Content uploaded by Hana Zickgraf
Author content
All content in this area was uploaded by Hana Zickgraf on Jun 27, 2017
Content may be subject to copyright.
Psychological Science
2014, Vol. 25(10) 1955 –1959
© The Author(s) 2014
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0956797614545131
pss.sagepub.com
Research Report
For almost 100 years, psychology as a discipline has
been jockeying to be classified as a natural science. The
natural sciences are often viewed as more prestigious
and more advanced than the social sciences, and psy-
chologists have strived to be viewed as “baby” natural
scientists rather than the most “scientific” social scien-
tists. In recent decades, psychology has appended the
word science to itself (psychological science) and to for-
mer subareas that have now become departments (e.g.,
cognitive science, neuroscience). This trend in nomen-
clature is also illustrated in the names of several rela-
tively new journals in psychology: Social Psychological
and Personality Science, Perspectives on Psychological
Science, Current Directions in Psychological Science,
and Clinical Psychological Science. And unlike text-
books in more secure natural sciences, introductory psy-
chology textbooks usually include a section on what
science is and what the scientific method is, implicitly or
explicitly claiming that psychology is a (natural) science.1
We think psychologists are defending their belief that
psychology is properly categorized as a natural science;
in contrast, physicists, biologists, and chemists, who are
more firmly placed in the natural sciences, rarely feel
compelled to defend their disciplines as natural sciences.
This phenomenon is an instance of a general social
phenomenon: the tendency for border or marginal mem-
bers of positively valenced groups to emphasize their
membership in those groups. We call this tendency to
emphasize membership at a positively valenced border
the tendency to create asymmetrical social Mach bands,
and ground this social phenomenon in some basic percep-
tual science. Borders are sources of particularly rich and
useful information, and perceptual systems enhance them.
Thus, in the visual system, at the border between a uni-
formly dark and uniformly light patch, the area perceived
to be brightest is just over the border on the more illumi-
nated side, whereas the area perceived to be darkest is just
545131PSSXXX10.1177/0956797614545131Rozin et al.Asymmetrical Social Mach Bands
research-article2014
Corresponding Author:
Paul Rozin, University of Pennsylvania, 3720 Walnut St., Philadelphia,
PA 19104
E-mail: rozin@psych.upenn.edu
Asymmetrical Social Mach Bands:
Exaggeration of Social Identities on the
More Esteemed Side of Group Borders
Paul Rozin, Sydney E. Scott, Hana F. Zickgraf, Flora Ahn, and
Hong Jiang
University of Pennsylvania
Abstract
Perceptual processes generally enhance borders, because of their high information value. Mach bands are an example
in vision. In the social world, borders are also of special significance; one side of a border is generally more esteemed
or valued than the other. We claim that entities (individuals, groups) that are just over the border on the positive side
tend to exaggerate their membership on the positive side (asymmetrical social Mach bands). We demonstrate this by
showing that (a) master’s-degree universities use the word university to describe themselves more than major graduate
universities do, (b) small international airports use the word international to describe themselves more than major
airports do, and (c) University of Pennsylvania students, who are affiliated with a “marginal” Ivy League school, use
the word Ivy to describe their school more than Harvard students do.
Keywords
social, borders, enhancement, groups, open data
Received 11/5/13; Revision accepted 6/21/14
1956 Rozin et al.
over the border on the darker side. The effect of these
Mach bands (Ratliff, 1965) is to accentuate borders.
Social Mach bands also serve to accentuate borders, but
they depart in character from visual Mach bands in that
they are often asymmetrical. Social categories are almost
always valenced, so that one side of a social border is
more esteemed than the other. Asymmetrical social Mach
bands arise as individuals, institutions, or other social enti-
ties create, maintain, and broadcast social identities—
identities derived from group membership (Tajfel &
Turner, 1979). We propose that individuals who are in a
favored group near the group’s boundary emphasize their
presence on the favored side of that boundary. Individuals
generally prefer to be in higher-status or more positively
valenced groups, both to enhance their self-esteem and to
project a more impressive self to others (Hogg & Abrams,
1988; Tajfel & Turner, 1979). We assume that the same
holds true for institutions and other groups.
