially Supported: Studies1 and 2
H3a and H3b Factor variance Invariant factor variances across countries Supported: Study 1 Supported: Study 1Rejected: Study 2 Rejected: Study 2
H4a and H4b Factor covariance Invariant factor covariances across countries Supported: Studies 1 and 2 Rejected: Studies 1 and 2H5a and H5b Scalar Invariant intercepts of scale items across
countriesRejected: Studies 1 and 2 Rejected: Studies 1 and 2
Table 3 Country comparison data
Panel A: 2003–2011 socio-economic data by country
Country GDP HDI Econ freedom Happiness Motor vehicles Poverty GDP growth
Tunisia $4600 0.753 2.1 5.9 71 12.8 0Brazil $12 400 0.792 2.0 7.5 81 10.8 2.8China $7783 0.755 1.5 6.3 10 29.8 9.5Russia $12 900 0.795 1.3 5.5 124 0.1 4.3Taiwan $21 900 0.882 2.7 6.2 297 0 5.2US $48 400 0.944 3.2 7.4 765 0 1.5
Data retrieved 30 May 2012.GDP, per capita (Central Intelligence Agency, 2011); HDI, human development index – includes life expectancy, literacy, education, living standards– range = 0.281–0.963 (United Nations Development Programme, 2011); Econ freedom, economic freedom rating (range = 0–3.55; NationMaster,2011); Poverty, % living on <$2/day (World Bank, 2011b).Happiness data (average on 10-point scale; Veenhoven, 2011).Motor vehicles per 1000 people (World Bank, 2011a).Annual GDP growth (Central Intelligence Agency, 2011).
Panel B: Cultural comparison data (source: http://www.geert-hofstede.com/)
Country PDI IDV MAS UAI LTO
Arab countries 80 38 52 68 n/aBrazil 69 38 49 76 65China 80 20 66 30 118Russia 94 38 36 95 n/aTaiwan 58 17 45 69 87US 40 91 62 46 29World averages 55 43 50 64 n/a
PDI, the power distance relationship, i.e. the extent of societal acceptance of power inequality; IDV, i.e. the cultural strength of ‘I’ (individualism andsmall families) vs. ‘we’ (collectivism and shared group identity) identification; MAS, masculinity, the extent of a cultural embrace of competition,achievement and success (traditionally masculine values) vs. quality of life and caring for others (traditionally feminine values); UAI, uncertaintyavoidance index, i.e. the extent of efforts to create cultural norms and institutions to manage and control change; LTO, long-term orientation, i.e. theextent to which the culture adopts a future and long-term vs. present and short-term planning focus; n/a, data not available.
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
677
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
Tang and colleagues’ Money Ethics Scale (e.g. Tang and Sutarso,2006).
The importance and value of FSE may vary with wealth andculture. For example, low-income individuals, who are (by defi-nition) more prevalent in poorer countries, have stronger needs forwealth than do their richer counterparts (Maslow, 1954). Conse-quently, FSE may be a more important skill in poorer thanwealthier countries due to its value in managing economic scar-city, i.e. of reducing the likelihood of individual poverty. Second,Heine et al. (1999) examine the universality of the need for posi-tive self-regard and find that positive self-regard is less importantin Japanese than North American culture. Hence, beliefs aboutone’s self-efficacy, including FSE, may be less important andconsequential in Asian than Western countries.
Financial autonomy
Within SDT, the need for autonomy concerns acting with a senseof choice, volition and self-determination. Financial autonomyis the belief that one’s financial decisions are volitional choicesthat reflect one’s interests and beliefs (Stone et al., 2010). Forexample, financial autonomy beliefs would likely facilitate thetransition from adolescent financial dependence to adult financialvolition. Alternatively, a non-working consumer whose partner hasa job may experience low financial autonomy if the non-workingpartner feels that their spouses’ job and position constrain theirfinancial volition and purchasing ability.
Although the cross-country generality of the financial autonomyconstruct is untested, the universality of the posited need forautonomy within SDT is controversial (e.g. see Chirkov et al.,2003). For example, Iyengar and Lepper (1999) and Kashimaet al. (1995) argue that, while autonomy needs are congruent withindividualistic culture, they contradict cultural norms of groupcohesion within collectivist cultures. Alternatively, SDT research-ers (e.g. Deci et al., 2001; Chirkov et al., 2003) argue thatautonomy is not synonymous with individualism and independ-ence; instead, individuals ‘. . . may autonomously embrace collec-tivist values and moral obligations (Deci et al., 2001, p. 940).’Thus, whether the financial autonomy construct is culturallyinvariant is, potentially, both a conceptual (e.g. what is financialautonomy?) and empirical (e.g. what is the relationship betweenfinancial autonomy and financial community beliefs?) question.
Financial community – trust and support
Within SDT, relatedness refers to the need for satisfying andsupportive social relationships. Extending this construct to thefinancial domain, Stone et al. (2010) propose and validate twomeasures of financial community beliefs: financial community-trust, and financial community-support. Financial community-trust is the belief that one can rely on significant others for helpwith financial issues and problems. Financial community-supportis the belief that financial resources can contribute to supportingcommunities and interpersonal relationships.
Deci et al. (2001, p. 940) argue that the need for relatedness ‘iscentral to those who hold collectivist values.’ Therefore, it is pos-sible that in collectivist cultures (e.g. China, Taiwan), financialcommunity needs may exceed those of individualist cultures(e.g. US).
Materialistic values
Research often assumes that financial values and goals are mate-rialistic and demonstrates the deleterious effects of materialisticvalues, and pursuing materialistic goals, on psychological,physical and relational health. For example, Burroughs andRindfleisch (2002) review the results of 18 studies that investi-gate the relationship between materialism and well-being. Allstudies provide evidence of a negative relationship betweenmaterialism and well-being. Cross-cultural replications confirmthe negative effects of materialism on well-being among Austral-ian, English, German, Romanian, Russian, Singaporean andSouth Korean participants (e.g. see Kasser et al., 2004). Alter-natively, cross-cultural research also suggests that materialismmay be less harmful in poorer than wealthier countries. Forexample, Tang et al.’s (2007) ‘love of money’ construct, which issimilar to materialism, positively correlates with unethicalbehaviour for employees in high and median, but not low, GDPcountries. Among employees in low GDP countries, the “love ofmoney”, i.e. materialism, is unrelated to unethical behavior.
Research has also explored the cross-cultural consistency ofsome materialism measurements. For example, Ger and Belk(1996) test the validity of five English-language measures of mate-rialism, using student participants in 12 countries. Results suggesthigher validity and reliability in measures of materialism amongWestern than non-Western respondents. Using responses from 14-to 17-year-olds from China, Japan and the US, Schaefer et al.(2004) also find some evidence of cross-country differences in thefactor structure of materialism construct among adolescents.Herein, we seek to contribute to the literature on materialism byshowing its relations to positive financial values, and by providingadditional evidence on its cross-cultural equivalence.
Financial altruism
Investigations rarely explore the nature of a functional relationshipto money and possessions. Most SDT research is also silent onfinancial values, such as financial altruism, that may positivelycontribute to the fulfilment of core psychological needs and per-sonal happiness. Financial altruism concerns the value of financialresources in relation to their ability to enhance one’s communityor interpersonal relationships (Stone et al., 2010; cf. De Cremer,2006; Francois and Vlassopoulos, 2008).
Evidence suggests that culturally specific relations may existbetween materialistic and altruistic values. For example, Tanget al. (2008) find that a financially materialist belief (i.e. the loveof money) positively predicts helping behaviour among Egyptians(r = 0.26) and Poles (r = 0.33), but is unrelated to helping amongAmericans and Taiwanese. Such results may obtain because, inless affluent countries, materialistic financial goals positivelyco-vary with altruistic values, such that concerns for providingcore material resources for significant others also represent altru-istic acts (e.g. providing for one’s children’s basic consumerneeds, e.g. shoes). In contrast, in affluent countries, materialisticgoals may often be competitive (e.g. to gain economic advantage;Frank, 2010, 2011). Hence, some evidence, although only indi-rectly related to the construct of financial altruism, suggests pos-sible country-specific influences on the relation betweenmaterialist and altruistic values. Herein, we seek to contribute by
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
678
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
providing evidence on the cultural specificity of, and relationsbetween, materialistic and altruistic financial values.
Study 1
Purpose
Study 1 seeks to establish the extent of cross-country invariance inthree measures of financial need beliefs (FSE, financial autonomyand financial community-trust) and two measures of financialvalues (materialism and financial altruism) in three studentsamples (China, Taiwan and the US). The CFA models appear inFigs 1 and 2. We explore the following specific types of invariancein the indicated order: configural (H1), metric (H2), factorvariance/covariance (H3/H4) and scalar (H5). Note that latentmean comparisons have meaning only after establishing metricand scalar invariance since, without equivalent factor loadings andintercepts, cross-country differences in scale items could be due torater bias, rather than true differences in underlying constructs(Chen et al., 2005).
Participants and procedures
Nine hundred eighty-eight undergraduate and graduate studentsfrom four Taiwanese and three US universities, and one Chineseuniversity, participated for partial, or additional, course credit.Chinese participants (54 men and 124 women), enrolled inNanjing University, ranged in age from 20 to 23 years (M = 21.79,SD = 0.76). Taiwanese participants (40 men and 180 women),ranged in age from 18 to 23 years (M = 19.41, SD = 1.33); USparticipants (318 men and 272 women), ranged in age from 18 to23 years (M = 20.66, SD = 1.37).
Native-speaking research collaborators translated the instru-ment. Following suggestions in the cross-cultural psychologyliterature (e.g. Brislin, 1970), translations to simplified and tradi-tional Chinese were followed by back-translations to English toconfirm syntactic and semantic correctness. Course instructorssolicited participation. Participants completed the instrument in
either English (US), simplified Chinese (China) or traditionalChinese (Taiwan). The Chinese and Taiwanese participants com-pleted the survey instrument on paper while the US participantscompleted the survey online.
Figure 1 (Study 1): Financial need beliefsconfirmatory factor analysis (CFA) model(financial self-efficacy, financial autonomy andfinancial community trust).
Figure 2 (Studies 1 and 2): Financial value confirmatory factor analysis(CFA) model (altruism and materialism).
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
679
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
Measures
The initial scale development process included the creation oridentification in existing literature of potential scale items, andmultiple rounds of pilot and pre-testing with participant groups
(not reported) to determine valid constructs (Stone et al., 2010).Constructs and the related measurement items, for both Studies 1and 2 appear in Table 4. To reduce instrument length, data collec-tion related to the construct of financial community-trust occurredin Study 1 (but not Study 2), while data collection related to the
Table 4 Constructs and scale itemsPanel A: Financial need beliefs
Construct: Financial self-efficacy (FSE)
Item label Item
FSE1 I am good at managing my moneyFSE2 I am satisfied with my ability to manage my moneyFSE3 Compared with other people, I think I do pretty well at making financial decisionsFSE4 I am pretty skilled at making financial decisionsFSE5 I budget my money very wellFSE6 I use my money very carefully
Construct: Financial autonomy
Item label Item
Auto1 I do not have a choice about making money decisions (reversed)Auto2 I make financial decisions because I have to, not because I want to (reversed).Auto3 In making financial decisions, I feel pushed, forced, and pressured (reversed).
Construct: Financial community – trust (measured only in Study 1)
Item label Item
Trust1 I can rely on other people to help me when I am in financial needTrust2 I can depend on other people for help with money problemsTrust3 I trust that the people I care most about will keep their financial commitments to
me
Construct: Financial community – support (measured only in Study 2)
Item label Item
Support1 I am willing to help the people I care most about financially if they need itSupport2 I feel that I can talk about money problems with my close friends and loved onesSupport3 Money is valuable because it can help you support the people that you love
Panel B: Financial values
Construct: Financial altruism
Item label Item
Altruism1 Either now or in the future, I intend to donate money to causes that I care aboutAltruism2 Money can be used for acts of kindness.Altruism3 Money is useful because it can help make the world a better placeAltruism4 Money, wisely used, can help build communitiesAltruism5 Donating money to charity is a waste of money (reverse) (deleted).
Construct: Materialism
Item label Item
Material1 I admire people who own expensive homes, cars and clothesMaterial2 Some of the important achievements in life include acquiring material
possessions.Material3 The things I own say a lot about how well I am doing in life.Material4 I like to own things that impress peopleMaterial5 I do not place much emphasis on the amount of material objects people own as
a sign of success (reversed) (deleted).Material6 I do not pay much attention to the material objects other people own (reversed)
(deleted).
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
680
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
construct of financial community-support occurred in Study 2 (butnot Study 1).
Financial need beliefs
Financial self-efficacy
Pilot testing resulted in a reliable six-item measure of FSE, createdor adapted from existing sources (Ryan et al., 1983; Tang, 1995;Mitchell and Mickel, 1999; Chatzky, 2003; ANZ Bank and ACNielsen, 2005), e.g. item, ‘I am good at managing my money.’Financial autonomy was measured as a three-item, reverse scoredmeasure, e.g. item, ‘I don’t have a choice about making moneydecisions.’ Hence, the reversed-scored financial autonomymeasure captures the extent to which a respondent believes thatexternal forces control or influence their financial life.
Financial community: trust
Financial community-trust was measured using a three-itemmeasure, e.g. item, ‘I can rely on other people to help me when Iam in financial need.’
Following the ‘rough guidelines’ criteria of Kline (2005),Cronbach’s alphas for the measured scales ranged from excellent(≥0.9) to weak (<0.7). Specifically, the FSE, financial autonomyand financial community-trust scales were, respectively, 0.92, 0.72and 0.63 for the Chinese sample, 0.93, 0.78 and 0.70 for theTaiwanese sample, and 0.95, 0.70 and 0.83 for the US sample.
Financial values
Financial altruism was measured using a five-item measure, e.g.item, ‘Either now or in the future, I intend to donate money tocauses that I care about.’We subsequently deleted the only reverse-scored item ‘Donating money to charity is a waste of money(reverse)’ due to its low factor loadings in the Chinese sample(0.19). Financial materialism was assessed using Richins andDawson’s six-item measure related to possession-defined success(PDS) instrument (Richins and Dawson, 1992; Chang and Arkin,2002), e.g. item, ‘I admire people who own expensive homes, cars,and clothes.’ We chose the PDS instrument based on its extensivevalidation (e.g. Burroughs and Rindfleisch, 2002; Wong et al.,2003; Richins, 2004) and brevity. Consistent with previousresearch, including the two reverse-scored scale items reducedvalidity (Wong et al., 2003); therefore, we report results for thefour-item scale. Cronbach’s alphas for the financial altruism andmaterialism ranged from very good (≥0.8) to weak (<0.7; Kline,2005). Specifically, Cronbach’s alphas for the financial altruismand materialism scales were, respectively, 0.68 and 0.66 for theChinese sample, 0.74 and 0.60 for the Taiwanese sample, and 0.77and 0.80 for the US sample.
Analyses
Multi-group CFA, using AMOS (Version 19.0, IBM SPSS Amos,Somers, NY, USA), provide maximum likelihood estimation(MLE) estimates.2 Following Burns et al. (2008), we use different
criteria to evaluate various types of cross-country invariance: (1)χ2 difference tests, which according to Cudeck and Browne (1983)and MacCallum et al. (1992), provide extremely stringent test ofinvariance; (2) comparing model fit indices including comparativefit index (CFI), standardized root mean squared residual (SRMR)and root mean square error of approximation (RMSEA); (3) exam-ining the modification indices (MIs) for the constrained modelsand ensure all MIs are smaller than critical values.3
Results
Financial need beliefs model: FSE-financial
autonomy-community trust
H1a (configural invariance)The configural or baseline model (see Fig. 1), with no cross-cultural invariance constraints, is estimated on all three samples.As indicated in Table 5, the adequate fit of the baseline model[χ2(144) = 218.26, CFI = 0.990, SRMR = 0.027 and RMSEA =0.023] provides evidence of configural invariance. For all threesamples, item loadings on FSE, financial autonomy andcommunity-trust are significant and above 0.30. H1a is thussupported.
H2a (metric invariance)The results also provide evidence of H2a or metric invariance.Specifically, although the χ2 increase significantly (Δχ2 = 60.40,Δd.f. = 18, P < 0.01) when factor loadings are constrained as equalbetween countries, the model fit indices do not suggest an inferiorfit (CFI = 0.984, SRMR = 0.027 and RMSEA = 0.027) comparedwith the unconstrained model. In addition, none of the MIs in themetric invariance model is greater than the critical value of 10.22.
H3a and H4a (factor variance/covariance invariance)We next estimate a CFA model with factor variances andcovariances constrained to equality across countries. Comparedwith the metric invariance model, the χ2 increase significantly(Δχ2 = 47.26, Δd.f. = 12, P < 0.01), but the model fit indices againare not inferior (CFI = 0.979, SRMR = 0.031 and RMSEA =0.030). In addition, only one MI (i.e. the variance of financialautonomy for the Taiwanese sample) is greater than the criticalvalue of 9.47. H3a and H4a are thus both supported. Across threesamples, the covariance between FSE and Autonomy (0.34,P < 0.01) is higher than that between FSE and community-trust(0.03, P = 0.62) and that between autonomy and community-trust(0.11, P = 0.01).
H5a (scalar invariance)The scalar invariance model provides a significantly poorer fit ofthe data compared with the factor variance/covariance invariancemodel (Δχ2 = 346.55, Δd.f. = 12, P < 0.01). The model fit indicesshow inferior fit with the exception of SRMR. Many MIs aregreater than the critical value of 10.22. Thus, H5a is rejected. Thissuggests that item mean differences are not necessarily caused by
2We also derive bootstrapping estimates of standard errors that correct fornon-normality. Since MLE and bootstrapping method provide similar sig-nificance results, we report only MLE results.
3With an acceptable family-wise error rate = 0.05, the alpha error for eachconstraint should be 0.05/n, where ‘n’ denotes the number of invarianceconstraints imposed (Huck, 2012).
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
681
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
latent constructs mean differences, which disallows mean differ-ences comparisons across countries (Chen et al., 2005). Forexample, individuals from different countries may interpret thescale items differently, or they may judge those items based onsocial comparison with their countrymen (Peng et al., 1997),which could artificially lead to lower or higher ratings in onecountry than another.
Financial value model: financial altruism-materialism
H1b (configural invariance)The baseline model (see Fig. 2), with no cross-country invarianceconstraints, is estimated on all three samples. As indicated inthe bottom half of Table 5, the adequate fit of the model[χ2(51) = 106.50, CFI = 0.972, SRMR = 0.044 and RMSEA =0.034] provides evidence of configural invariance for the financialvalue measurement model. For all three samples, item loadingson materialism and altruism are significant and above 0.38 withthe exception of the loading on altruism-1 for the Chinese sample(see below). H1b is thus supported.
H2b (metric invariance)The results also provide evidence of partial metric invariance. Thefactor loading on Altruism item no. 1 (i.e. ‘Either now or in thefuture, I intend to donate money to causes that I care about.’) isallowed to vary for the Chinese sample due to its significantlylower loading in this sample (0.24 in the Chinese sample com-pared with 0.67 in the Taiwanese sample and 0.61 in the USsample). Compared with the baseline model, this partial metricinvariance model derives a moderate increase in χ2 (Δχ2 = 38.69,Δd.f. = 11, P < 0.01), and the model fit indices are not inferior(CFI = 0.957, SRMR = 0.046 and RMSEA = 0.037). No MI in thepartial metric invariance model is greater than the critical valueof 9.47.
H3b and H4b (factor variance/covariance invariance)The evidence provides support for H3b or factor variance invari-ance. Specifically, when factor variances are constrained as equalbetween countries, the model fit indices are not inferior(CFI = 0.947, SRMR = 0.050 and RMSEA = 0.040) comparedwith the partial metric invariance model. None of the MI is greaterthan the critical value of 7.48. However, constraining the factorcovariance (between altruism and materialism) as equal betweencountries results in a significant MI. When factor covariance isestimated freely between countries, the standardized correlationbetween altruism and materialism is insignificant for the Americansample and Taiwanese sample (ρ = 0.07, P = 0.16, and ρ = 0.13,P = 0.18, respectively), but is significant and positive for theChinese sample (ρ = 0.57, P < 0.01). Hence, on average, in con-trast to US and Taiwanese students, the Chinese students whoscore high (low) on materialism also score high (low) on financialaltruism. H4b is thus rejected.
H5b (scalar invariance)The scalar invariance model provides a much poorer fit of the datacompared with the factor variance invariance model (Δχ2 =597.66, Δd.f. = 11, P < 0.01) and model fit indices are inadequate(CFI = 0.646, SRMR = 0.052 and RMSEA = 0.096). Moreover,many MIs are greater than 10. Thus, H5b is rejected and scalarinvariance is again not supported for the financial value measure-ment model, which disallows latent mean difference comparisonbetween countries.
Summary
Results (n = 988) from student respondents in three countries(China, Taiwan and the US) support configural, metric and factorvariance/covariance invariance in the financial need belief CFAmodel (see Fig. 1). Across three samples, we find strong correla-tion between FSE and financial autonomy need beliefs, but the
Table 5 (Study 1): Fit indices and change in fit for baseline models, metric invariance models, factor structure invariance models and scalar invariancemodels: for student samples
Model
Chi-square Absolute fit indices Comparative fit indices
χ2 d.f. P CFI SRMR RMSEA RMSEA, CI Δχ2 Δd.f. P ΔCFI ΔSRMR ΔRMSEA
1. FSE-autonomy-trustbaseline model
218.26 144 0.000 0.990 0.027 0.023 (0.016,0.029)
Metric invariance 278.66 162 0.000 0.984 0.027 0.027 (0.022,0.032) 60.40 18 <0.01 0.006 0 0.004Factor variance andcovariance invariance
325.92 174 0.000 0.979 0.031 0.030 (0.025,0.035) 47.26 12 <0.01 0.005 0.004 0.003
Scalar invariance 672.47 186 0.000 0.932 0.026 0.052 (0.047,0.056) 346.55 12 <0.01 0.047 -0.005 0.0222. Altruism-materialism
baseline model106.50 51 0.000 0.972 0.044 0.034 (0.025,0.042)
Metric invariance 173.65 63 0.000 0.943 0.047 0.043 (0.035,0.050)Partially metricinvariance*
145.19 62 0.000 0.957 0.046 0.037 (0.029,0.045) 38.69 11 <0.01 0.015 0.002 0.003
Factor variance invariance 169.02 66 0.000 0.947 0.050 0.040 (0.033,0.048) 23.83 4 <0.01 0.01 0.004 0.003Factor variance andcovariance invariance
189.93 68 0.000 0.937 0.054 0.043 (0.036,0.050) 44.74 6 <0.01 0.02 0.008 0.006
Scalar invariance 766.68 77 0.000 0.646 0.052 0.096 (0.090,0.102) 597.66 11 <0.01 0.301 0.002 0.056
*Release loading constraint on altruism item no. 1 for the Chinese sample.
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
682
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
correlation between these two need beliefs and financialcommunity-trust need belief is weak or insignificant. We also findsupport for configural, partial metric and factor variance invari-ance in the financial values CFA model (see Fig. 2). The resultssuggest cross-cultural differences in the structural relationbetween materialistic and altruistic financial values. For example,in our results, while financial altruism and materialism areuncorrelated among US and Taiwanese respondents, they posi-tively correlate among Chinese respondents. This finding can beinterpreted as consistent with Tang et al.’s (2008) results that thelove of money, which is similar to materialism, is unrelated tohelping behaviour among Americans and Taiwanese (i.e. wealthiercountries), but positively correlates to helping behaviour amongEgyptians and Poles (i.e. poorer countries). This result maysuggest country-specific differences, in the relation between finan-cial materialism and altruism, based on wealth. In less wealthycountries, materialism may correlate with altruism, since materi-alistic goals may be consistent with, and supportive of, relationswith loved ones. In contrast, in wealthier countries, financiallymaterialistic and altruistic goals may diverge because economicgoals are less supportive of one’s loved ones, and more motivatedby tournament competitions for economic outcomes (e.g. seeFrank, 2010, 2011).
These results may provide insight into the results of twostudies (Eastman et al., 1997; Podoshen et al., 2011), whichdocument higher levels of materialism among Chinese, thanAmerican, young adults. Based on these results, Podoshen et al.(2011) argues that young Chinese consumers may overspendon conspicuous goods and neglect care of their elders. Such per-ceptions may explain why the Chinese national legislaturerecently passed legislation requiring children to visit theirparents ‘often’ or face lawsuits, brought by parents, due to thechildren’s neglect (Associated Press, 2012). Alternatively, ourresults suggest that among young Chinese consumers (students),materialistic values need not contradict altruism, e.g. caring forfamily, including the elderly. Hence, materialistic values, amongyoung Chinese consumers, may provide no evidence of neglector irresponsibility.
Study 2
Purpose
Study 1 provides evidence, with student participants, ofconfigural, metric and factor variance/covariance invariance in thefinancial need belief CFA model, and configural, partial metric andfactor variance invariance in the financial values CFA model.Study 2 extends Study 1 to older, non-student samples in fivecountries, i.e. Brazil, China, Russia, Tunisia and the US. As inStudy 1, we estimate the financial need beliefs CFA model (seeFig. 3) and the financial values CFA model (see Fig. 2) simulta-neously on five non-student samples, i.e. the US sample (n = 196),the Chinese sample (n = 189), the Brazilian sample (n = 287), theRussian sample (n = 157) and the Tunisian sample (n = 130). Wetest the same cross-cultural invariances of these CFA models.
Participants and procedures
Data collection proceeded as in Study 1, except we requestednon-student participants in participating countries. Participationwas solicited, using a snowball sampling method, through research-ers’ friends and professional colleagues and was uncompensated.Given that most participants in Study 2 are personal or professionalassociates of university faculty, it is likely that participants are, onaverage, wealthier, and more educated than are the average resi-dents of the sampled countries. The Brazilian, Russian and Tunisianparticipants completed paper surveys while the Chinese and USparticipants completed an online survey. Some demographic differ-ences of respondents exist across the five samples (see Table 6A).For example, on average, the Brazilian respondents are younger,while the Chinese respondents on average have more years ofpost-secondary education and fewer years of work experience.
Measures
Study 2 measures are the same as Study 1 with one exception: wemeasured financial community-support instead of community-
Figure 3 (Study 2): Financial need beliefsconfirmatory factor analysis (CFA) model(financial self-efficacy, financial autonomy andfinancial community support).
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
683
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
trust to represent the need for relatedness proposed by SDT. Thisis because we speculate that providing financial support to one’scommunity may be of greater concern among non-students (whoare more likely to have full-time jobs and stable income) whiletrust in one’s financial community may be of greater concernamong students (who often lack full-time jobs and income). Meas-urement of financial community-support was with a three-itemmeasure, e.g. item, ‘I am willing to help the people I care mostabout financially if they need it.’
Some Study 2 measures evidence low Cronbach’s alphas incertain country samples, e.g. the financial community-supportmeasure in the Tunisian sample and the altruism measure in theRussian sample – see Table 6B. Cronbach’s alpha is based onintercorrelations between scale items and thus measures the inter-nal consistency of a scale (Cronbach, 1951). Therefore, lowCronbach’s alphas suggest that for some country samples, thescale items may capture multiple underlying constructs or factors.The CFA in the results section that follows provides additionalinsight into differing factor structures between countries.
Results
Financial need beliefs model: FSE-financial
autonomy-community support
The baseline model (see Fig. 3), with no cross-cultural invarianceconstraints, is estimated on all five samples. As indicated inTable 7, the adequate fit of the model [χ2(240) = 354.47,CFI = 0.978, SRMR = 0.061 and RMSEA = 0.022] provides evi-dence of configural invariance, which supports H1a. For all fivesamples, all item loadings are significant and above 0.31.
The results support H2a or metric invariance. Although theχ2 increases significantly when the factor loadings are con-strained as equal between countries, the model fit indices are
not inferior (see Table 7). Moreover, none of the MIs is greaterthan the critical value of 11.51. This result is consistent withStudy 1.
In contrast to Study 1, the factor variance invariance modelresults in a worse fit compared with the metric invariance model(Δχ2 = 177.03, Δd.f. = 12, P < 0.01, CFI = 0.935, SRMR = 0.071and RMSEA = 0.035). The highest MI associates with the con-straint imposed on the variance of autonomy for the Tunisiansample. Given the multiple cross-cultural constraints we have torelax in order to achieve partial factor variance invariance, weconclude that factor variance invariance is not established with ournon-student samples. H3a is thus rejected.
The model fit of the factor covariance invariance model does notworsen significantly compared with the metric invariance model(Δχ2 = 22.40, Δd.f. = 12, P = 0.03, CFI = 0.965, SRMR = 0.078and RMSEA = 0.025; see Table 7). The results thus support H4aor factor covariance invariance between countries. Similar withwhat we find in Study 1, the covariance between FSE andautonomy (0.25, P < 0.01) is higher than that between FSE andcommunity-support (0.06, P = 0.02) and that between autonomyand community-support (0.01, P = 0.41).
Similar to Study 1, support for H5a or scalar invariance was notobtained. The scalar invariance model results in a significantpoorer fit of the data compared with the factor covariance invari-ance model (Δχ2 = 954.5, Δd.f. = 36, P < 0.01, CFI = 0.785,SRMR = 0.086 and RMSEA = 0.060; see Table 7).
Financial value model: financial altruism-materialism
As indicated in the bottom half of Table 7, the adequate fit ofthe baseline model (see Fig. 2) [χ2(85) = 156.26, CFI = 0.945,SRMR = 0.047 and RMSEA = 0.030] provides evidence ofconfigural invariance in support of H1b. For all five samples, allitem loadings are significant and above 0.35.
Table 6 (Study 2)
Panel A: Demographic information for non-student respondents – mean (standard deviation) [median]
Country n Age Gender (% female) Post-secondary education (years) Work experience (years)
China 189 42.94% = 26 to 30;29.41% = 31 to 35;18.82% = 36 to 40*
48.55 42.20% = bachelor’s degree;56.65% = master’s or PhD*
8.24 (5.67) [7]
Brazil 287 32.71 (8.08) [31] 44.96 3.76 (2.06) [4] 13.17 (8.00) [12]Russia 157 37.10 (16.71) [34] 47.77 4.34 (1.07) [4] 13.27 (15.04) [7]Tunisia 130 36.27 (9.58) [34.5] 33.08 3.30 (2.34) [4] 11.93 (10.33) [8]US 196 38.82 (13.49) [37.5] 58.03 4.78 (1.60) [6] 14.92 (12.43) [12.5]
*The wording of the age and education questions differed in the Chinese instrument.
Panel B: Cronbach’s alphas by scale and by country
Country (sample)Self-efficacy(FSE) Autonomy
Community(support) Altruism Materialism
China (n = 189) 0.89 0.65 0.75 0.76 0.55Brazil (n = 287) 0.90 0.79 0.56 0.56 0.58Russia (n = 157) 0.85 0.77 0.70 0.55 0.74Tunisia (n = 130) 0.84 0.85 0.48 0.57 0.65US (n = 196) 0.95 0.77 0.58 0.64 0.84
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
684
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
We find some support for H2b or metric invariance in theseconstructs. Specifically, when the factor loadings are constrainedas equal between countries, χ2 increases moderately (Δχ2 = 60.29,Δd.f. = 24, P < 0.01) and SRMR as well as RMSEA are notinferior (SRMR = 0.052 and RMSEA = 0.032). Although CFIworsens (CFI = 0.917), none of the MIs is greater than 11.51, thecritical value.
Different from Study 1, we find no evidence of factor varianceinvariance between countries. The factor variance invariancemodel results in a much worse fit of data compared with the metricinvariance model (Δχ2 = 176.99, Δd.f. = 8, P < 0.01, CFI = 0.788,SRMR = 0.094 and RMSEA = 0.050; see Table 7). The greatestMIs associate to the constraints imposed on the variance of mate-rialism for the US and Brazilian samples. H3b is thus rejected.
Consistent with Study 1, H4b is rejected. Specifically, the modelfit of the factor covariance invariance model is worse comparedwith the metric invariance model (Δχ2 = 30.70, Δd.f. = 4, P < 0.01,CFI = 0.897, SRMR = 0.054 and RMSEA = 0.035; see Table 7).When factor covariance is estimated freely between countries, thestandardized correlation coefficient between altruism and materi-alism is 0.02 (P = 0.83), 0.28 (P = 0.02), 0.27 (P = 0.01), −0.50(P < 0.01) and 0.41 (P < 0.01) for the US, Chinese, Brazilian,Russian and Tunisian samples respectively. Recall that in Study 1,we find that their correlation is positive in less wealthy country(i.e. China) and insignificant in more wealthy countries (i.e.Taiwan and the US). Thus, with the exception of the Russiansample, these results on the correlation between altruism andmaterialism in Study 2 are largely consistent with those found inStudy 1. We discuss a potential reason for the negative correlationin the Russian sample in the section that follows.
Similar to Study 1, support for H5b or scalar invariance was notobtained.
Summary
The Study 2 results largely replicate those of Study 1 with theexceptions of the failure to establish factor variance invarianceacross the measurement models, and lower Cronbach’s alphas
for several constructs. Results from non-student participants(n = 959) in five countries support configural, metric and factorcovariance invariance of the financial need beliefs measurementmodel. A strong correlation obtains between the need beliefs offinancial autonomy and financial self-efficacy across the fivesamples; in contrast, the correlation between the two need beliefsand the need belief of financial community-support is either weakor insignificant. We also find support for configural invarianceand some evidence of metric invariance of the financial valuesmeasurement model. Moreover, consistent with Study 1, Study 2also suggests cross-cultural differences in the structural relationbetween financial altruism and materialism. Specifically, Study 2results suggest that among non-student respondents, financialaltruism and materialism: (1) positively correlate among Chinese,Brazilian and Tunisian respondents; (2) are uncorrelated amongUS respondents; and (3) negatively correlate among Russianrespondents. Although speculative, the negative correlationbetween financial altruism and materialism among Russianrespondents may result from a widespread mistrust of Russiancharitable organizations (Mersiyanova and Yakobson, 2010).Perhaps where consumers perceive charitable organizations asuntrustworthy, e.g. Russia, beliefs arise that ‘altruistic’ financialgoals are self-interested and materialistic. Such attitudes mayexplain the negative correlation between financial altruism andmaterialism among (only the) Russian respondents. Given ourspeculations regarding this result, we encourage research inves-tigating this relationship.
Discussion, limitations and conclusion
Discussion
Table 2 summarizes the results of the two investigations. Acrossboth studies, configural invariance obtains for both the financialneed belief and financial value measurement models. The resultsprovide evidence of similar factor structures, across differentcountries, between both student and non-student consumers.Metric invariance, which tests the equivalence of the factor
Table 7 (Study 2): Fit indices and change in fit for baseline models, metric invariance models, factor structure invariance models and scalar invariancemodels: for non-student samples
Model
Chi-square Absolute fit indices Comparative fit indices
χ2 d.f. P CFI SRMR RMSEA RMSEA, CI Δχ2 Δd.f. P ΔCFI ΔSRMR ΔRMSEA
1. FSE-autonomy-supportbaseline model
354.5 240 0.000 0.978 0.061 0.022 (0.017, 0.027)
Metric invariance 442.1 276 0.000 0.967 0.061 0.025 (0.021, 0.029) 87.58 36 <0.01 0.011 0.000 0.003Factor variance invariance 619.1 288 0.000 0.935 0.071 0.035 (0.031, 0.038) 177 12 <0.01 0.032 0.010 0.010Factor covarianceinvariance
464.5 288 0.000 0.965 0.078 0.025 (0.021, 0.030) 22.4 12 0.03 0.002 0.017 0.000
Scalar invariance 1419 324 0.000 0.785 0.086 0.060 (0.056, 0.063) 955 36 <0.01 0.182 0.025 0.0352. Altruism-materialism
baseline model156.3 85 0.000 0.945 0.047 0.030 (0.022, 0.037)
Metric invariance 216.6 109 0.000 0.917 0.052 0.032 (0.026, 0.038) 60.29 24 <0.01 0.028 0.005 0.002Factor variance invariance 393.5 117 0.000 0.788 0.094 0.050 (0.044, 0.055) 177 8 <0.01 0.129 0.042 0.018Factor covarianceinvariance
247.3 113 0.000 0.897 0.054 0.035 (0.029, 0.041) 30.70 4 <0.01 0.020 0.002 0.003
Scalar invariance 1309 141 0.000 0.104 0.078 0.093 (0.089, 0.098) 1062 28 <0.01 0.813 0.026 0.061
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
685
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
loadings across countries, obtains in both studies for the financialneed belief constructs. Partial metric invariance obtains for thefinancial value constructs among both student and non-studentrespondents. These results together suggest that one can use themeasures of financial need beliefs and values to understand con-sumers in different countries. However, due to the lack of scalarinvariance, we may not use those measures to conduct latent meancomparisons between countries. In other words, the mere differ-ences in scale scores across countries cannot be interpreted as realcross-country differences in financial need beliefs and financialvalues. Our results cast doubts on the validity of research thatcompares financial attitude scales (such as materialism) acrosscountries without establishing their measurement equivalence.
For the financial need belief constructs, Studies 1 and 2 dem-onstrate factor covariance invariance, i.e. equivalent factor corre-lations across countries. In other words, the various types offinancial need beliefs correlate with one another in similar waysacross countries. Specifically, although FSE and financialautonomy need beliefs moderately correlate with each other, thecorrelations between these two financial need beliefs and financialcommunity need beliefs are either insignificant or small.
Results for the financial value constructs, in both studies,suggest that while in wealthier countries (i.e. the US and Taiwan),financial altruism and materialism are uncorrelated, in lesswealthy countries (i.e. Brazil, China and Tunisia), with the excep-tion of Russia, these constructs positively correlate. These resultsmay imply that in less wealthy countries, pursuing wealth andpossessions are a means to enhance individuals’ community andinterpersonal relationships. This is consistent with cross-culturalresearch results indicating that in less wealthy countries, materi-alism is less extrinsic, i.e. more transcendent, than in wealthiercountries (Grouzet et al., 2005).
The results indicate greater cross-country commonality in con-sumers’ financial need beliefs than their financial values. Forexample, full metric invariance and factor covariance invarianceobtain in the financial need belief measurement model but not inthe financial value measurement model. These results largelyaccord with SDT’s postulation of common psychological needsacross cultures (e.g. Ryan et al., 1999; Deci et al., 2001; Grouzetet al., 2005). Specifically, we find evidence of cross-countryinvariance in financial needs beliefs, which are extensions of thethree core psychological needs posited by SDT. In contrast, wefind less support for cross-country invariance in the measurementof financial values that we posit, which do not directly tie to corepsychological needs according to SDT.
Implications for theory and practice
Our studies suggest configural, metric and factor covarianceinvariance in financial need beliefs, which we posited as exten-sions of the core psychological needs posited in SDT. Conceptu-ally, this finding supports a core assertion of SDT, i.e. theinvariance of psychological (human) needs. Pragmatically, it pro-vides empirical support for using need belief measures acrosscultures. However, scalar invariance does not obtain for the finan-cial need belief constructs. Hence, the results support the use offinancial need belief constructs in ‘nomological nets’, i.e. toexplore the relations among constructs (Steenkamp andBaumgartner, 1998).
In contrast, less support emerges for the cross-cultural stabilityof financial value constructs, which are not posited as arising fromSDT principles. Hence, our finding cautions researchers thatcross-cultural comparisons of financial values could be invalid.In addition, consistent with some previous investigations, theresults also suggest cross-country differences in relation betweenfinancially altruistic and materialistic values. Our results fail tosupport the belief that materialistic values harm individuals’ psy-chological well-being and communities. Instead, the resultssuggest culturally dependent relations between materialism andother constructs.
Limitations and conclusion
Our surveys utilized country contacts in both studies and a snow-ball sampling method in Study 2. Accordingly, our samples werenot randomly chosen from within-country sampling frames. Study1 intentionally chose a younger than average sample of consumers(students), while Study 2 intentionally chose an older than averagesample of non-student consumers. While the correlation of themedian age by country (Central Intelligence Agency, 2009) andthe average age of our Study 2 samples is high (r = 0.77), it islikely that our samples differ from the country populations inseveral important ways. Specifically, because our samples reliedon professors and professionals in the sampled countries, it islikely that our samples are wealthier, and more educated, than themedian populations. In addition, the lower public visibility ofwomen in Islamic countries likely explains the lower men towomen ratio (33% women to 67% men) in our Tunisian sample.Consequently, our samples likely represent the financial need andattitude beliefs held by wealthier, more educated citizens in thesampled countries. Although convenience samples are common incross-cultural psychology literature, we caution readers regardinginferring from our snowball samples to country populations.
Moreover, this investigation does not associate the five financialattitudes with consequence variables such as subjective well-being. Our investigation also ignores the potential predictors offinancial attitudes. It may be the case that cultural influencesdifferentially determine the antecedents and consequences offinancial need beliefs and financial values. For example, doespersonal wealth, relative to reference others, uniformly predictmaterialism and financial altruism across countries? Does finan-cial dependence, i.e. the absence of financial autonomy, influencepersonal happiness equally in wealthier and poorer countries?Further, the data herein are cross-sectional, not longitudinal.Cross-sectional data allows for observations related to correla-tional relations among the measured attitudes but not to stronginferences regarding the evolution or development of these atti-tudes. Hence, the investigation of the longitudinal developmentof these critical financial attitudes will likely be useful in articu-lating the cultural conditions that best enhance their growth anddevelopment.
Herein, two studies explored the extent of cross-country equiva-lence in financial need belief and financial value measurementmodels in student (Study 1) and non-student (Study 2) samples.Multi-group CFA results of both studies suggest configural, metricand factor covariance invariance in financial need beliefs.However, less support emerges for the cross-cultural stability offinancial value constructs. The results also show cross-country
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
686
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
differences in the relation between financially altruistic and mate-rialistic values.
AcknowledgementsThe authors thank the University of Kentucky, the Von AllmenSchool ofAccountancy, CA/Laurier Centre for theAdvancement ofAccounting Research and Education Research, and the GattonCollege of Business for financial support. Thanks also to ProfessorMei-Hwa Lin (National Chengchi University and Professor BeiDong (University of South Florida) for assistance with data collec-tion and the Virginia Commonwealth Faculty Excellence Fund.
ReferencesANZ Bank & AC Nielsen (2005) ANZ Survey of Adult Financial Lit-
eracy in Australia: Final Report. ANZ Bank, Melbourne.Associated Press (2012) New Chinese law requires adult kids to visit
parents, USA Today. [WWW document]. URL http://www.usatoday.com/story/news/world/2012/12/28/china-law-adult-children-visit-elderly-parents/1795721/ (accessed on 15 August 2013).
Baskerville, R.F. (2003) Hofstede never studied culture. Accounting,Organizations and Society, 28, 1–14.
Belk, R.W. & Bryce, W.J. (1986) Materialism and individual determin-ism in United States and Japanese print and television advertising.Advances in Consumer Research, 13, 568–572.
Belk, R.W. & Pollay, R.W. (1985) Materialism and magazine advertis-ing during the 20th-century. Advances in Consumer Research, 12,394–398.
Brislin, R.W. (1970) Back-translation for cross-cultural research.Journal of Cross-Cultural Psychology, 1, 185–216.
Burns, G.L., De Moura, M.A., Walsh, J.A., Desmul, C., Silpakit, C. &Sommers-Flanagan, J. (2008) Invariance and convergent and discrimi-nant validity between mothers’ and fathers’ ratings of oppositionaldefiant disorder toward adults, ADHD–HI, ADHD–IN, and academiccompetence factors within Brazilian, Thai, and American children.Psychological Assessment, 20, 121–130.
Burroughs, J.E. & Rindfleisch, A. (2002) Materialism and well-being: aconflicting values perspective. Journal of Consumer Research, 29,348–370.
Central Intelligence Agency (2009) The World Factbook. [WWW docu-ment]. URL https://www.cia.gov/library/publications/download/download-2009 (accessed on 15 August 2013).
Central Intelligence Agency (2011) The World Factbook 2011. [WWWdocument]. URL https://www.cia.gov/library/publications/download/download-2011/ (accessed on 15 August 2013).