Asymmetrical social Mach bands are also present in
the domain of sociolinguistics. Labov (1972) showed
that in formal speech, middle-class speakers of American
English, compared with both lower- and upper-class
speakers, demonstrate more precise and careful pro-
nunciation of the dialect associated with educated and
high-socioeconomic-status English speakers. This phe-
nomenon is known as hypercorrection. By using hyper-
corrected pronunciation borrowed from the upper
social classes, speakers on the border between upper
and lower classes emphasize their membership in the
upper classes.
In the domain of intergroup attitudes, research by
White and her colleagues explored the related phenom-
enon of horizontal hostility (White & Langer, 1999;
White, Schmitt, & Langer, 2006). They demonstrated that
members of “extreme” minority groups, who are often
proud of their distinction, show more negative attitudes
toward bordering, but less extreme, minority groups
than toward majority groups. Thus, Greek Communists
are more hostile toward Greek Progressive party mem-
bers than toward Greek Conservatives, members of var-
sity teams are more hostile toward members of
junior-varsity teams than toward members of intramural
teams, and conservative Jews are more hostile toward
reform Jews than toward nonpracticing Jews. This excel-
lent work by White and her colleagues served as a model
for the present work, because they described a phenom-
enon and then established its generality by showing that
it is manifest in a wide range of contexts. Here, we dem-
onstrate the phenomenon of asymmetrical social Mach
bands by showing that marginal members of three
groups (universities, international airports, and the Ivy
League) emphasize their group membership more than
do secure, quintessential members of the relevant
groups.
Study 1
Technically, any school of higher education that offers at
least one postbaccalaureate degree is a university (as
opposed to a college). We presume that it is prestigious
to be a university, and that institutions at the border of
the university category (i.e., those with few or no Ph.D.
programs) put greater emphasis on their status as a uni-
versity compared with schools that have more substantial
higher-degree programs.
Method
Listings of the top national universities and the top mas-
ter’s universities in the United States in 2012 were taken
from U.S. News & World Report (2012b, 2012c). According
to U.S. News, the 280 listed national universities “offer a
full range of undergraduate majors, master’s, and doctoral
degrees,” whereas the 626 listed master’s universities
(called regional universities) “offer a full range of under-
grad programs and some master’s programs but few doc-
toral programs” (U.S. News & World Report, 2012a). To be
included in our sample, a university had to meet the fol-
lowing criteria: First, its formal name had to include
University. Second, its names could not include a location
(e.g., “state,” “Michigan”). Universities with names includ-
ing locations, such as “University of Michigan,” are less
likely than others to refer to themselves in an abbreviated
form, such as “Michigan,” because the abbreviated form is
ambiguous and could be naming the location (the state of
Michigan) or the university (the University of Michigan).
Third, the university had to have an “About Us” Web page
where the university was referenced in the text at least
once.
Two undergraduate judges blind to our hypotheses
coded the “About Us” Web pages for the 55 national uni-
versities and 151 master’s universities included in our final
sample. Each self-reference of the specific university was
categorized as either a “university mention” (e.g., “Harvard
University” or “the University”) or an “other mention” (e.g.,
“Harvard”). Acronyms (e.g., “SMU”) were excluded,2 as
were other uses of the word university not in reference to
the specific institution. For each university’s Web page, the
percentage of self-references that included the word uni-
versity was calculated. The judges’ ratings were highly reli-
able (intraclass correlation coefficient, or ICC = .99).
Results
Universities at the border of the university category
emphasized their university identity more than arche-
typal universities did. On average, master’s universities
used the word university in 62.2% of self-references
(SD = 31.6%, n = 151), whereas the corresponding mean
Asymmetrical Social Mach Bands 1957
for national universities was 46.4% (SD = 31.0%, n = 55),
t(204) = 3.19, p = .002, 95% confidence interval (CI) for
the difference = [6.0%, 25.5%], d = 0.50.