Chang, L. & Arkin, R.M. (2002) Materialism as an attempt to cope withuncertainty. Psychology and Marketing, 19, 389–406.
Chatzky, J. (2003) You Don’t Have to Be Rich: Comfort, Happiness,and Financial Security on Your Own Terms. Portfolio, New York, NY.
Chen, F.F., Sousa, K.H. & West, S.G. (2005) Testing measurementinvariance of second-order factor models. Structural EquationModeling, 12, 471–492.
Chirkov, V., Ryan, R.M., Kim, Y. & Kaplan, U. (2003) Differentiatingautonomy from individualism and independence: a self-determinationtheory perspective on internalization of cultural orientations andwell-being. Journal of Personality and Social Psychology, 84,97–110.
Cronbach, L.J. (1951) Coefficient alpha and the internal structure oftests. Psychometrika, 16, 297–334.
Csikszentmihalyi, M. & Csikszentmihalyi, I.S. (2006) A Life WorthLiving: Contributions to Positive Psychology. Oxford UniversityPress, New York.
Cudeck, R. & Browne, M.W. (1983) Cross-validation of covariancestructures. Multivariate Behavioral Research, 18, 147–167.
De Cremer, D. (2006) The Social Psychology and Economics. LawrenceErlbaum, Mahwah, NJ.
Deci, E.L. & Ryan, R. (1985) Intrinsic Motivation and Self-Determination in Human Behavior. Plenum, New York, NY.
Deci, E.L., Ryan, R.M., Gagné, M., Leone, D.R., Usunov, J. &Kornazheva, B.P. (2001) Need satisfaction, motivation, and well-beingin the work organizations of a former eastern bloc country. Personal-ity and Social Psychology Bulletin, 27, 930–942.
Dunn, E.W., Aknin, L.B. & Norton, M.I. (2008) Spending money onothers promotes happiness. Science, 319, 1687–1688.
Eastman, J., Fredenberger, B., Campbell, D. & Calvert, S. (1997) Therelationship between status consumption and materialism: cross-cultural comparison of Chinese, Mexican and American students.Journal of Marketing Theory and Practice, 5, 52–65.
Engelberg, E. (2007) The perception of self-efficacy in coping witheconomic risks among young adults: an application of psychologicaltheory and research. International Journal of Consumer Studies, 31,95–101.
Francois, P. & Vlassopoulos, M. (2008) Financial altruism and the deliv-ery of social services. CESifo Economic Studies, 54, 22–54.
Frank, R.H. (2010) Luxury Fever : Weighing the Cost of Excess. Prince-ton University Press, Princeton, NJ.
Frank, R.H. (2011) The Darwin Economy: Liberty, Competition, and theCommon Good. Princeton University Press, Princeton, NJ.
Ger, G. & Belk, R.W. (1996) Cross-cultural differences in materialism.Journal of Economic Psychology, 17, 55–77.
Giacalone, R.A., Jurkiewicz, C.L. & Deckop, J.R. (2008) On ethics andsocial responsibility: the impact of materialism, postmaterialism, andhope. Human Relations, 61, 483–514.
Grouzet, F.M., Kasser, T., Ahuvia, A., Dols, J.M.F., Kim, Y., Lau, S.,Ryan, R.M., Saunders, S., Schmuck, P. & Sheldon, K.M. (2005) Thestructure of goal contents across 15 cultures. Journal of Personalityand Social Psychology, 89, 800–816.
Heine, S.J., Lehman, D.R., Markus, H.R. & Kitayama, S. (1999) Is therea universal need for positive self-regard. Psychological Review, 106,766–794.
Hofstede, G. (2001) Culture’s Consequences: Comparing Values,Behaviors, Institutions, and Organizations across Nations, 2nd edn.Sage, Thousand Oaks, CA.
Huck, S.W. (2012) Reading Statistics and Research, 6th edn. Pearson,Boston.
Iyengar, S.S. & Lepper, M.R. (1999) Rethinking the value of choice: acultural perspective on intrinsic motivation. Journal of Personalityand Social Psychology, 76, 349–366.
Kashima, Y., Yamaguchi, S., Kim, U., Choi, S., Gelfand, M. & Yuki, M.(1995) Culture, gender, and self: a perspective from individualism-collectivism research. Journal of Personality and Social Psychology,69, 925–937.
Kasser, T. (2002) The High Price of Materialism. MIT Press, Cam-bridge, MA.
Kasser, T. (2005) Frugality, generosity, and materialism in children andadolescents. In What Do Children Need to Flourish? Conceptualizingand Measuring Indicators of Positive Development (ed. by K. A.Moore & L. H. Lippman), pp. 357–373. Springer, New York.
Kasser, T. & Kanner, A. (2003) Psychology and Consumer Culture: TheStruggle for A Good Life in A Materialistic World. American Psycho-logical Association, Washington, DC.
Kasser, T., Ryan, R.M., Couchman, C.E. & Sheldon, K.M. (2004)Materialistic values: their causes and consequences. In Psychologyand the Culture of Consumption (ed. by T. Kasser & A. Kanner),pp. 11–28. American Psychological Association Press,New York, NY.
L. Guo et al. Consumer needs and values across cultures
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
687
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
Kline, R.B. (2005) Principles and Practices of Structural EquationModeling, 2nd edn. Guilford Press, New York.
Lastovicka, J.L., Bettencourt, L.A., Hughner, R.S. & Kuntze, R.J.(1999) Lifestyle of the tight and frugal: theory and measurement.Journal of Consumer Research, 26, 85–98.
MacCallum, R.C., Roznowski, M. & Necowitz, L.B. (1992) Modelmodifications in covariance structure analysis: the problem of capitali-zation on chance. Psychological Bulletin, 111, 490–504.
Manrai, L.A. & Manrai, A.K. (1996) Current issues in the cross-culturaland cross-national consumer research. Journal of International Con-sumer Marketing, 8, 9–22.
Maslow, A.H. (1954) Motivation and Personality. Harper, New York,NY.
Mersiyanova, I.V. & Yakobson, L.I. (2010) Features of charitable dona-tions and volunteer work in Russia, report for the ‘Boletín del Centrode Investigación de Economía y Sociedad’. [WWW document].URL http://www.grupcies.com/boletin/images/stories/PDFBoletin/articuloi_edic_86.pdf (accessed on 27 July 2011).
Mitchell, T.R. & Mickel, A.E. (1999) The meaning of money: anindividual-difference perspective. Academy of Management Review,24, 568–579.
Moller, A.C., Ryan, R.M. & Deci, E. (2006) Self-determination theoryand public policy: improving the quality of consumer decisionswithout using coercion. Journal of Public Policy & Marketing, 25,104–116.
Nationmaster (2011) Economy statistics. [WWW document].URL http://www.nationmaster.com/graph/eco_eco_ove-economy-overview (accessed on 15 August 2013).
Peng, K., Nisbett, R.E. & Wong, N. (1997) Validity problems comparingvalues across cultures and possible solutions. Psychological Methods,2, 329–344.
Podoshen, J.S., Li, L. & Zhang, J. (2011) Materialism and conspicuousconsumption in China: a cross-cultural examination. InternationalJournal of Consumer Studies, 35, 17–25.
Polak, E.L. & McCullough, M.E. (2006) Is gratitude an alternative tomaterialism. Journal of Happiness Studies, 7, 18, 343–360.
Richins, M.L. (2004) The material values scale: measurement propertiesand development of a short form. Journal of Consumer Research, 31,209–219.
Richins, M.L. & Dawson, S. (1992) A consumer values orientation formaterialism and its measurement – scale development and validation.Journal of Consumer Research, 19, 303–316.
Rindfleisch, A., Burroughs, J.E. & Denton, F. (1997) Family structure,materialism, and compulsive consumption. Journal of ConsumerResearch, 23, 312–325.
Ryan, R.M. & Deci, E.L. (2000) Self-determination theory and thefacilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68–78.
Ryan, R.M. & Deci, E.L. (2002) An overview of self-determinationtheory. In Handbook of Self-Determination Research (ed. by E. Deci& R. Ryan), pp. 3–33. University of Rochester Press, Rochester, NY.
Ryan, R.M., Mims, V. & Koestner, R. (1983) Relation of reward contin-gency and interpersonal context to intrinsic motivation: a review andtest using cognitive evaluation theory. Journal of Personality andSocial Psychology, 45, 736–750.
Ryan, R.M., Chirkov, V.I., Little, T.D., Sheldon, K.M., Timoshina, E. &Deci, E.L. (1999) The American dream in Russia: extrinsic aspira-tions and well-being in two cultures. Personality and Social Psychol-ogy Bulletin, 25, 1509–1524.
Schaefer, A.D., Hermans, C.M. & Parker, R.S. (2004) A cross-culturalexploration of materialism in adolescents. International Journal ofConsumer Studies, 28, 399–411.
Schroeder, J.E. & Dugal, S.S. (1995) Psychological correlates of thematerialism construct. Journal of Social Behavior and Personality,10, 243–253.
Sheldon, K.M. & Kasser, T. (2001) Goals, congruence, and positivewell-being: new empirical support for humanistic theories. Journal ofHumanistic Psychology, 41, 30–50.
Steenkamp, J.B.E.M. & Baumgartner, H. (1998) Assessing measurementinvariance in cross-national consumer research. Journal of ConsumerResearch, 25, 78–90.
Stone, D.N., Bryant, S.M. & Wier, B. (2010) Why are financialincentive effects unreliable? An extension of self-determinationtheory. Behavioral Research in Accounting, 22,105–132.
Tang, T.L.P. (1995) The development of a short money ethic scale: atti-tudes toward money and pay satisfaction revisited. Personality andIndividual Differences, 19, 809–816.
Tang, T.L.P. & Sutarso, T. (2006) The love of money and pay level sat-isfaction: measurement and functional equivalence in 29 geopoliticalentities around the world. Management and Organization Review, 2,423–452.
Tang, T.L.P., Sutarso, T., Akande, A., Allen, M.W., Alzubaidi, A.S.,Ansari, M.A., Arias, F.-G., Borg, M.G., Canova, L., Charles-Pauvers,B., Cheng, B.S., Chiu, R.K., Du, L.Z., Garber, I., Garcia de la Torre,C., Higgs, R.C., Ibrahim, A.H.S., Jen, C., Kazem, A.M., Kim, K.,Lim, V.K.G., Luna-Arocas, R., Malovics, E., Manganelli, A.M.,Moreira, A., Nnedum, A.U.O., Osagie, J.E., Osman-Gani, A., Pereira,F.C., Pholsward, R., Pitariu, H.D., Polic, M., Sardzoska, E., Skobic,P., Stembridge, A.F., Tang, T.L.N., Teo, T.S.H., Tombolani, M.,Trontelj, M., Urbain, C. & Vlerick, P. (2007) Doing Well by DoingGood: Does Economic Development Make a Difference. Paper pre-sented at the Academy of Management Annual Meetings, Philadel-phia, PA.
Tang, T.L.P., Sutarso, T., Davis, G.M.T.W., Dolinski, D., Ibrahim,A.H.S. & Wagner, S.L. (2008) To help or not to help? The goodsamaritan effect and the love of money on helping behavior. Journalof Business Ethics, 82, 865–887.
United Nations Development Programme (2011) Human DevelopmentReport 2011. [WWW document]. URL http://hdr.undp.org/en/media/HDR_2011_EN_Table1.pdf (accessed on 15 August 2013).
Veenhoven, R. (2011) States of Nations: Data File to Be Used for theCross-National Analysis of Happiness. World Database of Happiness,Erasmus University Rotterdam, the Netherlands. [WWW document].URL http://worlddatabaseofhappiness.eur.nl/statnat/statnat_fp.htm(accessed on 15 August 2013).
Wong, N., Rindfleisch, A. & Burroughs, J.E. (2003) Do reverse-wordeditems confound measures in cross-cultural consumer research? Thecase of the material values scale. Journal of Consumer Research, 30,72–91.
World Bank (2011a) Motor vehicles date. [WWW document].URL http://data.worldbank.org/indicator/IS.VEH.NVEH.P3/countries?order=wbapi_data_value_2010+wbapi_data_value+wbapi_data_value-last&sort=asc (accessed on 15 August 2013).
World Bank (2011b) Poverty gap. [WWW document]. URL http://search.worldbank.org/quickview?name=%3Cem%3EPoverty%3C%2Fem%3E+gap+at+%242+a+day+%28PPP%29+%28%25%29&id=SI.POV.GAP2&type=Indicators&cube_no=2&qterm=poverty+index(accessed on 15 August 2013).
Zwierlein, E. (1986) The problem of materialism in the philosophy ofSchopenhauer, Arthur – German – Primer, Hra. Philosophy andHistory, 19, 123–124.
Consumer needs and values across cultures L. Guo et al.
International Journal of Consumer Studies 37 (2013) 675–688© 2013 John Wiley & Sons Ltd
688
14706431, 2013, 6, Dow
nloaded from https://onlinelibrary.w
iley.com/doi/10.1111/ijcs.12047 by L
iberty University, W
iley Online L
ibrary on [15/11/2022]. See the Term
s and Conditions (https://onlinelibrary.w
iley.com/term
s-and-conditions) on Wiley O
nline Library for rules of use; O
A articles are governed by the applicable C
reative Com
mons L
icense
BUSI 604
Hofstede Analysis Comparison between the USA and “…” Example
Title: Hofstede Analysis Comparison between the USA and
Outline Example
Introduction
Seriously relevant graduate school research requires a question for which no ready answer is available. The research is conducted, to answer specific questions regarding a topic, problem, or issue for which the answers are not yet known. Let us focus on the concept of a topic. What do you want to know about a topic? Asking a topic as a question (or series of related questions) has several advantages:
Questions require answers. A topic is hard to cover completely because it typically encompasses too many related issues; but a question has an answer, even if it is ambiguous or controversial.
Questions give you a way of evaluating the evidence. A clearly stated question helps you decide which information will be useful. A broad topic may tempt you to stash away information that may be helpful, but you are not sure how. A question also makes it easier to know when you have enough information to stop your research and draft an answer.
A clear open-ended question calls for real research and thinking. Asking a question with no direct answer makes research and writing more meaningful to both you and your audience. Assuming that your research may solve significant problems or expand the knowledge base of a discipline involves you in more meaningful activity of community and scholarship.
In this course, the required research questions are open-ended and require a variety of accumulated data to develop answers. Your topic is a Business Cultural Dimensions Analysis of the nation you selected. You have been provided two specific research questions to guide you in the study of this topic which, if done well, will demonstrate you have attained an advanced measure of expertise in the topic. The research questions provide the framework of your analysis.
1. From the perspective of a Hofstede Analysis, what are the differences and similarities between and the USA?
2. What are the implications for USA businesses that wish to conduct business in ?
The example begins on the following page. Be sure to use the exact wordings in this outline for your APA level-headings.
EXAMPLE OUTLINEComment by Satterlee, Brian C (School of Business): Focus your research on the nation you selected at the beginning of the course.As you begin to research your paper, think of it as writing two highly inter-related papers. Remember, what you are really doing here is breaking the research assignment into smaller, more manageable components.
1. Research Question 1: From the perspective of a Hofstede Analysis, what are the differences and similarities between and the USA?Comment by Satterlee, Brian C (School of Business): The first “paper” is a report of the similarities and differences comparison regarding Hofstede Analyses of the nation you selected for Project 1 and the USA
· Hofstede Analysis of
· Hofstede Analysis of the USA
· Similarities
· Differences
2. Research Question 2: What are the implications for USA businesses that wish to conduct business in ?
· Conclusions regarding the similarities and differences
· How could USA managers use the above information to create and sustain competitive advantages when doing business with
· How might a Christian manager/leader prepare to address the differences and similarities while working in that nation?
Page 2 of 2
Criteria Ratings Points
ResearchQuestions1 & 2
70 to >62.0 pts
Advanced
All content requirements asstated in the instructionsare included in the paper.
62 to >49.0 pts
Proficient
Most content requirementsas stated in the instructionsare included in the paper.
49 to >0.0 pts
Developing
Some contentrequirements as stated inthe instructions areincluded in the paper.
0 pts
NotPresent
70 pts
Spelling,Grammar,CurrentAPAFormatting
30 to >26.0 pts
Advanced
Spelling and grammar arecorrect. Sentences arecomplete, clear, andconcise. Paragraphscontain appropriately variedsentence structures.Required Level headingsare perfect. 10 referencesare cited in current APAformat.
26 to >21.0 pts
Proficient
Some spelling and grammarerrors are present.Sentences are presentedwell. Paragraphs containsome varied sentencestructures. Required levelheadings are not perfect.Where applicable, 8references are cited withAPA formatting.
21 to >0.0 pts
Developing
Spelling and grammarerrors distract.Sentences areincomplete or unclear.Paragraphs are poorlyformed. Required levelheadings are notpresent. References areminimally or not cited incurrent APA format.
0 pts
NotPresent
30 pts
Total Points: 100
Hofstede Analysis Comparison between the USA and “…” GradingRubric | BUSI604_D03_202240
Personality Assesment
James W. Bland III
School of , Liberty University
Author Note
I have no known conflict of interest to disclose.
Correspondence concerning this article should be addressed to . Email:
Studies show that there are sixteen work personalities. Those personalities are known as the type of table, which the Publisher wrote, Consulting Psychologists Press, Inc. Palo Alto, Ca 94303. It shows how important it is to know identify your work type for a healthy, happy, and fulfilled work environment for yourself and others.
Keywords:
2
PERSONALITY ASSESSMENT2
Introverts, extroverts, sensors, intuitive
Personality Assessment
Keirsey. (n.d). Learn about the rational mastermind.
Daniel. (2017). Keirsey temperaments. temperaments/
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 1
Created by Christy Owen of Liberty University’s Online Writing Center
onlinewriting@liberty.edu; last date modified: November 7, 2021
Sample APA Paper: Professional Format for Graduate/Doctoral Students
Claudia S. Sample
School of Behavioral Sciences, Liberty University
Author Note
Claudia S. Sample (usually only included if author has an ORCID number)
I have no known conflict of interest to disclose.
Correspondence concerning this article should be addressed to Claudia S. Sample.
Email: cssample123456789@liberty.edu
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 2
Table of Contents
(Only Included for Easy Navigation; Hyperlinked for Quick Access)
Sample APA Paper: Professional Format for Graduate/Doctoral Students ……………………………… 6
Basic Rules of Scholarly Writing ……………………………………………………………………………………… 7
Brief Summary of Changes in APA-7 ………………………………………………………………………………… 8
Running Head, Author Note, and Abstract …………………………………………………………………………. 9
Basic Formatting Elements …………………………………………………………………………………………….. 10
Font ………………………………………………………………………………………………………………….. 10
Line Spacing ……………………………………………………………………………………………………… 10
Spaces After Punctuation …………………………………………………………………………………….. 11
Footnotes …………………………………………………………………………………………………………… 11
Heading Levels—Level 1 ………………………………………………………………………………………………. 11
Level 2 Heading …………………………………………………………………………………………………. 12
Level 3 Heading ………………………………………………………………………………………. 13
Level 4 Heading. Must be bolded and indented ½”. Add a period, one
space, and begin your content on the same line as shown here. ………………………………… 13
Level 5 Heading …………………………………………………………………. 13
Specific Elements of Academic Papers ……………………………………………………………………………. 13
Tables of Contents and Outlines …………………………………………………………………………… 13
Annotated Bibliographies ……………………………………………………………………………………. 14
Appendices ………………………………………………………………………………………………………… 14
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 3
Crediting Your Sources………………………………………………………………………………………………….. 15
Paraphrasing and Direct Quotes ……………………………………………………………………………. 15
Paraphrasing ……………………………………………………………………………………………. 16
Block Quotes …………………………………………………………………………………………… 16
How Often to Cite Your Source in Each Paragraph ………………………………………………… 17
Rule for Omitting the Year of Publication ……………………………………………………………… 17
Arranging the Order of Resources in Your Citations ………………………………………………. 17
Two Works by the Same Author in the Same Year …………………………………………………. 18
Two Works by Two Different Authors with the Same Last Name ……………………………. 18
Three or More Authors Cited In-Text ……………………………………………………………………. 18
Number of Authors in the Reference List ………………………………………………………………. 19
Numbers ………………………………………………………………………………………………………………………. 19
Displaying Titles of Works in-Text …………………………………………………………………………………. 19
Primary Sources versus Secondary Sources ……………………………………………………………………… 20
Personal Communications ……………………………………………………………………………………………… 20
Resources Canonically Numbered Sections (i.e., the Bible and Plays) …………………………………. 21
Bible and other Classical Works …………………………………………………………………………… 21
Plays …………………………………………………………………………………………………………………. 22
Lectures and PowerPoints ………………………………………………………………………………………………. 22
Dictionary Entries …………………………………………………………………………………………………………. 23
Changes in Reference Entries …………………………………………………………………………………………. 23
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 4
Electronic Sources ………………………………………………………………………………………………………… 24
Adding Color ……………………………………………………………………………………………………………….. 24
Self-Plagiarism ……………………………………………………………………………………………………………… 25
Final Formatting Tweaks ……………………………………………………………………………………………….. 26
Exhaustive Reference List Examples & Additional Helpful Resources ………………………………… 26
Conclusion …………………………………………………………………………………………………………………… 29
References ……………………………………………………………………………………………………………………. 30
Appendix ……………………………………………………………………………………………………………………… 40
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 5
Abstract
Begin your abstract at the left margin. This is the only paragraph that should not be indented.
Unless otherwise instructed, APA recommends an abstract be no more than 250 words. It should
generally not contain any citations or direct quotes. This should be a tight, concise summary of
the main points in your paper, not a step-by-step of what you plan to accomplish in your paper.
Avoid phrases such as “this paper will,” and just structure your sentences to say what you want
to say. The following three sentences exemplify a good abstract style: There are many
similarities and differences between the codes of ethics for the ACA and the AACC. Both include
similar mandates in the areas of —-, —, and —. However, each differs significantly in the areas
of —, —, and —. For more detailed information, see “Writing an Abstract” at
https://www.liberty.edu/casas/academic-success-center/wp-content/uploads/sites/28/2019/04/
Writing_an_Abstract_Revised_2012.pdf (note that you would not include any links in your
abstract). This is just now at 168 words, so eyeball how brief your abstract must be. Think of
your paper as a movie you want to sound enticing, and the abstract as the summary of the plot
you would share to draw people’s interest into wanting to come and see your movie. You want to
really hook and intrigue them. What you have to say is important! Remember to stay under 250,
words. Keywords highlight the search terms someone would use to find your paper in a database.