Study 2
We presume that airports that offer international flights
are in a higher-status category than airports that offer
only domestic flights. International airports in the United
States differ substantially in the number of international
flights offered—ranging from a single route to Canada to
dozens of international flights per day. We hypothesized
that small international airports emphasize their status as
international airports more than large ones do.
Method
A list of airports in the United States was obtained from the
Web site of the Federal Aviation Administration, or FAA
(2013). We limited our sample to the 96 airports with
International in their formal title and an “About Us” Web
page (in English) that included at least one self-reference.
The FAA Web site provided the total number of enplane-
ments for each airport in the 2012 calendar year and clas-
sified each airport as a nonhub (n = 23), small hub (n =
34), medium hub (n = 19), or large hub (n = 20). We
compared small hubs with large hubs. As a robustness
check, we also examined how airports’ self-references dif-
fered as a function of the total number of enplanements.
Two undergraduate judges, blind to our hypotheses,
coded the “About Us” Web page for each airport. Each
self-reference of the specific airport was categorized as
either an “international mention” (e.g., “Philadelphia
International Airport” or “Philadelphia International”) or
an “other mention” (e.g., “Philadelphia Airport” or
“Philadelphia,” as in “Philadelphia has four runways”).
Official acronyms (e.g., “PHL”) were excluded,3 as were
nonspecific uses of the word airport. For each airport’s
Web page, the percentage of self-references that included
the word international was calculated. The judges’ rat-
ings were highly reliable (ICC = .97).
Results
Small (n = 34) airports were more likely to emphasize
their status as an international airport than large air-
ports were (n = 20). On average, small airports used
the word international in 68.2% (SD = 30.3%) of self-
references, whereas the corresponding mean for large
airports was 31.4% (SD = 29.1%), t(52) = 4.38, p < .001,
95% CI for the difference = [20.0%, 53.8%], d = 1.24. To
ensure that our results were not due to the way the
FAA categorizes small and large airports, we examined
whether larger international airports (i.e., airports with
a larger number of annual enplanements) were less
likely to use international in self-references. We found
that they were: More enplanements in the 2012 calendar
year correlated with less frequent use of international
in self-references, r(94) = −.331, p = .001.
Study 3
We presume that membership in the Ivy League repre-
sents high status and is positively valenced. Pilot studies
showed that the University of Pennsylvania (Penn) is not
widely recognized as an Ivy university, whereas Harvard
is a quintessential Ivy. Of a sample of 204 American par-
ticipants recruited from Amazon’s Mechanical Turk, 83
(40.7%) mentioned Harvard in free-association responses
to “Ivy League,” whereas only 3 (1.5%) mentioned the
University of Pennsylvania, χ2(1, N = 204) = 78.01, p <
.001. We investigated whether students at the University
of Pennsylvania are more likely to think of the phrase Ivy
League in association with their school than students at
Harvard University are. Moreover, we investigated
whether this social-Mach-band effect would be present in
communicative, impression-management contexts only
or in both private and communicative contexts. From a
social identity approach, one might predict these effects
to occur in both communicative and private contexts.
When individuals construct, maintain, and broadcast
identities, their social identities from group membership
provide integral knowledge and emotional value (Tajfel
& Turner 1979). Self-categorization and social identity
construction change how people perceive themselves
and others (Hogg & Abrams, 1988; Tajfel & Turner, 1979).
Method
A total of 54 Harvard and 53 Penn students participated
in the study. We employed a 2 (university: Harvard, Penn) ×
2 (context: private, public) between-subjects, full-factorial
design. Two female research assistants at each institution
asked undergraduates leaving introductory psychology
lectures in the spring 2012 semester and social psychol-
ogy lectures in the fall 2012 semester to fill out a psychol-
ogy research survey. Participants were presented with
one of two forms that requested, “Please write down 7
things you think of [when you think of your university /
when you describe your university to other people]. You
may write in words or phrases” (emphasis in original). At
random, half the participants received the private-context
(“think of”) form, and half received the public-context
(“describe”) form.
Results
Penn students were more likely to mention “Ivy League”
or “Ivy” in describing their university than Harvard stu-
dents were, but directing individuals to answer in a
1958 Rozin et al.
public context, if anything, decreased “Ivy” mentions. In
the public condition, none of 30 Harvard students men-
tioned “Ivy,” whereas 9 of 33 Penn students (27.3%) did.
In the private condition, 4 of 24 Harvard students (16.7%)
mentioned “Ivy,” whereas 7 of 20 Penn students (35%)
did.
In a binomial logistic regression model with Firth bias
correction,4 being a Penn student (rather than a Harvard
student) predicted higher probability of mentioning “Ivy”
(β = 3.16, p = .001), and private context increased the
probability of mentioning “Ivy” (β = 2.59, p = .03). There
was also a marginally significant interaction such that
Penn students were less affected by being in a private
context than Harvard students were (β = −2.23, p = .1).
We also conducted Pearson’s chi-square tests with a Yates
continuity correction. A test collapsing across the two
contexts revealed that Penn students were more likely to
mention “Ivy” than Harvard students were (4 of 54
Harvard students and 16 of 53 Penn students mentioned
“Ivy”), χ2(1, N = 107) = 7.70, p = .006, 95% CI for the dif-
ference = [6.7%, 38.8%], but a test collapsing across uni-
versities indicated that context (private vs. public) did not
affect the frequency of “Ivy” mentions, χ2(1, N = 107) =
1.32, p > .1, 95% CI for the difference = [−7.0%, 28.1%].
Discussion
We have demonstrated the existence of asymmetrical
social Mach bands at one individual and two institutional
social borders. We have also presented an initial analysis
of the mechanisms by which social Mach bands arise; the
results of Study 3 suggest that self-construal is of greater
importance than social communication.
We believe that effects like those illustrated here are
widespread, although it is sometimes difficult to collect
the appropriate data. For example, we believe that the
following instances of the asymmetrical social-Mach-
band effect are likely to occur: (a) greater display of
officer status by lieutenants as opposed to colonels;
(b) greater use of doctor titles by osteopaths, dentists,
and chiropractors as opposed to medical doctors; (c) greater
display of team membership by members of junior-
varsity as opposed to varsity teams; (d) greater display of
sorority or fraternity membership, or honor-society mem-
bership, by new than by continuing members; (e) greater
display of wealth by the nouveau riche than by “old
money”; and (f) greater display of affiliation with presti-
gious universities (e.g., jackets) by freshmen than by
seniors.
Our results run counter to similarity-attraction theory.
Individuals near the border of a highly esteemed group
are presumably more similar to marginal out-group mem-
bers than to individuals at the center of the highly
esteemed group, and similarity breeds attraction (Byrne,
1971). Therefore, this theory suggests that border mem-
bers would deemphasize the boundary to create social
bonds to the more similar, and therefore more attractive,
out-group members. Such an effect might be expected
when the valence difference between the two relevant
groups is small or when the positively valenced group
feels that its optimal distinctiveness is secure (Brewer,
1991, 2003).
We leave open questions about the relative importance
of self-esteem and public-esteem motivations for asym-
metrical social Mach bands. We also leave open the
importance of a third motivation, uncertainty reduction
(an epistemic motivation to understand the self and the
social world; Hogg, 2000). We recognize that the contrast
between border and central members observed in our
data is not necessarily due to marginal members enhanc-
ing their border status, but could instead be due to central
members deemphasizing their social identity. Finally, we
have not actually demonstrated the asymmetry of the
social Mach bands; we did not test whether individuals at
the border of the less positively valenced group also over-
emphasize their membership in that group, compared
with members further from the border in that group. Our
aim in this set of studies was to identify one feature of
group dynamics, link it to a principle in perception, and
provide some evidence for its existence and extent.