Keywords: main words, primary, necessary, search terms
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 6
Sample APA Paper: Professional Format for Graduate/Doctoral Students
The title of your paper goes on the top line of the first page of the body (American
Psychological Association [APA], 2019, section 2.11). It should be centered, bolded, and in title
case (all major words—usually those with four+ letters—should begin with a capital letter)—see
p. 51 of your Publication Manual of the American Psychological Association: Seventh Edition
(APA, 2019; hereinafter APA-7). It must match the title that is on your title page (see last line on
p. 32). As shown in the previous sentence, use brackets to denote an abbreviation within
parentheses (bottom of p. 159). Write out the full name of an entity or term the first time
mentioned before using its acronym (see citation in first sentence in this paragraph), and then use
the acronym throughout the body of the paper (section 6.25).
There are many changes in APA-7. One to mention here is that APA-7 allows writers to
include subheadings within the introductory section (APA, 2019, p. 47). Since APA-7 now
regards the title, abstract, and term “References” to all be Level-1 headings, a writer who opts to
include headings in his or her introduction must begin with Level-2 headings as shown above
(see section 2.27) for any divisions within the introductory section.
If you do choose to include headings in your introduction section (which is optional), be
sure to include two or more subheadings, since APA (2019) forbids stand-alone heading levels.
A second notable change in APA-7 is that writers are no longer required to cite their source every
single sentence that content from it is mentioned (section 8.1). As demonstrated in this paper,
since all of the content (other than the examples included for illustration and reference-entry
variation purposes) comes directly from the APA-7 itself, citations to the APA-7 are only
included for the first instance in each paragraph. Section and/or page numbers are included
parenthetically throughout for the sake of students who desire to know exactly where the stated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 7
rule appears in the APA-7 itself. In your academic papers, however, it is critical to include the
required author(s) and year, as applicable, for all citations that are included; this may include
more than one citation for each resource per paragraph, as required to avoid any confusion about
the source of that content.
Basic Rules of Scholarly Writing
Most beginning students have difficulty learning how to write papers and also format
papers correctly using the seventh edition of the APA manual. However, the Liberty University
Online Writing Center’s (OWC) mission includes helping students learn how to be autonomous,
proficient writers. The OWC also provides students with templates to help them with basic
formatting elements, but this sample paper is designed to help graduate and doctoral students
learn to master APA rules and formatting on their own, which will prove helpful as they progress
in their studies and work toward future publication in scholarly journals.
For the purpose of instruction, this paper will use second person (you, your), but third
person (this author) must be used in most student papers. First person (I, me, we, us, our) is not
generally permitted in academic papers. Students should refrain from using first or second person
in college courses (even though the APA manual encourages this in other writing venues) unless
the assignment instructions clearly permit such (as in the case of personal reflection sections or
life histories). If in doubt, students should clarify with their professors.
APA-7 delineates separate rules and guidelines between “student” and “professional”
writers (APA, 2019). Because a primary purpose of graduate and doctoral studies is to prepare
those students to publish professionally, Liberty University has decided to have undergraduate
students follow APA-7’s guidelines for “student papers,” and graduate/doctoral students follow
APA-7’s guidelines for “professional papers.” Separate templates are available for each level.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 8
This sample paper illustrates and discusses the rules and formatting of professional papers, as
required for all Liberty University graduate and doctoral courses using APA-7 style.
Brief Summary of Changes in APA-7
Most of these changes will be discussed in more detail below; this is just a very brief
overview here. APA-7 reverts back to only one space after closing punctuation in the body of the
paper (APA-6 required two spaces; APA, 2019, section 6.1). Student (undergraduate) papers no
longer include a running head or abstract (sections 2.2 and 2.8); professional (graduate/doctoral)
papers require an abstract but the running head is now the same on all pages (the added phrase
“Running head:” from APA-6 has been eliminated; see section 2.8). Title pages are different for
both student and professional formats. The title of a paper is no longer limited to 12 words
(section 2.4).
Citations of all resources with three or more authors now use the first author’s last name
and the term et al. (APA, 2019, section 8.17). Reference entries must name up to the first 19
authors before adding an ampersand and ellipsis (up from APA-6’s six authors; section 9.8).
APA-7 omits the phrase DOI and instead standardizes DOIs to be presented in hyperlink format
(i.e., https://doi.org/10.1177%2F1524838017742386; section 9.35). Formatting guidelines for
annotated bibliographies are included in APA-7 (section 9.51), as well as expanded and
standardized reference entry examples. As discussed above, it is no longer necessary to cite a
source every single time you refer to content gleaned from it as long as it is clear the content
comes from that source (section 8.1); APA-7 also expanded the specific location noted in the
citation to include page, paragraph, section (as used throughout this sample paper, to direct the
student to the exact relevant content), chapter, timestamp, etc. (section 8.13).
APA-7 allows for “self-plagiarism” (clarified and defined below). It also invites writers to
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 9
highlight the most relevant work first, rather than just present all works in alphabetical order
(APA, 2019, section 8.12).
Heading-level formatting has changed, and APA-7 provides more flexibility in font and
line spacing (APA, 2019). The Bible must now be included in the reference list and its citations
must include the editor’s details and year (section 8.28); there are also new rules for dictionary
entries. Publisher city and state details are omitted from all reference entries except those
involving presentations or conferences, as is the phrase “retrieved from.” Hyperlinks should be
live, but they may be either presented as blue underlining or plain black text.
Running Head, Author Note, and Abstract
APA (2019) delineates separate formatting requirements for what it terms “student” and
“professional” papers. Its descriptions for those labels, however, suggests that it regards
undergraduate-level writing to fall within the student purview, and graduate/doctoral-level
writing (including dissertations and theses) to fall within the professional purview. Since a
significant goal in graduate and post-graduate studies is preparing those students to publish in
scholarly journals at and beyond graduation, it makes sense to train those students in the
formatting that is required for professionals. As such, Liberty University has opted to require the
APA-7’s “student” version format for all undergraduate assignments using APA, and its
“professional” version for all graduate and doctoral assignments. To that end, this being the
sample paper for professional formatting, it includes the additional elements required for such: a
running head (same on all pages), an author’s note, and an abstract. Graduate and doctoral
students will use this format. Note that the first “paragraph” under the author’s note is generally
only included if the author has an ORCID number, which most students will not have. However,
it is included in this sample paper and the corresponding template because the purpose of these
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 10
resources is to prepare students to publish manuscripts post-graduation. The student’s full
address, however, is intentionally omitted from the Liberty University template and this sample
paper for privacy and safety reasons, since student papers are often unfortunately published
online and disclosing their home addresses could pose safety risks.
Basic Formatting Elements
Font
APA-7 does not prescribe a specific font or size (APA, 2019, section 2.19) but rather
allows for some choice (e.g., 12-point Times New Romans, 11-point Calibri, 11-point Arial, 11-
point Georgia, or 10-point Lucinda Sans Unicode). Most journals and academic institutions will
have a preference, however, as even APA-7 acknowledges on p. 44. For this reason—and
because font size can easily be changed if an editor interested in publishing a student’s work
prefers a different font—Liberty University recommends that students use 12-point Times New
Romans or 11-point Calibri font for the body text in all academic papers. Data in charts, figures,
and tables should be presented in 8- to 14-point size in either Calibri, Arial, or Lucinda Sans
Unicode font. Students are not permitted to use any fonts such as script, calligraphy, poster,
decorative, or others not found in published scholarly journals. Since APA-7 itself authorizes a
variety of fonts and sizes, assignments will be gauged by word count rather than page count.
Word count constitutes the number of words within the body of the paper, and excludes the title
page, abstract, reference list, appendices, and other supplemental resources.
Line Spacing
APA-7 adds extra/blank lines on the title page (APA, 2019, sections 2.5, 2.7, 2.21). It also
specifies that spacing in tables and figures may be single-, 1-1/2-, or double-spaced; equations
can be triple- or quadruple-spaced. Footnotes, when used at the bottom of a page, should be
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 11
single-spaced (section 2.21).
Spaces After Punctuation
APA-7 reverts back to just one space after closing punctuation in the body of the paper, as
well as in reference entries (APA, 2019, section 6.1). Ordinarily, it would be improper to have a
paragraph with only one sentence, though APA itself asserts that for its purposes “sentences and
paragraphs of any length are technically allowed.”1
Footnotes
This leads to another new rule in APA-7, one allowing the inclusion of footnotes (APA,
2019, section 2.13). Footnotes should be use very sparingly and are appropriate to include
information such as that in the prior section to alert the reader to supplemental material that is
available online for that thought. Though APA-7 authorizes placement of footnote content either
at the bottom of the page (as in this sample paper) or on a separate page after the reference list
(section 2.21), Liberty University recommends that student place them, when used, at the bottom
of the page, as shown here.
Heading Levels—Level 1
This sample paper uses primarily two levels of headings (Levels 1 and 2). APA style,
however, has five heading levels, which will be demonstrated briefly for visual purposes. See
section 2.27 of your APA-7 (APA, 2019) for more details on heading levels and formatting. In
APA-7, all heading levels are now bolded and in title case (capitalize each major word—usually
those with four or more letters, including hyphenated compound words). Do not capitalize
articles (a, an, the) in headings unless they begin a title or follow a colon. Level 1 headings are
centered, with the content falling on the line beneath each, in standard paragraph format.
1 See https://blog.apastyle.org/apastyle/2016/05/index.html
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 12
Many students misunderstand that you progress from Level 1 to Level 2 to Level 3 to
Level 4 to Level 5, but that is not correct. In fact, your paper may have only Level 1 headings, or
just Levels 1 and 2. The rule of thumb is that you must have at least two of each heading level
that you use, otherwise omit that heading level.
Headings are basically styling ways of organizing your paper, without using an outline
format. APA specifies five levels of headings; you would likely never use Level 5 and only very
rarely use Level 4 as a student. Think of each level as the different levels in an outline. Roman
numerals, for example, would be Level 1 headings. Capital letters would be Level 2 headings.
Numerals would be Level 3 headings. Lowercase letters would be Level 4. And lowercase
Roman numerals would be Level 5. You must always have two or more of each subheading, but
you do not need every level. You start with Level 1 and work down from that (but not
consecutive 1-2-3-4-5). Under a Level 1, you would either have two+ Level 2 headings or none
at all (just one big section in paragraphs before the next Level 1 section).
Special note about conclusion sections: Please note that some of the sample papers
published by APA to demonstrate proper APA-7 format (including the “professional” sample on
pp. 50-60 of the APA-7 manual) depict the “Conclusion” section with a Level-2 heading. This is
limited to empirical papers that are being submitted for publication in scholarly journals, as those
conclusions pertain to the “Discussion” sections in such papers and are not conclusions of the
overall papers themselves. Conclusions in academic papers at Liberty University will be Level 1
headings (including dissertations and theses, which are divided by chapters, unlike journal article
manuscripts).
Level 2 Heading
Level 2 headings are left-justified (APA, 2019, p. 48). The supporting information is
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 13
posed in standard paragraph form beneath it. Never use only one of any level of heading. You
must use two or more of any level you use, though not every paper will require more than one
level. The heading levels are simply demonstrated here for visual purposes, but you would
always have two or more of each under a larger heading, as shown throughout all the other
sections of this sample paper.
Level 3 Heading
Level 3 headings are bolded, left-justified, and italicized; the content falls on the line
underneath, as with Levels 1 and 2.
Level 4 Heading. Must be bolded and indented ½”. Add a period, one space, and begin
your content on the same line as shown here.
Level 5 Heading. Same as Level 4, but also italicized. Despite heavy writing experience,
this author has never used Level 5 headings.
Specific Elements of Academic Papers
Tables of Contents and Outlines
APA (2019) does not regulate every type of paper and some elements in various
assignments are not addressed in the APA-7 manual, including outlines and tables of content. In
those cases, follow your professor’s instructions and the grading rubric for the content and
format of the outline or annotations, and use standard APA formatting for all other elements
(such as running head, title page, body, reference list, 1″ margins, double-spacing, permitted
font, etc.). Note that most academic papers will not require a table of contents, nor would one be
appropriate. One was included in this paper simply for ease-of-access so students could go
directly to the content they want to see. Generally speaking, no table of contents would be
necessary for papers less than 20 pages of content, unless otherwise required by your professor.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 14
That being said, when organizing outlines in APA format, set your headings up in the
proper levels (making sure there are at least two subheadings under each level), and then use
those to make the entries in the outline. As discussed above, Level 1 headings become uppercase
Roman numerals (I, II, III), Level 2 headings become capital letters (A, B, C), Level 3 headings
become numbers (1, 2, 3), Level 4 headings become lowercase letters (a, b, c), and Level 5
headings become lowercase Roman numerals (i, ii, iii). Many courses now require “working
outlines,” which are designed to have the bones and foundational framework of the paper in
place (such as title page, abstract, body with title, outline/heading divisions, supporting content
with citations, and references), without the full “meat” that fills out and forms a completed paper.
Annotated Bibliographies
Many Liberty University courses also now require students to prepare and submit an
annotated bibliography as a foundational step to building a research paper. There is significant
merit in these assignments, as they teach students to critique the resources they have found and
rationalize why each is relevant for their paper’s focus. APA (2019) includes a section on
annotated bibliographies (9.51; see the example provided on p. 308). The appendix attached to
this sample paper also includes a sample annotated bibliography.
Appendices
Appendices, if any, are attached after the reference list (APA, 2019, section 2.14). You
must refer to them (i.e., “callout”) in the body of your paper so that your reader knows to look
there (see the yellow-highlighted callouts to Table 1 on p. 54 and to Footnote 1 on p. 55 of your
APA-7 for visuals on how this should appear in your paper). The word “Appendix” is singular;
use it to refer to individual appendices. APA-7 regards it as a Level 1 heading so it should be
bolded. I attached a sample Annotated Bibliography as a visual aid (see Appendix). You will see
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 15
that I included the title “Appendix” at the top of the page and formatted it in standard APA
format beneath that. Because I only included one appendix, it is simply titled as such. If there are
more appendices, assign a letter to each and denote each by that: “Appendix A” and “Appendix
B.”
Crediting Your Sources
Paraphrasing and Direct Quotes
Paraphrasing is rephrasing another’s idea in one’s own words by changing the wording
sufficiently without altering the meaning (remember not to just change a word here or there or
rearrange the order of the original source’s wording). Quoting is using another’s exact words.
Both need to be cited; failure to do so constitutes plagiarism. Include the author(s) and year for
paraphrases, and the author(s), year, and page or paragraph number for direct quotes. APA-7 also
expands this to include figure number, time stamp, etc.—whatever detail is necessary to get the
reader directly to that content. Page numbers should be used for any printed material (books,
articles, etc.), and paragraph numbers should be used in the absence of page numbers (online
articles, webpages, etc.; see APA, 2019, section 8.13). Use p. for one page and pp. (not italicized
in your paper) for more than one (section 8.25). Use para. for one paragraph and paras. (also not
italicized in your paper) for two or more (section 8.28). For example: (Perigogn & Brazel, 2012,
pp. 12–13) or (Liberty University, 2019, para. 8). Section 8.23 of the APA (2019) manual
specifies that it is not necessary to include a page or paragraph number for paraphrases (just for
direct quotes), but writers may choose to do so to help their readers find that content in the cited
resource.
When naming authors in the text of the sentence itself (called a narrative citation), use the
word “and” to connect them. For example, Perigogn and Brazel (2012) contemplated that . . .
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 16
Use an ampersand (&) in place of the word “and” in parenthetical citations and reference lists:
(Perigogn & Brazel, 2012).
Paraphrasing
Only use quotes when the original text cannot be said as well in your own words or
changing the original wording would change the author’s meaning. You cannot simply change
one word and omit a second; if you paraphrase, the wording must be substantially different, but
with the same meaning. Regardless, you would need to cite the resource you took that
information from. For example, Benoit et al. (2010) wrote that “although, a link between
attachment and posttraumatic stress disorder (PTSD) symptoms has been established, the
mechanisms involved in this link have not yet been identified” (p. 101). A paraphrase for that
quote might be: A link between dysfunctional attachment and the development of PTSD has
been made, though there is insufficient data to determine exactly how this mechanism works
(Benoit et al., 2010).
Block Quotes
Quotes that are 40 or more words must be blocked, with the left margin of the entire
quote indented ½ inch. Maintain double-spacing of block quotes. APA prefers that you introduce
quotes but note that the punctuation falls at the end of the direct quote, with the page number
outside of that (which is contrary to punctuation for non-blocked quotes). For example, Alone
(2008) claims:2
Half of a peanut butter sandwich contains as much bacteria as the wisp of the planet
Mars. Thus, practicality requires that Mrs. Spotiker nibble one bit at a time until she is
assured that she will not perish from ingesting it too quickly. (p. 13)
2 Note that there are no quotation marks for block quotes, as shown in the example.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 17
Usually quotes within quotes use single quotation marks; however, use double quotation marks
for quotes within blocked quotes, since there are no other quotation marks involved. Also
understand that direct quotes should be used sparingly in scholarly writing; paraphrasing is much
preferred in APA format (APA, 2019, section 8.23), as it demonstrates that you read, understood,
and assimilated other writers’ content into one cohesive whole.
How Often to Cite Your Source in Each Paragraph
As already mentioned above, APA’s (2019) new official rule is that you no longer must
cite your source every single time you refer to material you gleaned from it (section 8.1). It is
now acceptable to cite your source the first time you refer to content from it in your paragraph,
and then not again in that same paragraph unless your phrasing does not make the source of your
content clear. This is demonstrated throughout this sample paper.
Rule for Omitting the Year of Publication
That being said, APA (2019) has clarified its special rule that excludes the year of
publication in subsequent narrative in-text citations (when you name the authors in the text of the
sentence itself), after the first narrative citation in each paragraph. It should continue to appear in
all parenthetical citations (see section 8.16). For example, Alone (2008) portrays imagery of Mrs.
Spotiker. This includes her devouring a peanut butter sandwich (Alone, 2008). Alone conveys
this through the lens of astronomy. Note that the year of publication was omitted from the second
narrative citation (underlined for visual purposes).
Arranging the Order of Resources in Your Citations
If the material you cited was referred to in multiple resources, separate different sets of
authors with semicolons, arranged in the order they appear (alphabetically by the first author’s
last name) in the reference list (i.e., Carlisle, n.d.-a; Prayer, 2015) (APA, 2019, section 8.12).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 18
APA-7 now invites writers to prioritize or highlight one or more sources as most prominent or
relevant for that content by placing “those citations first within parentheses in alphabetical order
and then insert[ing] a semicolon and a phrase, such as ‘see also,’ before the first of the remaining
citations” (APA., 2019, p. 263)—i.e., (Cable, 2013; see also Avramova, 2019; De Vries et al.,
2013; Fried & Polyakova, 2018). Periods are placed after the closing parenthesis, except with
indented (blocked) quotes.
Two Works by the Same Author in the Same Year
Authors with more than one work published in the same year are distinguished by lower-
case letters after the years, beginning with a (APA, 2019, section 8.19). For example, Double
(2008a) and Double (2008b) would refer to resources by the same author published in 2008.
When a resource has no date, use the term n.d. followed by a dash and the lowercase letter (i.e.,
Carlisle, n.d.-a and Carlisle, n.d.-b; see APA, 2019, section 8.19).
Two Works by Two Different Authors with the Same Last Name
Citations in the body of the paper should include only the last names, unless you have
two or more resources authored by individuals with the same last name in the same year (or are
citing a personal communication). When there are two different authors with the same last name
but different first names who published in the same year, include the first initials: Brown, J.
(2009) and Brown, M. (2009) (APA, 2019, section 8.20).
Three or More Authors Cited In-Text
When referring to material that comes from three or more authors, APA-7 now requires
that all citations name just the first author’s last name followed by the words et al. (without
italics) (APA, 2019, section 8.17). Et al. is a Latin abbreviation for et alii, meaning “and others,”
which is why the word “al.” has a period, whereas “et” does not. Alone et al. (2011) stipulated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 19
that peacocks strut. Every single time I refer to their material, I would apply APA-7’s rule: Alone
et al. (2011) or (Alone et al., 2011). Since et al. denotes plural authors, the verb must be plural to
match, too: Alone et al. (2011) are… This applies to all citations within the body of the paper
with three or more authors.
Number of Authors in the Reference List
For resources with 20 or fewer authors in the reference list, write out all of the authors’
last names with first and middle initials, up to and including the 20th author (APA, 2019, section
9.8). APA-7 has a special rule for resources with 21 or more authors: Write out the first 19
authors’ last names with initials, insert an ellipsis (…) in place of the ampersand (&), and finish
it with the last name and initials of the last author. See example #4 provided on page 317 of your
APA-7, as well as this paper’s reference list for visuals of these variances (Acborne et al. 2011;
Kalnay et al., 1996).
Numbers
Numbers one through nine must be written out in word format (APA, 2019, section 6.33),
with some exceptions (such as ages—see section 6.32). Numbers 10 and up must be written out
in numerical format (section 6.32). Always write out in word format any number that begins a
sentence (section 6.33).
Displaying Titles of Works in-Text
The names of journals, books, plays, and other long works, if mentioned in the body of
the paper, are italicized in title case (APA, 2019, section 6.17). Titles of articles, lectures, poems,
chapters, website articles, and songs should be in title case, encapsulated by quotation marks
(section 6.7). The year of publication should follow the author’s name, whether in narrative or
parenthetical format: Perigogn and Brazel (2012) anticipated…, or (Perigogn & Brazel, 2012).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 20
The page or paragraph number must follow after the direct quote. Second (2015) asserted that
“paper planes can fly to the moon” (p. 13). You can restate that with a parenthetical citation as:
“Paper planes can fly to the moon” (Second, 2015, p. 13). Second (2011) is another resource by
the same author in a different year.
Primary Sources versus Secondary Sources
APA (2019) strongly advocates against using secondary sources; rather, it favors you
finding and citing the original (primary) resource whenever possible (section 8.6). On the rare
occasion that you do find it necessary to cite from a secondary source, both the primary (who
said it) and secondary (where the quote or idea was mentioned) sources should be included in the
in-text citation information. If the year of publication is known for both resources, include both
years in the citation (section 8.6). Only the secondary source should be listed in the reference
section, however. Use “as cited in” (without the quotation marks) to indicate the secondary
source. For example, James Morgan hinted that “goat milk makes the best ice cream” (as cited in
Alone, 2008, p. 117). Morgan is the primary source (he said it) and Alone is the secondary
source (he quoted what Morgan said). Only the secondary source is listed in the reference section
(Alone, and not Morgan) because if readers want to confirm the quote, they know to go to page
117 of Alone’s book.
Personal Communications
APA (2019) rationalizes the exclusion of references for information obtained through
personal communication (such as an interview, email, telephone call, postcard, text message, or
letter) in the reference list because your readers will not be able to go directly to those sources
and verify the legitimacy of the material. Instead, these items are cited only in the body of the
paper. You must include the individual’s first initial, his or her last name, the phrase “personal
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 21
communication” (without the quotation marks), and the full date of such communication (section
8.9). As with other citations, such citations may be either narrative or parenthetical. For example,
L. Applebaum advised him to dip pretzel rolls in cheese fondue (personal communication, July
13, 2015). The alternative is that he was advised to dip pretzel rolls in cheese fondue (L.
Applebaum, personal communication, July 13, 2015). Note that there is no entry for Applebaum
in the reference list below.
Resources Canonically Numbered Sections (i.e., the Bible and Plays)
These resources should be cited in book format (APA, 2019, Section 9.42). The Bible and
other religious works are generally regarded as having no author; an annotated version would be
treated as having an editor. Include republished dates as necessary. The OWC will publish a list
of reference entries for various Bible versions on its APA Quick Guide webpage.
Bible and other Classical Works
Works such as the Bible, ancient Greek or Roman works, and other classical works like
Shakespeare must be cited in the body of the paper (APA, 2019, section 8.28). APA-7 now also
requires that they be included in the reference list, too (section 9.42), which is a significant
change from APA-6. Republished dates are included as well (see section 9.41). As such, you
would add a parenthetical phrase at the end of your reference entry with the original publication
details; note that there should be no punctuation following such parenthetical content at the end
of a reference entry (the reference entries depicting this in the reference list below are correctly
punctuated).
Citations for the Bible will include the Bible version’s name in the author’s position (as
an anonymous work), original and republished years, and then the book chapter/verse (spelled
out) in place of the page number (i.e., King James Bible, 1769/2017, Genesis 3:8)—see sections
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 22
8.28 and 9.42. Note that APA (2019) requires book titles to be italicized in every venue,
including citations and reference entries. Because Liberty University is a distinctly-Christian
institution and many of its courses require biblical integration, most if not all of its students will
cite the Bible in virtually every course. The examples provided on pp. 274 and 325 of APA-7 are:
(note the italics in each)
Narrative citation: King James Bible (1769/2017)
Parenthetical citation: (King James Bible, 1769/2017, Song of Solomon 8:6)
Reference entry: King James Bible. (2017). King James Bible Online.
https://www.kingjamesbibleonline.org/ (Original work published 1769)
Plays
When citing plays, “cite the act, scene, and line(s), in a single string, separated by
periods. For example, ‘1.3.36-37’ refers to Act 1 Scene 3, Lines 36-37” (APA, 2019, section
8.28; see also example #37 on p. 325).
Lectures and PowerPoints
APA (2019) has expanded and standardized its rules for citations and reference entries in
an effort to best credit the original sources. It now includes rules for crediting content in course
or seminar handouts, lecture notes, and PowerPoint presentations (see #102 on p. 347). When
citing a PowerPoint presentation, include the slide number rather than the page number. For
purposes of Liberty University course presentations and lectures (which are not readily available
to the public), reference each as a video lecture with the URL (if available) for the presentation,
naming the presenter(s) in the author’s position. Include the course number, lecture title, and
enough details for others to identify it within that course, in a sort of book format, naming
Liberty University as publisher. Peters (2012) is an example of this in the reference list of this
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 23
paper. If the presenter for a Liberty University class lecture is not named, credit Liberty
University as the author; see Liberty University (2020) in the reference list below as an example.
Dictionary Entries
In keeping with its efforts to standardize reference entries, APA (2019) now requires
citation and referencing of word definitions from dictionaries to follow the same rules for
chapters in an edited book (see #47 and #48 on p. 328; section 8.13). As such, you will now
name either the individual, group, or corporate author of the dictionary in the author’s place (e.g.,
Merriam-Webster, n.d.). If you searched online, include the retrieval date and the URL to the
exact webpage. If you used a hard copy book, include the publisher details. The in-text citation
in the body of the paper would follow standard author/year format (e.g., Merriam-Webster, n.d.).
Changes in Reference Entries
There are a number of notable changes in APA-7 from past versions. For the most part,
these simplify and unify the formats to be more consistent across the different resource venues.
Some of these have already been discussed above (i.e., naming up to 19 authors’ names before
adding an ellipsis, and crediting authors and editors of classical works and dictionaries). Other
changes include italicizing names of webpages and website resources in the reference list (APA,
2019, section 6.22), as well as book titles even when named in the author’s position (such as
King James Bible). The city and state locations of publishers are no longer required; only include
those details “for works that are associated with a specific location, such as conference
presentations” (p. 297, section 9.31). Issue numbers are required for all journal articles that have
such, regardless of what page number each issue begins with (section 9.25). If two or more
publishers are listed on the copyright page, include all of them in the order listed, separated by
semicolons (section 9.29). Omit the word Author in the publisher’s place when it is the same as
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 24
the author (section 9.24).
Electronic Sources
Note that since the APA 6th edition was published in 2010, great strides have been made
in online and electronic resource accessibility, and APA’s position on electronic resources has
shifted to embrace this. More and more resources are available electronically through the
Internet. The advent of this increased availability has resulted in APA-7’s effort to standardize
the formatting of resources, which in turn simplifies them to some extent. All reference entries
follow the same basic details: Author(s), year of publication, name of resource, and location
details (i.e., either journal name/volume/issue/page numbers, or book publisher, or webpage).
APA (2019) requires inclusion of a Digital Object Identifier (DOI) in the references
whenever available (section 9.34); if not, then a webpage, if available. In keeping with its
unification of resources, APA-7 now standardizes all DOIs and URLs to be presented in
https://doi.org/10.1177%2F1524838017742386 format. The phrase “Retrieved from” is now
excluded except when the content may have changed (such as dictionary entries, Twitter profiles,
Facebook pages; see section 9.16). APA-7 requires all hyperlinks to be active (so your reader can
click on one to go directly to that webpage), but they may appear as either blue-underlined text
or simple black text (section 9.35). There should be no period after any URL. APA-7 no longer
requires authors to break long URLs with soft returns (hold down the Shift key and press the
Enter key) at forward slashes, periods, or underscores to avoid unsightly spacing gaps, but it may
be best to do so in academic papers.
Adding Color
Though APA (2019) authorizes writers to include the use of color in photographs and
figures (section 7.26), Liberty University discourages this in academic papers. It risks becoming
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 25
distracting for both students in their quest to be creative, and professors in their quest to focus on
academic content.
Self-Plagiarism
APA (2019) also invites writers to repurpose some of their work in future papers.
Specifically, APA-7 states that:
In specific circumstances, authors may wish to duplicate their previously used works
without quotation marks or citation …, feeling that extensive self-referencing is
undesirable or awkward and that rewording may lead to inaccuracies. When the
duplicated material is limited in scope, this approach is permissible. (p. 8.3)
APA-7 adds “Do not use quotation marks or block quotation formatting around your own
duplicated material” (p. 256).
Liberty University, however, has stringent rules against self-plagiarism, as do many
scholarly journals. Liberty University students receive grades for their class papers; those who
have received feedback and a grade from a prior professor on a prior paper have an advantage
over their classmates, both in having the benefit of that feedback/grade and in not having to write
a whole paper from scratch during the subsequent class. Student papers are also submitted to
SafeAssign to deter plagiarism. For these reasons, Liberty University expressly forbids students
using significant portions of a prior paper in a subsequent course (either a retake of the same
course or a new class altogether). It is conceivable that students who are building their
knowledge base in a subject matter—particularly at the graduate and post-graduate levels—
would reasonably justify incorporating brief excerpts from past papers into current ones. In such
case, Liberty University authorizes students to utilize APA-7’s disclosure (i.e., “I have previously
discussed”), along with a citation to the prior class paper and a reference entry (i.e., Owen, 2012;
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 26
Yoo et al., 2016). Such self-references and re-use of content from prior papers should be used
sparingly and disclosed fully in the current paper; that content should not constitute a significant
portion of any academic assignment, however.
Final Formatting Tweaks
The templates provided by Liberty University are already formatted with proper spacing,
margins, heading level structure, and hanging indents, as necessary. With the exceptions of the
title page, figures, and equations, papers in APA format should be double-spaced throughout,
with no extra spacing between lines. Academic papers at Liberty University should also be in
one of the accepted fonts throughout (recommended: Times New Romans, 12-point font).
Sometimes when you format your paper or cut-and-paste material into it, things get skewed. One
quick way to ensure that your paper appears correct in these regards is to do a final formatting
tweak after you have completed your paper. Hold down the “Ctrl” button and press the “A” key,
which selects and highlights all of the text in your paper. Then go to the Home tab in Microsoft
Word and make sure that whichever acceptable font/size you choose to use is selected in the Font
box. Next, click on the arrow at the bottom of the Paragraph tab. Set your spacing before and
after paragraphs to “0 pt” and click the “double” line spacing. The extra spacing required on the
title page is already programmed into the template and should not change even when you
complete these actions.
Exhaustive Reference List Examples & Additional Helpful Resources
The reference list at the end of this paper includes an example of a myriad of different
sources and how each is formatted in proper APA-7 format. One example of each of the primary
types of resources will be included in the reference list, as cited in the body of this paper.
Remember that, for purposes of this paper only, many of the sources cited in the body of the
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 27
paper were provided for illustrative purposes only and thus are fictional, so you will not be able
to locate them if you searched online. Nevertheless, in keeping with APA-7 style, all resources
cited in the body of the paper are included in the reference list and vice versa (except for personal
communications, per APA-7’s published exceptions). Be absolutely sure that every resource cited
in the body of your paper is also included in your reference list (and vice versa), excepting only
those resources with special rules, such as personal communications and primary sources you
could not access directly.
The reference list in this paper is fairly comprehensive and will include a book by one
author who also appears as one of many authors in another resource (Alone, 2008; Alone et al.,
2011); chapters in edited books (Balsam et al., 2019; Haybron, 2008; Perigogn & Brazel, 2012;
Weinstock et al., 2003); electronic version of book (Strong & Uhrbrock, 1923); electronic only
book (O’Keefe, n.d.); edited books with and without DOIs, with multiple publishers (Hacker
Hughes, 2017; Schmid, 2017); work in an anthology (Lewin, 1999); journal articles (Andrews,
2016; Carlisle, n.d.-a, n.d.-b; De Vries R. et al., 2013; McCauley & Christiansen, 2019);
newspaper article (Goldman, 2018; Guarino, 2017); online webpages (Liberty University, 2019;
Prayer, 2015); resource with corporate author as publisher (American Psychological Association,
2019); resources by two authors with the same last name but different first names in the same
year of publication (Brown, J., 2009; Brown, M., 2009); two resources by same author in the
same year (Double, 2008a, 2008b; Carlisle, n.d.-a, n.d.-b); two resources by the same author in
different years (Second, 2011, 2015); resource with 20 authors (maximum allowed by APA-7
before special rule applies) (Acborne et al., 2011); resource with 21 or more authors (Kalnay et
al., 1996); dictionary entries (American Psychological Association, n.d.; Graham, 2019;
Merriam-Webster, n.d.); Liberty University class lecture using course details (Peters, 2012);
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 28
PowerPoint slides or lecture notes, not including course details (Canan & Vasilev, 2019); citing a
student’s paper submitted in a prior class, in order to avoid self-plagiarism (Owen, 2012);
unpublished manuscript with a university cited (Yoo et al., 2016); code of ethics (American
Counseling Association, 2014); diagnostic manual (American Psychiatric Association, 2013);
religious and classical works, including the Bible (Aristotle, 350 BC/1994; King James Bible,
1769/2017; Shakespeare, 1623/1995); dissertation or thesis (Hollander, 2017; Hutcheson, 2012);
review of a book (Schatz, 2000); video (Forman, 1975); podcast (Vedentam, 2015); recorded
webinar (Goldberg, 2018); YouTube or other streaming video (University of Oxford, 2018); clip
art or stock image (GDJ, 2018); map (Cable, 2013); photograph (McCurry, 1985); data set (Pew
Research Center, 2018); measurement instrument (Friedlander et al., 2002); manual for a test,
scale, or inventory (Tellegen & Ben-Porah, 2011); test, scale, or inventory itself (Project
Implicit, n.d.); report by a government agency or other organization (National Cancer Institute,
2018); report by individual authors at a government agency or other organization (Fried &
Polyakova, 2018); annual report (U.S. Securities and Exchange Commission, 2017); conference
session (Fistek et al., 2017); and webpages (Avramova, 2019; Centers for Disease Control and
Prevention, 2018; National Nurses United, n.d.; U.S. Census Bureau, n.d.).
Lastly, below are a few webpages that address critical topics, such as how to avoid
plagiarism and how to write a research paper. Be sure to check out Liberty University’s Online
Writing Center (https://www.liberty.edu/online/casas/writing-center/) for more tips and tools, as
well as its Facebook page (https://www.facebook.com/groups/LUOWritingCenter). Remember
that these links are only provided for your easy access and reference throughout this sample
paper, but web links and URLs should never be included in the body of scholarly papers; just in
the reference list. Writing a research paper (https://www.youtube.com/watch?v=zaa-PTexW2E
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 29
or https://www.youtube.com/watch?v=KNT6w8t3zDY and avoiding plagiarism
(https://www.youtube.com/watch?v=VeCrUINa6nU).
Conclusion
The conclusion to your paper should provide your readers with a concise summary of the
main points of your paper (though not via cut-and-pasted sentences used above). It is a very
important element, as it frames your whole ideology and gives your readers their last impression
of your thoughts. Be careful not to introduce new content in your conclusion.
After your conclusion, if you are not using the template provided by the Online Writing
Center, insert a page break at the end of the paper so that the reference list begins at the top of a
new page. Do this by holding down the “Ctrl” key and then clicking the “Enter” key. You will go
to an entirely new page in order to start the reference list. The word “References” (not in
quotation marks) should be centered and bolded. Items in the reference list are presented
alphabetically by the first author’s last name and are formatted with hanging indents (the
second+ lines of each entry are indented 1/2” from the left margin). APA authorizes the use of
singular “Reference” if you only have one resource.3 Students would, of course, NOT include
any color-coding or footnotes in their reference entries. However, for the sake of clarity and
ease in identifying what each entry represents, each one included in the reference list of this
sample paper is color-coordinated to its corresponding footnote, with a brief description of what
each depicts.
3 https://apastyle.apa.org/instructional-aids/creating-reference-list.pdf
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 30
References
Acborne, A., Finley, I., Eigen, K., Ballou, P., Gould, M. C., Blight, D., Callum, M., Feist, M.,
Carroll, J. E., Drought, J., Kinney, P., Owen, C., Owen, K., Price, K., Harlow, K.,
Edwards, K., Fallow, P., Pinkley, O., Finkel, F., & Gould, P. P. (2011). The emphasis of
the day after tomorrow. Strouthworks. 4
Alone, A. (2008). This author wrote a book by himself. Herald Publishers. 5
Alone, A., Other, B., & Other, C. (2011). He wrote a book with others, too: Arrange
alphabetically with the sole author first, then the others. Herald Publishers. 6
American Counseling Association. (2014). 2014 ACA code of ethics.
https://www.counseling.org/knowledge-center 7
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). https://www.doi.org/10.1176/appi.books.9780890425596 8
American Psychological Association. (n.d.). Positive transference. In APA dictionary of
psychology. Retrieved August 31, 2019, from https://dictionary.apa.org/positive-
transference 9
American Psychological Association. (2019). Publication manual of the American Psychological
Association (7th ed.). 10
4 Resource with 20 authors (maximum allowed by APA before special rule applies).
5 Entry by author who also appears as one of many authors in another resource (single author
appears first in list).
6 Multiple authors appear after same single-author resource.
7 Code of ethics.
8 Diagnostic manual.
9 Entry in a dictionary, thesaurus, or encyclopedia, with group author.
10 Resource with corporate author as publisher.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 31
Andrews, P. M. (2016). Congruence matters. Educational Leadership, 63(6), 12-15. 11
Aristotle. (1994). Poetics (S. H. Butcher, Trans.). The internet Classics Archive.
http://classics.mit.edu/Aristotle/poetics.html (Original work published ca. 350 B.C.E.) 12
Avramova, N. (2019, January 3). The secret to a long, happy, heathy life? Think age-positive.
CNN. https://www.cnn.com/2019/01/03/health/respect-toward-elderly-leads-to-long-life-
intl/index.html 13
Balsam, K. F., Martell, C. R., Jones, K. P., & Safren, S. A. (2019). Affirmative cognitive
behavior therapy with sexual and gender minority people. In G. Y. Iwamasa & P. A.
Hays (Eds.), Culturally responsive cognitive behavior therapy: Practice a supervision
(2nd ed., pp. 287-314). American Psychological Association.
https://doi.org/10.1037/0000119-012 14
Benoit, M., Bouthillier, D., Moss, E., Rousseau, C., & Brunet, A. (2010). Emotion regulation
strategies as mediators of the association between level of attachment security and PTSD
symptoms following trauma in adulthood. Anxiety, Stress, and Coping, 23(1), 101-118.
http://dx.doi.org/10.1080/10615800802638279
Brown, J. (2009). Ardent anteaters. Brockton.
Brown, M. (2009). Capricious as a verb. Journal of Grammatical Elements, 28(6), 11-12. 15
11 Journal article without DOI, from most academic research databases or print version.
12 Ancient Greek or Roman work.
13 Webpage on a news website.
14 Chapter in an edited book with DOI.
15 Resources by two authors with the same last name but different first names in the same year of
publication. Arrange alphabetically by the first initials.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 32
Cable, D. (2013). The racial dot map [Map]. University of Virginia, Weldon Cooper Center for
Public Service. https://demographics.coopercenter.org/Racial-Dot-Map 16
Canan, E., & Vasilev, J. (2019, May 22). [Lecture notes on resource allocation]. Department of
Management Control and Information Systems, University of Chile. https:// uchilefau.
academia.edu/ElseZCanan 17
Carlisle, M. A. (n.d.-a). Erin and the perfect pitch. Journal of Music, 21(3), 16-17. http:// make-
sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include 18
Carlisle, M. A. (n.d.-b). Perfect pitch makes sweet music. Journal of Music, 24(8), 3-6. http://
make-sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include
Centers for Disease Control and Prevention. (2018, January 23). People at high risk of
developing flu-related complications. https://www.cdc.gov/flu/highrisk/index.htm 19
De Vries R., Nieuwenhuijze, M., Buitendijk, S. E., & the members of Midwifery Science Work
Group. (2013). What does it take to have a strong and independent profession of
midwifery? Lessons from the Netherlands. Midwifery, 29(10), 1122-1128.
https://doi.org/10.1016/j.midw.2013.07.007 20
Double, C. (2008a). This is arranged alphabetically by the name of the title. Peters.
Double, C. (2008b). This is the second (“the” comes after “arranged”). Peters. 21
16 Map.
17 PowerPoint slides or lecture notes.
18 Online journal article with a URL and no DOI; also depicts one of two resources by the same
author with no known publication date.
19 Webpage on a website with a group author.
20 Journal article with a DOI, combination of individual and group authors.
21 Two resources by same author in the same year. Arrange alphabetically by the title and then
add lowercase letters (a and b, respectively here) to the year.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 33
Fistek, A., Jester, E., & Sonnenberg, K. (2017, July 12-15). Everybody’s got a little music in
them: Using music therapy to connect, engage, and motivate [Conference session].
Autism Society National Conference, Milwaukee, WI, United States.
https://asa.confex.com/asa/2017/webprogramarchives/Session9517.html 22
Forman, M. (Director). (1975). One flew over the cuckoo’s nest [Film]. United Artists. 23
Fried, D., & Polyakova, A. (2018). Democratic defense against disinformation. Atlantic Council.
https://www.atlanticcouncil.org/in-depth-research-reports/report/democratic-defense-
against-disinformation/ 24
Friedlander, M. L., Escudero, V., & Heatherton, L. (2002). E-SOFTA: System for observing
family therapy alliances [Software and training videos] [Unpublished instrument].
http://www.softa-soatif.com/ 25
GDJ. (2018). Neural network deep learning prismatic [Clip art]. Openclipart.
https://openclipart.org/detail/309343/neural-network-deep-learning-prismatic 26
Goldberg, J. F. (2018). Evaluating adverse drug effects [Webinar]. American Psychiatric
Association. https://education.psychiatry.org/Users/ProductDetails.aspx?
ActivityID=6172 27
Goldman, C. (2018, November 28). The complicate calibration of love, especially in adoption.
22 Conference session.
23 Video.
24 Report by individual authors at a government agency or other organization.
25 Measurement instrument.
26 Clip art or stock image.
27 Webinar, recorded.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 34
Chicago Tribune. 28
Graham, G. (2019). Behaviorism. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy
(Summer 2019 ed.). Stanford University.
https://plato.stanford.edu/archives/sum2019/entries/behaviorism 29
Guarino, B. (2017, December 4). How will humanity react to alien life? Psychologists have some
predictions. The Washington Post. https://www.washingtonpost.com/news/speaking-of-
science/wp/2017/12/04/how-will-humanity-react-to-alien-life-psychologists-have-some-
predictions/ 30
Hacker Hughes, J. (Eds.). (2017). Military veteran psychological health and social care:
Contemporary approaches. Routledge. 31
Haybron, D. M. (2008). Philosophy and the science of subjective well-being. In M. Eid & R. J.