Author Contributions
P. Rozin developed the study concept, participated in the
design of the individual studies, and supervised data collection
and analysis. Study 1 was designed by F. Ahn, S. E. Scott, and
H. F. Zickgraf. Study 2 was designed by H. Jiang, S. E. Scott,
and H. F. Zickgraf. Study 3 was designed by S. E. Scott and
H.F. Zickgraf. S. E. Scott and H. F. Zickgraf collected and ana-
lyzed the data for all three studies. The manuscript was drafted
by P. Rozin, S. E. Scott, and H. F. Zickgraf.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
Funding
This research was supported by funds from the Positive
Psychology Center at the University of Pennsylvania.
Open Practices
All data and some materials have been made publicly available
via Open Science Framework and can be accessed at https://
osf.io/28ejf/. Some of the materials are not posted because of
copyright concerns. The complete Open Practices Disclosure
for this article can be found at http://pss.sagepub.com/content/
by/supplemental-data. This article has received the badge for
Asymmetrical Social Mach Bands 1959
Open Data. More information about the Open Practices badges
can be found at https://osf.io/tvyxz/wiki/view/ and http://pss
.sagepub.com/content/25/1/3.full.
Notes
1. We conducted two pilot studies concerning the frequency
of the use of the word science in journal titles and department
names and mentions of science-related words (e.g., experiment,
empirical) in the indices of introductory textbooks in anthro-
pology, sociology, psychology, biology, chemistry, and physics.
Psychology and biology tended to score highest of all disci-
plines on all three measures.
2. We also conducted an analysis in which acronyms were
included as self-references in the “other” category and obtained
the same pattern of results as reported here.
3. We also conducted an analysis in which acronyms were
included as self-references in the “other” category and obtained
the same pattern of results as reported here.
4. We applied Firth’s (1993) bias-reduction method because one
of our cells had zero observations, which resulted in an estima-
tion problem of complete separation.
References
Brewer, M. B. (1991). The social self: On being the same and dif-
ferent at the same time. Personality and Social Psychology
Bulletin, 17, 475–482.
Brewer, M. B. (2003). Optimal distinctiveness, social iden-
tity, and the self. In M. R. Leary & J. P. Tangney (Eds.),
Handbook of self and identity (pp. 480–491). New York,
NY: Guilford Press.
Byrne, D. E. (1971). The attraction paradigm. New York, NY:
Academic Press.
Federal Aviation Administration. (2013). Passenger board-
ing (enplanement) and all-cargo data for U.S. airports.
Retrieved from http://www.faa.gov/airports/planning_
capacity/passenger_allcargo_stats/passenger/
Firth, D. (1993). Bias reduction of maximum likelihood esti-
mates. Biometrika, 80, 27–38.
Hogg, M. A. (2000). Subjective uncertainty reduction through
self-categorization: A motivational theory of social iden-
tity processes. European Review of Social Psychology, 11,
223–255.
Hogg, M. A., & Abrams, D. (1988). Social identifications: A
social psychology of intergroup relations and group pro-
cesses. London, England: Routledge.
Labov, W. (1972). Sociolinguistic patterns. Philadelphia: Univer-
sity of Pennsylvania Press.
Ratliff, F. (1965). Mach bands: Quantitative studies on neural
networks in the retina. San Francisco, CA: Holden-Day.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of
intergroup conflict. In W. G. Austin & S. Worchel (Eds.),
The social psychology of intergroup relations (pp. 33–47).
Monterey, CA: Brooks Cole.
U.S. News & World Report. (2012a). Education: Colleges.
Retrieved from http://colleges.usnews.rankingsandreviews
.com/best-colleges
U.S. News & World Report. (2012b). National university rank-
ings. Retrieved from http://colleges.usnews.rankingsandre
views.com/best-colleges/rankings/national-universities
U.S. News & World Report. (2012c). Regional university rank-
ings. Retrieved from http://colleges.usnews.rankingsandre
views.com/best-colleges/rankings/regional-universities
White, J. B., & Langer, E. J. (1999). Horizontal hostility: Relations
between similar minority groups. Journal of Social Issues,
55, 537–559.
White, J. B., Schmitt, M. T., & Langer, E. J. (2006). Horizontal
hostility: Multiple minority groups and differentiation from
the mainstream. Group Processes & Intergroup Relations, 9,
339–358.