Larsen (Eds.), The science of subjective well-being (pp. 17-43). Guilford Press. 32
Hollander, M. M. (2017). Resistance to authority: Methodological innovations and new lessons
from the Milgram experiment (Publication No. 10289373) [Doctoral dissertation,
University of Wisconsin-Madison]. ProQuest Dissertations and Theses Global. 33
Hutcheson, V. H. (2012). Dealing with dual differences: Social coping strategies of gifted and
lesbian, gay, bisexual, transgender, and queer adolescents [Master’s thesis, The College
28 Newspaper article without DOI, from most academic research databases or print version
29 Entry in a dictionary, thesaurus, or encyclopedia, with individual author.
30 Online newspaper article.
31 Edited book without a DOI, from most academic research databases or print version.
32 Book chapter, print version.
33 Doctoral dissertation, from an institutional database.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 35
of William & Mary]. William & Mary Digital Archive.
https://scholarworks.wm.edu/etd/1539272210/ 34
Kalnay, E., Kanimitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S.,
White, G., Whollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J.,
Mo, K. C., Ropelewski, C., Wang, J., Leetmaa, A., … Joseph, D. (1996). The
NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological
Society, 77(3), 437-471. http://doi.org/ fg6rf9 35
King James Bible. (2017). King James Bible Online. https://www.kingjamesbibleonline.org/
(Original work published 1769) 36
Lewin, K. (1999). Group decision and social change. In M. Gold (Ed.), The complex social
scientist: A Kurt Lewin reader (pp. 265-284). American Psychological Association.
https://doi.org/10.1037/10319-010 (Original work published 1948) 37
Liberty University. (2019). The online writing center. https://www.liberty.edu/online/casas/
writing-center/ 38
Liberty University. (2020). BIOL 102: Human biology. Week one, lecture two: Name of class
lecture. https://learn.liberty.edu 39
34 Thesis or dissertation, from the web (not in a database).
35 Resource with 21 or more authors. Note the ellipse (…) in place of the ampersand (&).
36 Religious work.
37 Work in an anthology.
38 Online webpage with URL.
39 Liberty University class lecture with no presenter named.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 36
McCauley, S. M., & Christiansen, M. H. (2019). Language learning as language use: A cross-
linguistic model of child language development. Psychological Review, 126(1), 1-51.
https://doi.org/10.1037/rev0000126 40
McCurry, S. (1985). Afghan girl [Photograph]. National Geographic.
https://www.nationalgeographic.com/magazine/national-geographic-magazine-50-years-
of-covers/#/ngm-1985-jun-714.jpg 41
Merriam-Webster. (n.d.). Heuristic. In Merriam-Webster.com dictionary. Retrieved 01/02/2020,
from http://www.m-w.com/dictionary/heuristic 42
National Cancer Institute. (2018). Facing forward: Life after cancer treatment (NIH Publication
No. 18-2424). U.S. Department of Health and Human Services, National Institutes of
Health. https://www.cancer.gov/publications/patient-education/life-after-treatment.pdf 43
National Nurses United. (n.d.). What employers should do to protect nurses from Zika.
https://www.nationalnursesunited.org/pages/what-employers-should-do-to-protect-rns-
from-zika 44
O’Keefe, E. (n.d.). Egoism & the crisis in Western values. http:// www. onlineoriginals.com/
showitem.asp?itemID-135 45
40 Typical journal article with doi.
41 Photograph.
42 Dictionary entry.
43 Report by a government agency or other organization.
44 Webpage on a website with no date.
45 Electronic only book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 37
Owen, C. (2012, Spring). Behavioral issues resulting from attachment have spiritual
implications [Unpublished manuscript]. COUN502, Liberty University. 46
Perigogn, A. U., & Brazel, P. L. (2012). Captain of the ship. In J. L. Auger (Ed.) Wake up in the
dark (pp. 108-121). Shawshank Publications. 47
Peters, C. (2012). COUN 506: Integration of spirituality and counseling. Week one, lecture two:
Defining integration: Key concepts. Liberty University.
https://podcasts.apple.com/us/podcast/defining-integration-key-
concepts/id427907777?i=1000092371727 48
Pew Research Center. (2018). American trend panel Wave 26 [Data set].
https://www.pewsocialtrends.org/dataset/american-trends-panel-wave-26 49
Prayer. (2015). http:// www exact-webpage 50
Project Implicit. (n.d.). Gender–Science IAT. https://implicit.harvard.edu/implici/taketest.html 51
Schatz, B. R. (2000, November 17). Learning by text or context? [Review of the book The social
life of information, by J. S. Brown & P. Duguid]. Science, 290, 1304.
https://doi.org/10.1126/science.290.5495.1304 52
Schmid, H.-J. (Ed.). (2017). Entrenchment and the psychology of language learning: How we
reorganize ad adapt linguistic knowledge. American Psychological Association; De
46 Citing a student’s paper submitted in a prior class, in order to avoid self-plagiarism.
47 Chapter from an edited book.
48 Liberty University class lecture using course details.
49 Data set.
50 Online resource with no named author. Title of webpage is in the author’s place.
51 Test, scale, or inventory itself.
52 Review of a book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 38
Gruyter Mouton. https://doi.org/10.1037/15969-000 53
Second, M. P. (2011). Same author arranged by date (earlier first). Journal Name, 8, 12-13.
Second, M. P. (2015). Remember that earlier date goes first. Journal Name, 11(1), 18. 54
Shakespeare, W. (1995). Much ado about nothing (B. A. Mowat & P. Werstine, Eds.).
Washington Square Press. (Original work published 1623) 55
Strong, E. K., Jr., & Uhrbrock, R. S. (1923). Bibliography on job analysis. In L. Outhwaite
(Series Ed.), Personnel research series: Vol. 1, Job analysis and the curriculum (pp. 140-
146). https://doi.org/10.1037/10762-000 56
Tellegen, A., & Ben-Porah, Y. S. (2011). Minnesota Multiphasic Personality Inventory–2
Restructured Form (MPI-2-RF): Technical manual. Pearson. 57
U.S. Census Bureau. (n.d.). U.S. and world population clock. U.S. Department of Commerce.
Retrieved July 3, 2019, from https://www.census.gov/popclock 58
U.S. Securities and Exchange Commission. (2017). Agency financial report: Fiscal year 2017.
https://www.sec.gov/files/sec-2017-agency-financial-report.pdf 59
University of Oxford. (2018, December 6). How do geckos walk on water? [Video]. YouTube.
https://www.youtube.com/watch?v=qm1xGfOZJc8 60
53 Edited book with a DOI, with multiple publishers.
54 Two resources by the same author, in different years. Arrange by the earlier year first.
55 Shakespeare.
56 Electronic version of book chapter in a volume in a series
57 Manual for a test, scale, or inventory.
58 Webpage on a website with a retrieval date.
59 Annual report.
60 YouTube or other streaming video.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 39
Vedentam, S. (Host). (2015-present). Hidden brain [Audio podcast]. NPR. https://www.npr.org/
series/423302056/hidden-brain 61
Weinstock, R., Leong, G. B., & Silva, J. A. (2003). Defining forensic psychiatry: Roles and
responsibilities. In R. Rosner (Ed.), Principles and practice of forensic psychiatry (2nd
ed., pp. 7-13). CRC Press. 62
Yoo, J., Miyamoto, Y., Rigotti, A., & Ryff, C. (2016). Linking positive affect to blood lipids: A
cultural perspective [Unpublished manuscript]. Department of Psychology, University of
Wisconsin-Madison. 63
61 Podcast.
62 Chapter in an edited book without a DOI, from most academic research databases or print
version.
63 Unpublished manuscript with a university cited.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 40
Appendix
Annotated Bibliography
Cross, D. & Purvis, K. (2008). Is maternal deprivation the root of all evil? Avances en
Psycologia Latinoamericana, 26(1), 66-81.
Weaving spiritual applications throughout the article, the authors incorporate a plethora
of references to substantiate that maltreatment has a direct connection to attachment
disorders. They provide articulate and heavily-supported reasoning, detailing the specific
causes of maternal deprivation individually and then incorporating them in a broader
sense to answer the article’s title in the affirmative.
Feldman, R. (2007), Mother-infant synchrony and the development of moral orientation in
childhood and adolescence: Direct and indirect mechanisms of developmental continuity.
American Journal of Orthopsychiatry, 77(4), 582-597.
This longitudinal study tracked 31 Israeli children from ages 3 months to 13 years
(infancy to adolescence). There were direct parallels noted between increased
attachment/coherence and the child’s moral cognition, empathy development, and verbal
IQ. Toddlers who were able to regulate their own behavior later proved to excel in lead-
lag structures and language skills.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 1
Created by Christy Owen of Liberty University’s Online Writing Center
onlinewriting@liberty.edu; last date modified: February 7, 2022
Sample APA Paper: Professional Format for Graduate/Doctoral Students
Claudia S. Sample
School of Behavioral Sciences, Liberty University
Author Note
Claudia S. Sample (usually only included if author has an ORCID number)
I have no known conflict of interest to disclose.
Correspondence concerning this article should be addressed to Claudia S. Sample.
Email: cssample123456789@liberty.edu
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 2
Table of Contents
(Only Included for Easy Navigation; Hyperlinked for Quick Access)
Sample APA Paper: Professional Format for Graduate/Doctoral Students ……………………………… 6
Basic Rules of Scholarly Writing ……………………………………………………………………………………… 7
Brief Summary of Changes in APA-7 ………………………………………………………………………………… 8
Running Head, Author Note, and Abstract …………………………………………………………………………. 9
Basic Formatting Elements …………………………………………………………………………………………….. 10
Font ………………………………………………………………………………………………………………….. 10
Line Spacing ……………………………………………………………………………………………………… 10
Spaces After Punctuation …………………………………………………………………………………….. 11
Footnotes …………………………………………………………………………………………………………… 11
Heading Levels—Level 1 ………………………………………………………………………………………………. 11
Level 2 Heading …………………………………………………………………………………………………. 12
Level 3 Heading ………………………………………………………………………………………. 13
Level 4 Heading. Must be bolded and indented ½”. Add a period, one
space, and begin your content on the same line as shown here. ………………………………… 13
Level 5 Heading …………………………………………………………………. 13
Specific Elements of Academic Papers ……………………………………………………………………………. 13
Tables of Contents and Outlines …………………………………………………………………………… 13
Annotated Bibliographies ……………………………………………………………………………………. 14
Appendices ………………………………………………………………………………………………………… 14
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 3
Crediting Your Sources………………………………………………………………………………………………….. 15
Paraphrasing and Direct Quotes ……………………………………………………………………………. 15
Paraphrasing ……………………………………………………………………………………………. 16
Block Quotes …………………………………………………………………………………………… 16
How Often to Cite Your Source in Each Paragraph ………………………………………………… 17
Rule for Omitting the Year of Publication ……………………………………………………………… 17
Arranging the Order of Resources in Your Citations ………………………………………………. 17
Two Works by the Same Author in the Same Year …………………………………………………. 18
Two Works by Two Different Authors with the Same Last Name ……………………………. 18
Three or More Authors Cited In-Text ……………………………………………………………………. 18
Number of Authors in the Reference List ………………………………………………………………. 19
Numbers ………………………………………………………………………………………………………………………. 19
Displaying Titles of Works in-Text …………………………………………………………………………………. 19
Primary Sources versus Secondary Sources ……………………………………………………………………… 20
Personal Communications ……………………………………………………………………………………………… 20
Resources Canonically Numbered Sections (i.e., the Bible and Plays) …………………………………. 21
Bible and other Classical Works …………………………………………………………………………… 21
Plays …………………………………………………………………………………………………………………. 22
Lectures and PowerPoints ………………………………………………………………………………………………. 22
Dictionary Entries …………………………………………………………………………………………………………. 23
Changes in Reference Entries …………………………………………………………………………………………. 23
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 4
Electronic Sources ………………………………………………………………………………………………………… 24
Adding Color ……………………………………………………………………………………………………………….. 24
Self-Plagiarism ……………………………………………………………………………………………………………… 25
Final Formatting Tweaks ……………………………………………………………………………………………….. 26
Exhaustive Reference List Examples & Additional Helpful Resources ………………………………… 26
Conclusion …………………………………………………………………………………………………………………… 29
References ……………………………………………………………………………………………………………………. 30
Appendix ……………………………………………………………………………………………………………………… 40
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 5
Abstract
Begin your abstract at the left margin. This is the only paragraph that should not be indented.
Unless otherwise instructed, APA recommends an abstract be no more than 250 words. It should
generally not contain any citations or direct quotes. This should be a tight, concise summary of
the main points in your paper, not a step-by-step of what you plan to accomplish in your paper.
Avoid phrases such as “this paper will,” and just structure your sentences to say what you want
to say. The following three sentences exemplify a good abstract style: There are many
similarities and differences between the codes of ethics for the ACA and the AACC. Both include
similar mandates in the areas of —-, —, and —. However, each differs significantly in the areas
of —, —, and —. For more detailed information, see “Writing an Abstract” at
https://www.liberty.edu/casas/academic-success-center/wp-content/uploads/sites/28/2019/04/
Writing_an_Abstract_Revised_2012.pdf (note that you would not include any links in your
abstract). This is just now at 168 words, so eyeball how brief your abstract must be. Think of
your paper as a movie you want to sound enticing, and the abstract as the summary of the plot
you would share to draw people’s interest into wanting to come and see your movie. You want to
really hook and intrigue them. What you have to say is important! Remember to stay under 250,
words. Keywords highlight the search terms someone would use to find your paper in a database.
Keywords: main words, primary, necessary, search terms
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 6
Sample APA Paper: Professional Format for Graduate/Doctoral Students
The title of your paper goes on the top line of the first page of the body (American
Psychological Association [APA], 2019, section 2.11). It should be centered, bolded, and in title
case (all major words—usually those with four+ letters—should begin with a capital letter)—see
p. 51 of your Publication Manual of the American Psychological Association: Seventh Edition
(APA, 2019; hereinafter APA-7). It must match the title that is on your title page (see last line on
p. 32). As shown in the previous sentence, use brackets to denote an abbreviation within
parentheses (bottom of p. 159). Write out the full name of an entity or term the first time
mentioned before using its acronym (see citation in first sentence in this paragraph), and then use
the acronym throughout the body of the paper (section 6.25).
There are many changes in APA-7. One to mention here is that APA-7 allows writers to
include subheadings within the introductory section (APA, 2019, p. 47). Since APA-7 now
regards the title, abstract, and term “References” to all be Level-1 headings, a writer who opts to
include headings in his or her introduction must begin with Level-2 headings as shown above
(see section 2.27) for any divisions within the introductory section.
If you do choose to include headings in your introduction section (which is optional), be
sure to include two or more subheadings, since APA (2019) forbids stand-alone heading levels.
A second notable change in APA-7 is that writers are no longer required to cite their source every
single sentence that content from it is mentioned (section 8.1). As demonstrated in this paper,
since all of the content (other than the examples included for illustration and reference-entry
variation purposes) comes directly from the APA-7 itself, citations to the APA-7 are only
included for the first instance in each paragraph. Section and/or page numbers are included
parenthetically throughout for the sake of students who desire to know exactly where the stated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 7
rule appears in the APA-7 itself. In your academic papers, however, it is critical to include the
required author(s) and year, as applicable, for all citations that are included; this may include
more than one citation for each resource per paragraph, as required to avoid any confusion about
the source of that content.
Basic Rules of Scholarly Writing
Most beginning students have difficulty learning how to write papers and also format
papers correctly using the seventh edition of the APA manual. However, the Liberty University
Online Writing Center’s (OWC) mission includes helping students learn how to be autonomous,
proficient writers. The OWC also provides students with templates to help them with basic
formatting elements, but this sample paper is designed to help graduate and doctoral students
learn to master APA rules and formatting on their own, which will prove helpful as they progress
in their studies and work toward future publication in scholarly journals.
For the purpose of instruction, this paper will use second person (you, your), but third
person (this author) must be used in most student papers. First person (I, me, we, us, our) is not
generally permitted in academic papers. Students should refrain from using first or second person
in college courses (even though the APA manual encourages this in other writing venues) unless
the assignment instructions clearly permit such (as in the case of personal reflection sections or
life histories). If in doubt, students should clarify with their professors.
APA-7 delineates separate rules and guidelines between “student” and “professional”
writers (APA, 2019). Because a primary purpose of graduate and doctoral studies is to prepare
those students to publish professionally, Liberty University has decided to have undergraduate
students follow APA-7’s guidelines for “student papers,” and graduate/doctoral students follow
APA-7’s guidelines for “professional papers.” Separate templates are available for each level.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 8
This sample paper illustrates and discusses the rules and formatting of professional papers, as
required for all Liberty University graduate and doctoral courses using APA-7 style.
Brief Summary of Changes in APA-7
Most of these changes will be discussed in more detail below; this is just a very brief
overview here. APA-7 reverts back to only one space after closing punctuation in the body of the
paper (APA-6 required two spaces; APA, 2019, section 6.1). Student (undergraduate) papers no
longer include a running head or abstract (sections 2.2 and 2.8); professional (graduate/doctoral)
papers require an abstract but the running head is now the same on all pages (the added phrase
“Running head:” from APA-6 has been eliminated; see section 2.8). Title pages are different for
both student and professional formats. The title of a paper is no longer limited to 12 words
(section 2.4).
Citations of all resources with three or more authors now use the first author’s last name
and the term et al. (APA, 2019, section 8.17). Reference entries must name up to the first 19
authors before adding an ampersand and ellipsis (up from APA-6’s six authors; section 9.8).
APA-7 omits the phrase DOI and instead standardizes DOIs to be presented in hyperlink format
(i.e., https://doi.org/10.1177%2F1524838017742386; section 9.35). Formatting guidelines for
annotated bibliographies are included in APA-7 (section 9.51), as well as expanded and
standardized reference entry examples. As discussed above, it is no longer necessary to cite a
source every single time you refer to content gleaned from it as long as it is clear the content
comes from that source (section 8.1); APA-7 also expanded the specific location noted in the
citation to include page, paragraph, section (as used throughout this sample paper, to direct the
student to the exact relevant content), chapter, timestamp, etc. (section 8.13).
APA-7 allows for “self-plagiarism” (clarified and defined below). It also invites writers to
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 9
highlight the most relevant work first, rather than just present all works in alphabetical order
(APA, 2019, section 8.12).
Heading-level formatting has changed, and APA-7 provides more flexibility in font and
line spacing (APA, 2019). The Bible must now be included in the reference list and its citations
must include the editor’s details and year (section 8.28); there are also new rules for dictionary
entries. Publisher city and state details are omitted from all reference entries except those
involving presentations or conferences, as is the phrase “retrieved from.” Hyperlinks should be
live, but they may be either presented as blue underlining or plain black text.
Running Head, Author Note, and Abstract
APA (2019) delineates separate formatting requirements for what it terms “student” and
“professional” papers. Its descriptions for those labels, however, suggests that it regards
undergraduate-level writing to fall within the student purview, and graduate/doctoral-level
writing (including dissertations and theses) to fall within the professional purview. Since a
significant goal in graduate and post-graduate studies is preparing those students to publish in
scholarly journals at and beyond graduation, it makes sense to train those students in the
formatting that is required for professionals. As such, Liberty University has opted to require the
APA-7’s “student” version format for all undergraduate assignments using APA, and its
“professional” version for all graduate and doctoral assignments. To that end, this being the
sample paper for professional formatting, it includes the additional elements required for such: a
running head (same on all pages), an author’s note, and an abstract. Graduate and doctoral
students will use this format. Note that the first “paragraph” under the author’s note is generally
only included if the author has an ORCID number, which most students will not have. However,
it is included in this sample paper and the corresponding template because the purpose of these
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 10
resources is to prepare students to publish manuscripts post-graduation. The student’s full
address, however, is intentionally omitted from the Liberty University template and this sample
paper for privacy and safety reasons, since student papers are often unfortunately published
online and disclosing their home addresses could pose safety risks.
Basic Formatting Elements
Font
APA-7 does not prescribe a specific font or size (APA, 2019, section 2.19) but rather
allows for some choice (e.g., 12-point Times New Romans, 11-point Calibri, 11-point Arial, 11-
point Georgia, or 10-point Lucinda Sans Unicode). Most journals and academic institutions will
have a preference, however, as even APA-7 acknowledges on p. 44. For this reason—and
because font size can easily be changed if an editor interested in publishing a student’s work
prefers a different font—Liberty University recommends that students use 12-point Times New
Romans or 11-point Calibri font for the body text in all academic papers. Data in charts, figures,
and tables should be presented in 8- to 14-point size in either Calibri, Arial, or Lucinda Sans
Unicode font. Students are not permitted to use any fonts such as script, calligraphy, poster,
decorative, or others not found in published scholarly journals. Since APA-7 itself authorizes a
variety of fonts and sizes, assignments will be gauged by word count rather than page count.
Word count constitutes the number of words within the body of the paper, and excludes the title
page, abstract, reference list, appendices, and other supplemental resources.
Line Spacing
APA-7 adds extra/blank lines on the title page (APA, 2019, sections 2.5, 2.7, 2.21). It also
specifies that spacing in tables and figures may be single-, 1-1/2-, or double-spaced; equations
can be triple- or quadruple-spaced. Footnotes, when used at the bottom of a page, should be
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 11
single-spaced (section 2.21).
Spaces After Punctuation
APA-7 reverts back to just one space after closing punctuation in the body of the paper, as
well as in reference entries (APA, 2019, section 6.1). Ordinarily, it would be improper to have a
paragraph with only one sentence, though APA itself asserts that for its purposes “sentences and
paragraphs of any length are technically allowed.”1
Footnotes
This leads to another new rule in APA-7, one allowing the inclusion of footnotes (APA,
2019, section 2.13). Footnotes should be use very sparingly and are appropriate to include
information such as that in the prior section to alert the reader to supplemental material that is
available online for that thought. Though APA-7 authorizes placement of footnote content either
at the bottom of the page (as in this sample paper) or on a separate page after the reference list
(section 2.21), Liberty University recommends that student place them, when used, at the bottom
of the page, as shown here.
Heading Levels—Level 1
This sample paper uses primarily two levels of headings (Levels 1 and 2). APA style,
however, has five heading levels, which will be demonstrated briefly for visual purposes. See
section 2.27 of your APA-7 (APA, 2019) for more details on heading levels and formatting. In
APA-7, all heading levels are now bolded and in title case (capitalize each major word—usually
those with four or more letters, including hyphenated compound words). Do not capitalize
articles (a, an, the) in headings unless they begin a title or follow a colon. Level 1 headings are
centered, with the content falling on the line beneath each, in standard paragraph format.
1 See https://blog.apastyle.org/apastyle/2016/05/index.html
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 12
Many students misunderstand that you progress from Level 1 to Level 2 to Level 3 to
Level 4 to Level 5, but that is not correct. In fact, your paper may have only Level 1 headings, or
just Levels 1 and 2. The rule of thumb is that you must have at least two of each heading level
that you use, otherwise omit that heading level.
Headings are basically styling ways of organizing your paper, without using an outline
format. APA specifies five levels of headings; you would likely never use Level 5 and only very
rarely use Level 4 as a student. Think of each level as the different levels in an outline. Roman
numerals, for example, would be Level 1 headings. Capital letters would be Level 2 headings.
Numerals would be Level 3 headings. Lowercase letters would be Level 4. And lowercase
Roman numerals would be Level 5. You must always have two or more of each subheading, but
you do not need every level. You start with Level 1 and work down from that (but not
consecutive 1-2-3-4-5). Under a Level 1, you would either have two+ Level 2 headings or none
at all (just one big section in paragraphs before the next Level 1 section).
Special note about conclusion sections: Please note that some of the sample papers
published by APA to demonstrate proper APA-7 format (including the “professional” sample on
pp. 50-60 of the APA-7 manual) depict the “Conclusion” section with a Level-2 heading. This is
limited to empirical papers that are being submitted for publication in scholarly journals, as those
conclusions pertain to the “Discussion” sections in such papers and are not conclusions of the
overall papers themselves. Conclusions in academic papers at Liberty University will be Level 1
headings (including dissertations and theses, which are divided by chapters, unlike journal article
manuscripts).
Level 2 Heading
Level 2 headings are left-justified (APA, 2019, p. 48). The supporting information is
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 13
posed in standard paragraph form beneath it. Never use only one of any level of heading. You
must use two or more of any level you use, though not every paper will require more than one
level. The heading levels are simply demonstrated here for visual purposes, but you would
always have two or more of each under a larger heading, as shown throughout all the other
sections of this sample paper.
Level 3 Heading
Level 3 headings are bolded, left-justified, and italicized; the content falls on the line
underneath, as with Levels 1 and 2.
Level 4 Heading. Must be bolded and indented ½”. Add a period, one space, and begin
your content on the same line as shown here.
Level 5 Heading. Same as Level 4, but also italicized. Despite heavy writing experience,
this author has never used Level 5 headings.
Specific Elements of Academic Papers
Tables of Contents and Outlines
APA (2019) does not regulate every type of paper and some elements in various
assignments are not addressed in the APA-7 manual, including outlines and tables of content. In
those cases, follow your professor’s instructions and the grading rubric for the content and
format of the outline or annotations, and use standard APA formatting for all other elements
(such as running head, title page, body, reference list, 1″ margins, double-spacing, permitted
font, etc.). Note that most academic papers will not require a table of contents, nor would one be
appropriate. One was included in this paper simply for ease-of-access so students could go
directly to the content they want to see. Generally speaking, no table of contents would be
necessary for papers less than 20 pages of content, unless otherwise required by your professor.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 14
That being said, when organizing outlines in APA format, set your headings up in the
proper levels (making sure there are at least two subheadings under each level), and then use
those to make the entries in the outline. As discussed above, Level 1 headings become uppercase
Roman numerals (I, II, III), Level 2 headings become capital letters (A, B, C), Level 3 headings
become numbers (1, 2, 3), Level 4 headings become lowercase letters (a, b, c), and Level 5
headings become lowercase Roman numerals (i, ii, iii). Many courses now require “working
outlines,” which are designed to have the bones and foundational framework of the paper in
place (such as title page, abstract, body with title, outline/heading divisions, supporting content
with citations, and references), without the full “meat” that fills out and forms a completed paper.
Annotated Bibliographies
Many Liberty University courses also now require students to prepare and submit an
annotated bibliography as a foundational step to building a research paper. There is significant
merit in these assignments, as they teach students to critique the resources they have found and
rationalize why each is relevant for their paper’s focus. APA (2019) includes a section on
annotated bibliographies (9.51; see the example provided on p. 308). The appendix attached to
this sample paper also includes a sample annotated bibliography.
Appendices
Appendices, if any, are attached after the reference list (APA, 2019, section 2.14). You
must refer to them (i.e., “callout”) in the body of your paper so that your reader knows to look
there (see the yellow-highlighted callouts to Table 1 on p. 54 and to Footnote 1 on p. 55 of your
APA-7 for visuals on how this should appear in your paper). The word “Appendix” is singular;
use it to refer to individual appendices. APA-7 regards it as a Level 1 heading so it should be
bolded. I attached a sample Annotated Bibliography as a visual aid (see Appendix). You will see
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 15
that I included the title “Appendix” at the top of the page and formatted it in standard APA
format beneath that. Because I only included one appendix, it is simply titled as such. If there are
more appendices, assign a letter to each and denote each by that: “Appendix A” and “Appendix
B.”
Crediting Your Sources
Paraphrasing and Direct Quotes
Paraphrasing is rephrasing another’s idea in one’s own words by changing the wording
sufficiently without altering the meaning (remember not to just change a word here or there or
rearrange the order of the original source’s wording). Quoting is using another’s exact words.
Both need to be cited; failure to do so constitutes plagiarism. Include the author(s) and year for
paraphrases, and the author(s), year, and page or paragraph number for direct quotes. APA-7 also
expands this to include figure number, time stamp, etc.—whatever detail is necessary to get the
reader directly to that content. Page numbers should be used for any printed material (books,
articles, etc.), and paragraph numbers should be used in the absence of page numbers (online
articles, webpages, etc.; see APA, 2019, section 8.13). Use p. for one page and pp. (not italicized
in your paper) for more than one (section 8.25). Use para. for one paragraph and paras. (also not
italicized in your paper) for two or more (section 8.28). For example: (Perigogn & Brazel, 2012,
pp. 12–13) or (Liberty University, 2019, para. 8). Section 8.23 of the APA (2019) manual
specifies that it is not necessary to include a page or paragraph number for paraphrases (just for
direct quotes), but writers may choose to do so to help their readers find that content in the cited
resource.
When naming authors in the text of the sentence itself (called a narrative citation), use the
word “and” to connect them. For example, Perigogn and Brazel (2012) contemplated that . . .
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 16
Use an ampersand (&) in place of the word “and” in parenthetical citations and reference lists:
(Perigogn & Brazel, 2012).
Paraphrasing
Only use quotes when the original text cannot be said as well in your own words or
changing the original wording would change the author’s meaning. You cannot simply change
one word and omit a second; if you paraphrase, the wording must be substantially different, but
with the same meaning. Regardless, you would need to cite the resource you took that
information from. For example, Benoit et al. (2010) wrote that “although, a link between
attachment and posttraumatic stress disorder (PTSD) symptoms has been established, the
mechanisms involved in this link have not yet been identified” (p. 101). A paraphrase for that
quote might be: A link between dysfunctional attachment and the development of PTSD has
been made, though there is insufficient data to determine exactly how this mechanism works
(Benoit et al., 2010).
Block Quotes
Quotes that are 40 or more words must be blocked, with the left margin of the entire
quote indented ½ inch. Maintain double-spacing of block quotes. APA prefers that you introduce
quotes but note that the punctuation falls at the end of the direct quote, with the page number
outside of that (which is contrary to punctuation for non-blocked quotes). For example, Alone
(2008) claims:2
Half of a peanut butter sandwich contains as much bacteria as the wisp of the planet
Mars. Thus, practicality requires that Mrs. Spotiker nibble one bit at a time until she is
assured that she will not perish from ingesting it too quickly. (p. 13)
2 Note that there are no quotation marks for block quotes, as shown in the example.
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 17
Usually quotes within quotes use single quotation marks; however, use double quotation marks
for quotes within blocked quotes, since there are no other quotation marks involved. Also
understand that direct quotes should be used sparingly in scholarly writing; paraphrasing is much
preferred in APA format (APA, 2019, section 8.23), as it demonstrates that you read, understood,
and assimilated other writers’ content into one cohesive whole.
How Often to Cite Your Source in Each Paragraph
As already mentioned above, APA’s (2019) new official rule is that you no longer must
cite your source every single time you refer to material you gleaned from it (section 8.1). It is
now acceptable to cite your source the first time you refer to content from it in your paragraph,
and then not again in that same paragraph unless your phrasing does not make the source of your
content clear. This is demonstrated throughout this sample paper.
Rule for Omitting the Year of Publication
That being said, APA (2019) has clarified its special rule that excludes the year of
publication in subsequent narrative in-text citations (when you name the authors in the text of the
sentence itself), after the first narrative citation in each paragraph. It should continue to appear in
all parenthetical citations (see section 8.16). For example, Alone (2008) portrays imagery of Mrs.
Spotiker. This includes her devouring a peanut butter sandwich (Alone, 2008). Alone conveys
this through the lens of astronomy. Note that the year of publication was omitted from the second
narrative citation (underlined for visual purposes).
Arranging the Order of Resources in Your Citations
If the material you cited was referred to in multiple resources, separate different sets of
authors with semicolons, arranged in the order they appear (alphabetically by the first author’s
last name) in the reference list (i.e., Carlisle, n.d.-a; Prayer, 2015) (APA, 2019, section 8.12).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 18
APA-7 now invites writers to prioritize or highlight one or more sources as most prominent or
relevant for that content by placing “those citations first within parentheses in alphabetical order
and then insert[ing] a semicolon and a phrase, such as ‘see also,’ before the first of the remaining
citations” (APA., 2019, p. 263)—i.e., (Cable, 2013; see also Avramova, 2019; De Vries et al.,
2013; Fried & Polyakova, 2018). Periods are placed after the closing parenthesis, except with
indented (blocked) quotes.
Two Works by the Same Author in the Same Year
Authors with more than one work published in the same year are distinguished by lower-
case letters after the years, beginning with a (APA, 2019, section 8.19). For example, Double
(2008a) and Double (2008b) would refer to resources by the same author published in 2008.
When a resource has no date, use the term n.d. followed by a dash and the lowercase letter (i.e.,
Carlisle, n.d.-a and Carlisle, n.d.-b; see APA, 2019, section 8.19).
Two Works by Two Different Authors with the Same Last Name
Citations in the body of the paper should include only the last names, unless you have
two or more resources authored by individuals with the same last name in the same year (or are
citing a personal communication). When there are two different authors with the same last name
but different first names who published in the same year, include the first initials: Brown, J.
(2009) and Brown, M. (2009) (APA, 2019, section 8.20).
Three or More Authors Cited In-Text
When referring to material that comes from three or more authors, APA-7 now requires
that all citations name just the first author’s last name followed by the words et al. (without
italics) (APA, 2019, section 8.17). Et al. is a Latin abbreviation for et alii, meaning “and others,”
which is why the word “al.” has a period, whereas “et” does not. Alone et al. (2011) stipulated
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 19
that peacocks strut. Every single time I refer to their material, I would apply APA-7’s rule: Alone
et al. (2011) or (Alone et al., 2011). Since et al. denotes plural authors, the verb must be plural to
match, too: Alone et al. (2011) are… This applies to all citations within the body of the paper
with three or more authors.
Number of Authors in the Reference List
For resources with 20 or fewer authors in the reference list, write out all of the authors’
last names with first and middle initials, up to and including the 20th author (APA, 2019, section
9.8). APA-7 has a special rule for resources with 21 or more authors: Write out the first 19
authors’ last names with initials, insert an ellipsis (…) in place of the ampersand (&), and finish
it with the last name and initials of the last author. See example #4 provided on page 317 of your
APA-7, as well as this paper’s reference list for visuals of these variances (Acborne et al. 2011;
Kalnay et al., 1996).
Numbers
Numbers one through nine must be written out in word format (APA, 2019, section 6.33),
with some exceptions (such as ages—see section 6.32). Numbers 10 and up must be written out
in numerical format (section 6.32). Always write out in word format any number that begins a
sentence (section 6.33).
Displaying Titles of Works in-Text
The names of journals, books, plays, and other long works, if mentioned in the body of
the paper, are italicized in title case (APA, 2019, section 6.17). Titles of articles, lectures, poems,
chapters, website articles, and songs should be in title case, encapsulated by quotation marks
(section 6.7). The year of publication should follow the author’s name, whether in narrative or
parenthetical format: Perigogn and Brazel (2012) anticipated…, or (Perigogn & Brazel, 2012).
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 20
The page or paragraph number must follow after the direct quote. Second (2015) asserted that
“paper planes can fly to the moon” (p. 13). You can restate that with a parenthetical citation as:
“Paper planes can fly to the moon” (Second, 2015, p. 13). Second (2011) is another resource by
the same author in a different year.
Primary Sources versus Secondary Sources
APA (2019) strongly advocates against using secondary sources; rather, it favors you
finding and citing the original (primary) resource whenever possible (section 8.6). On the rare
occasion that you do find it necessary to cite from a secondary source, both the primary (who
said it) and secondary (where the quote or idea was mentioned) sources should be included in the
in-text citation information. If the year of publication is known for both resources, include both
years in the citation (section 8.6). Only the secondary source should be listed in the reference
section, however. Use “as cited in” (without the quotation marks) to indicate the secondary
source. For example, James Morgan hinted that “goat milk makes the best ice cream” (as cited in
Alone, 2008, p. 117). Morgan is the primary source (he said it) and Alone is the secondary
source (he quoted what Morgan said). Only the secondary source is listed in the reference section
(Alone, and not Morgan) because if readers want to confirm the quote, they know to go to page
117 of Alone’s book.
Personal Communications
APA (2019) rationalizes the exclusion of references for information obtained through
personal communication (such as an interview, email, telephone call, postcard, text message, or
letter) in the reference list because your readers will not be able to go directly to those sources
and verify the legitimacy of the material. Instead, these items are cited only in the body of the
paper. You must include the individual’s first initial, his or her last name, the phrase “personal
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 21
communication” (without the quotation marks), and the full date of such communication (section
8.9). As with other citations, such citations may be either narrative or parenthetical. For example,
L. Applebaum advised him to dip pretzel rolls in cheese fondue (personal communication, July
13, 2015). The alternative is that he was advised to dip pretzel rolls in cheese fondue (L.
Applebaum, personal communication, July 13, 2015). Note that there is no entry for Applebaum
in the reference list below.
Resources Canonically Numbered Sections (i.e., the Bible and Plays)
These resources should be cited in book format (APA, 2019, Section 9.42). The Bible and
other religious works are generally regarded as having no author; an annotated version would be
treated as having an editor. Include republished dates as necessary. The OWC will publish a list
of reference entries for various Bible versions on its APA Quick Guide webpage.
Bible and other Classical Works
Works such as the Bible, ancient Greek or Roman works, and other classical works like
Shakespeare must be cited in the body of the paper (APA, 2019, section 8.28). APA-7 now also
requires that they be included in the reference list, too (section 9.42), which is a significant
change from APA-6. Republished dates are included as well (see section 9.41). As such, you
would add a parenthetical phrase at the end of your reference entry with the original publication
details; note that there should be no punctuation following such parenthetical content at the end
of a reference entry (the reference entries depicting this in the reference list below are correctly
punctuated).
Citations for the Bible will include the Bible version’s name in the author’s position (as
an anonymous work), original and republished years, and then the book chapter/verse (spelled
out) in place of the page number (i.e., King James Bible, 1769/2017, Genesis 3:8)—see sections
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 22
8.28 and 9.42. Note that APA (2019) requires book titles to be italicized in every venue,
including citations and reference entries. Because Liberty University is a distinctly-Christian
institution and many of its courses require biblical integration, most if not all of its students will
cite the Bible in virtually every course. The examples provided on pp. 274 and 325 of APA-7 are:
(note the italics in each)
• Narrative citation: King James Bible (1769/2017)
• Parenthetical citation: (King James Bible, 1769/2017, Song of Solomon 8:6)
• Reference entry: King James Bible. (2017). King James Bible Online.
https://www.kingjamesbibleonline.org/ (Original work published 1769)
Plays
When citing plays, “cite the act, scene, and line(s), in a single string, separated by
periods. For example, ‘1.3.36-37’ refers to Act 1 Scene 3, Lines 36-37” (APA, 2019, section
8.28; see also example #37 on p. 325).
Lectures and PowerPoints
APA (2019) has expanded and standardized its rules for citations and reference entries in
an effort to best credit the original sources. It now includes rules for crediting content in course
or seminar handouts, lecture notes, and PowerPoint presentations (see #102 on p. 347). When
citing a PowerPoint presentation, include the slide number rather than the page number. For
purposes of Liberty University course presentations and lectures (which are not readily available
to the public), reference each as a video lecture with the URL (if available) for the presentation,
naming the presenter(s) in the author’s position. Include the course number, lecture title, and
enough details for others to identify it within that course, in a sort of book format, naming
Liberty University as publisher. Peters (2012) is an example of this in the reference list of this
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 23
paper. If the presenter for a Liberty University class lecture is not named, credit Liberty
University as the author; see Liberty University (2020) in the reference list below as an example.
Dictionary Entries
In keeping with its efforts to standardize reference entries, APA (2019) now requires
citation and referencing of word definitions from dictionaries to follow the same rules for
chapters in an edited book (see #47 and #48 on p. 328; section 8.13). As such, you will now
name either the individual, group, or corporate author of the dictionary in the author’s place (e.g.,
Merriam-Webster, n.d.). If you searched online, include the retrieval date and the URL to the
exact webpage. If you used a hard copy book, include the publisher details. The in-text citation
in the body of the paper would follow standard author/year format (e.g., Merriam-Webster, n.d.).
Changes in Reference Entries
There are a number of notable changes in APA-7 from past versions. For the most part,
these simplify and unify the formats to be more consistent across the different resource venues.
Some of these have already been discussed above (i.e., naming up to 19 authors’ names before
adding an ellipsis, and crediting authors and editors of classical works and dictionaries). Other
changes include italicizing names of webpages and website resources in the reference list (APA,
2019, section 6.22), as well as book titles even when named in the author’s position (such as
King James Bible). The city and state locations of publishers are no longer required; only include
those details “for works that are associated with a specific location, such as conference
presentations” (p. 297, section 9.31). Issue numbers are required for all journal articles that have
such, regardless of what page number each issue begins with (section 9.25). If two or more
publishers are listed on the copyright page, include all of them in the order listed, separated by
semicolons (section 9.29). Omit the word Author in the publisher’s place when it is the same as
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 24
the author (section 9.24).
Electronic Sources
Note that since the APA 6th edition was published in 2010, great strides have been made
in online and electronic resource accessibility, and APA’s position on electronic resources has
shifted to embrace this. More and more resources are available electronically through the
Internet. The advent of this increased availability has resulted in APA-7’s effort to standardize
the formatting of resources, which in turn simplifies them to some extent. All reference entries
follow the same basic details: Author(s), year of publication, name of resource, and location
details (i.e., either journal name/volume/issue/page numbers, or book publisher, or webpage).
APA (2019) requires inclusion of a Digital Object Identifier (DOI) in the references
whenever available (section 9.34); if not, then a webpage, if available. In keeping with its
unification of resources, APA-7 now standardizes all DOIs and URLs to be presented in
https://doi.org/10.1177%2F1524838017742386 format. The phrase “Retrieved from” is now
excluded except when the content may have changed (such as dictionary entries, Twitter profiles,
Facebook pages; see section 9.16). APA-7 requires all hyperlinks to be active (so your reader can
click on one to go directly to that webpage), but they may appear as either blue-underlined text
or simple black text (section 9.35). There should be no period after any URL. APA-7 no longer
requires authors to break long URLs with soft returns (hold down the Shift key and press the
Enter key) at forward slashes, periods, or underscores to avoid unsightly spacing gaps, but it may
be best to do so in academic papers.
Adding Color
Though APA (2019) authorizes writers to include the use of color in photographs and
figures (section 7.26), Liberty University discourages this in academic papers. It risks becoming
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 25
distracting for both students in their quest to be creative, and professors in their quest to focus on
academic content.
Self-Plagiarism
APA (2019) also invites writers to repurpose some of their work in future papers.
Specifically, APA-7 states that:
In specific circumstances, authors may wish to duplicate their previously used works
without quotation marks or citation …, feeling that extensive self-referencing is
undesirable or awkward and that rewording may lead to inaccuracies. When the
duplicated material is limited in scope, this approach is permissible. (p. 8.3)
APA-7 adds “Do not use quotation marks or block quotation formatting around your own
duplicated material” (p. 256).
Liberty University, however, has stringent rules against self-plagiarism, as do many
scholarly journals. Liberty University students receive grades for their class papers; those who
have received feedback and a grade from a prior professor on a prior paper have an advantage
over their classmates, both in having the benefit of that feedback/grade and in not having to write
a whole paper from scratch during the subsequent class. Student papers are also submitted to
SafeAssign to deter plagiarism. For these reasons, Liberty University expressly forbids students
using significant portions of a prior paper in a subsequent course (either a retake of the same
course or a new class altogether). It is conceivable that students who are building their
knowledge base in a subject matter—particularly at the graduate and post-graduate levels—
would reasonably justify incorporating brief excerpts from past papers into current ones. In such
case, Liberty University authorizes students to utilize APA-7’s disclosure (i.e., “I have previously
discussed”), along with a citation to the prior class paper and a reference entry (i.e., Owen, 2012;
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 26
Yoo et al., 2016). Such self-references and re-use of content from prior papers should be used
sparingly and disclosed fully in the current paper; that content should not constitute a significant
portion of any academic assignment, however.
Final Formatting Tweaks
The templates provided by Liberty University are already formatted with proper spacing,
margins, heading level structure, and hanging indents, as necessary. With the exceptions of the
title page, figures, and equations, papers in APA format should be double-spaced throughout,
with no extra spacing between lines. Academic papers at Liberty University should also be in
one of the accepted fonts throughout (recommended: Times New Romans, 12-point font).
Sometimes when you format your paper or cut-and-paste material into it, things get skewed. One
quick way to ensure that your paper appears correct in these regards is to do a final formatting
tweak after you have completed your paper. Hold down the “Ctrl” button and press the “A” key,
which selects and highlights all of the text in your paper. Then go to the Home tab in Microsoft
Word and make sure that whichever acceptable font/size you choose to use is selected in the Font
box. Next, click on the arrow at the bottom of the Paragraph tab. Set your spacing before and
after paragraphs to “0 pt” and click the “double” line spacing. The extra spacing required on the
title page is already programmed into the template and should not change even when you
complete these actions.
Exhaustive Reference List Examples & Additional Helpful Resources
The reference list at the end of this paper includes an example of a myriad of different
sources and how each is formatted in proper APA-7 format. One example of each of the primary
types of resources will be included in the reference list, as cited in the body of this paper.
Remember that, for purposes of this paper only, many of the sources cited in the body of the
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 27
paper were provided for illustrative purposes only and thus are fictional, so you will not be able
to locate them if you searched online. Nevertheless, in keeping with APA-7 style, all resources
cited in the body of the paper are included in the reference list and vice versa (except for personal
communications, per APA-7’s published exceptions). Be absolutely sure that every resource cited
in the body of your paper is also included in your reference list (and vice versa), excepting only
those resources with special rules, such as personal communications and primary sources you
could not access directly.
The reference list in this paper is fairly comprehensive and will include a book by one
author who also appears as one of many authors in another resource (Alone, 2008; Alone et al.,
2011); chapters in edited books (Balsam et al., 2019; Haybron, 2008; Perigogn & Brazel, 2012;
Weinstock et al., 2003); electronic version of book (Strong & Uhrbrock, 1923); electronic only
book (O’Keefe, n.d.); edited books with and without DOIs, with multiple publishers (Hacker
Hughes, 2017; Schmid, 2017); work in an anthology (Lewin, 1999); journal articles (Andrews,
2016; Carlisle, n.d.-a, n.d.-b; De Vries R. et al., 2013; McCauley & Christiansen, 2019);
newspaper article (Goldman, 2018; Guarino, 2017); online webpages (Liberty University, 2019;
Prayer, 2015); resource with corporate author as publisher (American Psychological Association,
2019); resources by two authors with the same last name but different first names in the same
year of publication (Brown, J., 2009; Brown, M., 2009); two resources by same author in the
same year (Double, 2008a, 2008b; Carlisle, n.d.-a, n.d.-b); two resources by the same author in
different years (Second, 2011, 2015); resource with 20 authors (maximum allowed by APA-7
before special rule applies) (Acborne et al., 2011); resource with 21 or more authors (Kalnay et
al., 1996); dictionary entries (American Psychological Association, n.d.; Graham, 2019;
Merriam-Webster, n.d.); Liberty University class lecture using course details (Peters, 2012);
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 28
PowerPoint slides or lecture notes, not including course details (Canan & Vasilev, 2019); citing a
student’s paper submitted in a prior class, in order to avoid self-plagiarism (Owen, 2012);
unpublished manuscript with a university cited (Yoo et al., 2016); code of ethics (American
Counseling Association, 2014); diagnostic manual (American Psychiatric Association, 2013);
religious and classical works, including the Bible (Aristotle, 350 BC/1994; King James Bible,
1769/2017; Shakespeare, 1623/1995); dissertation or thesis (Hollander, 2017; Hutcheson, 2012);
review of a book (Schatz, 2000); video (Forman, 1975); podcast (Vedentam, 2015); recorded
webinar (Goldberg, 2018); YouTube or other streaming video (University of Oxford, 2018); clip
art or stock image (GDJ, 2018); map (Cable, 2013); photograph (McCurry, 1985); data set (Pew
Research Center, 2018); measurement instrument (Friedlander et al., 2002); manual for a test,
scale, or inventory (Tellegen & Ben-Porah, 2011); test, scale, or inventory itself (Project
Implicit, n.d.); report by a government agency or other organization (National Cancer Institute,
2018); report by individual authors at a government agency or other organization (Fried &
Polyakova, 2018); annual report (U.S. Securities and Exchange Commission, 2017); conference
session (Fistek et al., 2017); and webpages (Avramova, 2019; Centers for Disease Control and
Prevention, 2018; National Nurses United, n.d.; U.S. Census Bureau, n.d.).
Lastly, below are a few webpages that address critical topics, such as how to avoid
plagiarism and how to write a research paper. Be sure to check out Liberty University’s Online
Writing Center (https://www.liberty.edu/online/casas/writing-center/) for more tips and tools, as
well as its Facebook page (https://www.facebook.com/groups/LUOWritingCenter). Remember
that these links are only provided for your easy access and reference throughout this sample
paper, but web links and URLs should never be included in the body of scholarly papers; just in
the reference list. Writing a research paper (https://www.youtube.com/watch?v=zaa-PTexW2E
SAMPLE APA-7 PAPER FOR GRADUATE/DOCTORAL STUDENTS 29
or https://www.youtube.com/watch?v=KNT6w8t3zDY and avoiding plagiarism
(https://www.youtube.com/watch?v=VeCrUINa6nU).
Conclusion
The conclusion to your paper should provide your readers with a concise summary of the
main points of your paper (though not via cut-and-pasted sentences used above). It is a very
important element, as it frames your whole ideology and gives your readers their last impression
of your thoughts. Be careful not to introduce new content in your conclusion.
After your conclusion, if you are not using the template provided by the Online Writing
Center, insert a page break at the end of the paper so that the reference list begins at the top of a
new page. Do this by holding down the “Ctrl” key and then clicking the “Enter” key. You will go
to an entirely new page in order to start the reference list. The word “References” (not in
quotation marks) should be centered and bolded. Items in the reference list are presented
alphabetically by the first author’s last name and are formatted with hanging indents (the
second+ lines of each entry are indented 1/2” from the left margin). APA authorizes the use of
singular “Reference” if you only have one resource.3 Students would, of course, NOT include
any color-coding or footnotes in their reference entries. However, for the sake of clarity and
ease in identifying what each entry represents, each one included in the reference list of this
sample paper is color-coordinated to its corresponding footnote, with a brief description of what
each depicts.
3 https://apastyle.apa.org/instructional-aids/creating-reference-list.pdf
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 30
References
Acborne, A., Finley, I., Eigen, K., Ballou, P., Gould, M. C., Blight, D., Callum, M., Feist, M.,
Carroll, J. E., Drought, J., Kinney, P., Owen, C., Owen, K., Price, K., Harlow, K.,
Edwards, K., Fallow, P., Pinkley, O., Finkel, F., & Gould, P. P. (2011). The emphasis of
the day after tomorrow. Strouthworks. 4
Alone, A. (2008). This author wrote a book by himself. Herald Publishers. 5
Alone, A., Other, B., & Other, C. (2011). He wrote a book with others, too: Arrange
alphabetically with the sole author first, then the others. Herald Publishers. 6
American Counseling Association. (2014). 2014 ACA code of ethics.
https://www.counseling.org/knowledge-center 7
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). https://www.doi.org/10.1176/appi.books.9780890425596 8
American Psychological Association. (n.d.). Positive transference. In APA dictionary of
psychology. Retrieved August 31, 2019, from https://dictionary.apa.org/positive-
transference 9
American Psychological Association. (2019). Publication manual of the American Psychological
Association (7th ed.). 10
4 Resource with 20 authors (maximum allowed by APA before special rule applies).
5 Entry by author who also appears as one of many authors in another resource (single author
appears first in list).
6 Multiple authors appear after same single-author resource.
7 Code of ethics.
8 Diagnostic manual.
9 Entry in a dictionary, thesaurus, or encyclopedia, with group author.
10 Resource with corporate author as publisher.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 31
Andrews, P. M. (2016). Congruence matters. Educational Leadership, 63(6), 12-15. 11
Aristotle. (1994). Poetics (S. H. Butcher, Trans.). The internet Classics Archive.
http://classics.mit.edu/Aristotle/poetics.html (Original work published ca. 350 B.C.E.) 12
Avramova, N. (2019, January 3). The secret to a long, happy, heathy life? Think age-positive.
CNN. https://www.cnn.com/2019/01/03/health/respect-toward-elderly-leads-to-long-life-
intl/index.html 13
Balsam, K. F., Martell, C. R., Jones, K. P., & Safren, S. A. (2019). Affirmative cognitive
behavior therapy with sexual and gender minority people. In G. Y. Iwamasa & P. A.
Hays (Eds.), Culturally responsive cognitive behavior therapy: Practice a supervision
(2nd ed., pp. 287-314). American Psychological Association.
https://doi.org/10.1037/0000119-012 14
Benoit, M., Bouthillier, D., Moss, E., Rousseau, C., & Brunet, A. (2010). Emotion regulation
strategies as mediators of the association between level of attachment security and PTSD
symptoms following trauma in adulthood. Anxiety, Stress, and Coping, 23(1), 101-118.
http://dx.doi.org/10.1080/10615800802638279
Brown, J. (2009). Ardent anteaters. Brockton.
Brown, M. (2009). Capricious as a verb. Journal of Grammatical Elements, 28(6), 11-12. 15
11 Journal article without DOI, from most academic research databases or print version.
12 Ancient Greek or Roman work.
13 Webpage on a news website.
14 Chapter in an edited book with DOI.
15 Resources by two authors with the same last name but different first names in the same year of
publication. Arrange alphabetically by the first initials.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 32
Cable, D. (2013). The racial dot map [Map]. University of Virginia, Weldon Cooper Center for
Public Service. https://demographics.coopercenter.org/Racial-Dot-Map 16
Canan, E., & Vasilev, J. (2019, May 22). [Lecture notes on resource allocation]. Department of
Management Control and Information Systems, University of Chile. https:// uchilefau.
academia.edu/ElseZCanan 17
Carlisle, M. A. (n.d.-a). Erin and the perfect pitch. Journal of Music, 21(3), 16-17. http:// make-
sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include 18
Carlisle, M. A. (n.d.-b). Perfect pitch makes sweet music. Journal of Music, 24(8), 3-6. http://
make-sure-it-goes-to-the-exact-webpage-of-the-source-otherwise-don’t-include
Centers for Disease Control and Prevention. (2018, January 23). People at high risk of
developing flu-related complications. https://www.cdc.gov/flu/highrisk/index.htm 19
De Vries R., Nieuwenhuijze, M., Buitendijk, S. E., & the members of Midwifery Science Work
Group. (2013). What does it take to have a strong and independent profession of
midwifery? Lessons from the Netherlands. Midwifery, 29(10), 1122-1128.
https://doi.org/10.1016/j.midw.2013.07.007 20
Double, C. (2008a). This is arranged alphabetically by the name of the title. Peters.
Double, C. (2008b). This is the second (“the” comes after “arranged”). Peters. 21
16 Map.
17 PowerPoint slides or lecture notes.
18 Online journal article with a URL and no DOI; also depicts one of two resources by the same
author with no known publication date.
19 Webpage on a website with a group author.
20 Journal article with a DOI, combination of individual and group authors.
21 Two resources by same author in the same year. Arrange alphabetically by the title and then
add lowercase letters (a and b, respectively here) to the year.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 33
Fistek, A., Jester, E., & Sonnenberg, K. (2017, July 12-15). Everybody’s got a little music in
them: Using music therapy to connect, engage, and motivate [Conference session].
Autism Society National Conference, Milwaukee, WI, United States.
https://asa.confex.com/asa/2017/webprogramarchives/Session9517.html 22
Forman, M. (Director). (1975). One flew over the cuckoo’s nest [Film]. United Artists. 23
Fried, D., & Polyakova, A. (2018). Democratic defense against disinformation. Atlantic Council.
https://www.atlanticcouncil.org/in-depth-research-reports/report/democratic-defense-
against-disinformation/ 24
Friedlander, M. L., Escudero, V., & Heatherton, L. (2002). E-SOFTA: System for observing
family therapy alliances [Software and training videos] [Unpublished instrument].
http://www.softa-soatif.com/ 25
GDJ. (2018). Neural network deep learning prismatic [Clip art]. Openclipart.
https://openclipart.org/detail/309343/neural-network-deep-learning-prismatic 26
Goldberg, J. F. (2018). Evaluating adverse drug effects [Webinar]. American Psychiatric
Association. https://education.psychiatry.org/Users/ProductDetails.aspx?
ActivityID=6172 27
Goldman, C. (2018, November 28). The complicate calibration of love, especially in adoption.
22 Conference session.
23 Video.
24 Report by individual authors at a government agency or other organization.
25 Measurement instrument.
26 Clip art or stock image.
27 Webinar, recorded.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 34
Chicago Tribune. 28
Graham, G. (2019). Behaviorism. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy
(Summer 2019 ed.). Stanford University.
https://plato.stanford.edu/archives/sum2019/entries/behaviorism 29
Guarino, B. (2017, December 4). How will humanity react to alien life? Psychologists have some
predictions. The Washington Post. https://www.washingtonpost.com/news/speaking-of-
science/wp/2017/12/04/how-will-humanity-react-to-alien-life-psychologists-have-some-
predictions/ 30
Hacker Hughes, J. (Eds.). (2017). Military veteran psychological health and social care:
Contemporary approaches. Routledge. 31
Haybron, D. M. (2008). Philosophy and the science of subjective well-being. In M. Eid & R. J.
Larsen (Eds.), The science of subjective well-being (pp. 17-43). Guilford Press. 32
Hollander, M. M. (2017). Resistance to authority: Methodological innovations and new lessons
from the Milgram experiment (Publication No. 10289373) [Doctoral dissertation,
University of Wisconsin-Madison]. ProQuest Dissertations and Theses Global. 33
Hutcheson, V. H. (2012). Dealing with dual differences: Social coping strategies of gifted and
lesbian, gay, bisexual, transgender, and queer adolescents [Master’s thesis, The College
28 Newspaper article without DOI, from most academic research databases or print version
29 Entry in a dictionary, thesaurus, or encyclopedia, with individual author.
30 Online newspaper article.
31 Edited book without a DOI, from most academic research databases or print version.
32 Book chapter, print version.
33 Doctoral dissertation, from an institutional database.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 35
of William & Mary]. William & Mary Digital Archive.
https://scholarworks.wm.edu/etd/1539272210/ 34
Kalnay, E., Kanimitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S.,
White, G., Whollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J.,
Mo, K. C., Ropelewski, C., Wang, J., Leetmaa, A., … Joseph, D. (1996). The
NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological
Society, 77(3), 437-471. http://doi.org/ fg6rf9 35
King James Bible. (2017). King James Bible Online. https://www.kingjamesbibleonline.org/
(Original work published 1769) 36
Lewin, K. (1999). Group decision and social change. In M. Gold (Ed.), The complex social
scientist: A Kurt Lewin reader (pp. 265-284). American Psychological Association.
https://doi.org/10.1037/10319-010 (Original work published 1948) 37
Liberty University. (2019). The online writing center. https://www.liberty.edu/online/casas/
writing-center/ 38
Liberty University. (2020). BIOL 102: Human biology. Week one, lecture two: Name of class
lecture. https://learn.liberty.edu 39
34 Thesis or dissertation, from the web (not in a database).
35 Resource with 21 or more authors. Note the ellipse (…) in place of the ampersand (&).
36 Religious work.
37 Work in an anthology.
38 Online webpage with URL.
39 Liberty University class lecture with no presenter named.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 36
McCauley, S. M., & Christiansen, M. H. (2019). Language learning as language use: A cross-
linguistic model of child language development. Psychological Review, 126(1), 1-51.
https://doi.org/10.1037/rev0000126 40
McCurry, S. (1985). Afghan girl [Photograph]. National Geographic.
https://www.nationalgeographic.com/magazine/national-geographic-magazine-50-years-
of-covers/#/ngm-1985-jun-714.jpg 41
Merriam-Webster. (n.d.). Heuristic. In Merriam-Webster.com dictionary. Retrieved 01/02/2020,
from http://www.m-w.com/dictionary/heuristic 42
National Cancer Institute. (2018). Facing forward: Life after cancer treatment (NIH Publication
No. 18-2424). U.S. Department of Health and Human Services, National Institutes of
Health. https://www.cancer.gov/publications/patient-education/life-after-treatment.pdf 43
National Nurses United. (n.d.). What employers should do to protect nurses from Zika.
https://www.nationalnursesunited.org/pages/what-employers-should-do-to-protect-rns-
from-zika 44
O’Keefe, E. (n.d.). Egoism & the crisis in Western values. http:// www. onlineoriginals.com/
showitem.asp?itemID-135 45
40 Typical journal article with doi.
41 Photograph.
42 Dictionary entry.
43 Report by a government agency or other organization.
44 Webpage on a website with no date.
45 Electronic only book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 37
Owen, C. (2012, Spring). Behavioral issues resulting from attachment have spiritual
implications [Unpublished manuscript]. COUN502, Liberty University. 46
Perigogn, A. U., & Brazel, P. L. (2012). Captain of the ship. In J. L. Auger (Ed.) Wake up in the
dark (pp. 108-121). Shawshank Publications. 47
Peters, C. (2012). COUN 506: Integration of spirituality and counseling. Week one, lecture two:
Defining integration: Key concepts. Liberty University.
https://podcasts.apple.com/us/podcast/defining-integration-key-
concepts/id427907777?i=1000092371727 48
Pew Research Center. (2018). American trend panel Wave 26 [Data set].
https://www.pewsocialtrends.org/dataset/american-trends-panel-wave-26 49
Prayer. (2015). http:// www exact-webpage 50
Project Implicit. (n.d.). Gender–Science IAT. https://implicit.harvard.edu/implici/taketest.html 51
Schatz, B. R. (2000, November 17). Learning by text or context? [Review of the book The social
life of information, by J. S. Brown & P. Duguid]. Science, 290, 1304.
https://doi.org/10.1126/science.290.5495.1304 52
Schmid, H.-J. (Ed.). (2017). Entrenchment and the psychology of language learning: How we
reorganize ad adapt linguistic knowledge. American Psychological Association; De
46 Citing a student’s paper submitted in a prior class, in order to avoid self-plagiarism.
47 Chapter from an edited book.
48 Liberty University class lecture using course details.
49 Data set.
50 Online resource with no named author. Title of webpage is in the author’s place.
51 Test, scale, or inventory itself.
52 Review of a book.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 38
Gruyter Mouton. https://doi.org/10.1037/15969-000 53
Second, M. P. (2011). Same author arranged by date (earlier first). Journal Name, 8, 12-13.
Second, M. P. (2015). Remember that earlier date goes first. Journal Name, 11(1), 18. 54
Shakespeare, W. (1995). Much ado about nothing (B. A. Mowat & P. Werstine, Eds.).
Washington Square Press. (Original work published 1623) 55
Strong, E. K., Jr., & Uhrbrock, R. S. (1923). Bibliography on job analysis. In L. Outhwaite
(Series Ed.), Personnel research series: Vol. 1, Job analysis and the curriculum (pp. 140-
146). https://doi.org/10.1037/10762-000 56
Tellegen, A., & Ben-Porah, Y. S. (2011). Minnesota Multiphasic Personality Inventory–2
Restructured Form (MPI-2-RF): Technical manual. Pearson. 57
U.S. Census Bureau. (n.d.). U.S. and world population clock. U.S. Department of Commerce.
Retrieved July 3, 2019, from https://www.census.gov/popclock 58
U.S. Securities and Exchange Commission. (2017). Agency financial report: Fiscal year 2017.
https://www.sec.gov/files/sec-2017-agency-financial-report.pdf 59
University of Oxford. (2018, December 6). How do geckos walk on water? [Video]. YouTube.
https://www.youtube.com/watch?v=qm1xGfOZJc8 60
53 Edited book with a DOI, with multiple publishers.
54 Two resources by the same author, in different years. Arrange by the earlier year first.
55 Shakespeare.
56 Electronic version of book chapter in a volume in a series
57 Manual for a test, scale, or inventory.
58 Webpage on a website with a retrieval date.
59 Annual report.
60 YouTube or other streaming video.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 39
Vedentam, S. (Host). (2015-present). Hidden brain [Audio podcast]. NPR. https://www.npr.org/
series/423302056/hidden-brain 61
Weinstock, R., Leong, G. B., & Silva, J. A. (2003). Defining forensic psychiatry: Roles and
responsibilities. In R. Rosner (Ed.), Principles and practice of forensic psychiatry (2nd
ed., pp. 7-13). CRC Press. 62
Yoo, J., Miyamoto, Y., Rigotti, A., & Ryff, C. (2016). Linking positive affect to blood lipids: A
cultural perspective [Unpublished manuscript]. Department of Psychology, University of
Wisconsin-Madison. 63
61 Podcast.
62 Chapter in an edited book without a DOI, from most academic research databases or print
version.
63 Unpublished manuscript with a university cited.
SAMPLE APA-7 PAPER FOR GRADUATE STUDENTS 40
Appendix
Annotated Bibliography
Cross, D., & Purvis, K. (2008). Is maternal deprivation the root of all evil? Avances en
Psycologia Latinoamericana, 26(1), 66-81. https://doi.org/10.1037/0002-9432.77.4.582
Weaving spiritual applications throughout the article, the authors incorporate a plethora
of references to substantiate that maltreatment has a direct connection to attachment
disorders. They provide articulate and heavily supported reasoning, detailing the specific
causes of maternal deprivation individually and then incorporating them in a broader
sense to answer the article’s title in the affirmative.
Feldman, R. (2007). Mother-infant synchrony and the development of moral orientation in
childhood and adolescence: Direct and indirect mechanisms of developmental continuity.
American Journal of Orthopsychiatry, 77(4), 582-597.
This longitudinal study tracked 31 Israeli children from ages 3 months to 13 years
(infancy to adolescence). There were direct parallels noted between increased
attachment/coherence and the child’s moral cognition, empathy development, and verbal
IQ. Toddlers who were able to regulate their own behavior later proved to excel in lead-
lag structures and language skills.