Академический Документы
Профессиональный Документы
Культура Документы
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press and Comparative and International Education Society are collaborating with
JSTOR to digitize, preserve and extend access to Comparative Education Review.
http://www.jstor.org
This study identifies a theoretical mechanism that could potentially affect public university admissions standards in a context of demographic change. I explore how demographic changes at a prestigious public university in the United States affect individuals evaluations of college applications. Responding to a line graph that randomly
displays a freshmen enrollment trend toward a white plurality or an Asian American
plurality, white student evaluators lower their minimum class rank standard for admitting
white applicants when exposed to an Asian American plurality trend. They also raise
the minimum test percentile standard for admitting Asian American applicants. Notably,
plurality trend does not affect Asian American student evaluators minimum standards
for recommending admission. Applications differed only by applicant race.
The declining proportion of whites in the United States population represents a possible sea change in the future meaning of diversity for higher
education administrators and staff. The US Census Bureau projects that in
2042, the country will become one in which current minority groups will
constitute the majority of the population. More recent demographic projections from the 2010 Census suggest that this shift may likely arrive well
before 2040 (Frey 2011b). William H. Frey (2011a) proposes that, among
infants, the United States has already passed the demographic tipping point,
with over half of infants under the age of 1 hailing from a nonwhite ethnoracial group. As this nonwhite infant population ages, the school-aged
I thank Lawrence D. Bobo, Hazel Markus, Alicia Simmons, Mary Murphy, MarYam Hamedani, and
members of the Stanford Sociology Departments Workshop on Social Psychology, especially Karen Cook
and Cecilia Ridgeway, for their feedback and support. I would particularly like to thank Amber Davis,
Alexandra Harwin, Whitney Martin, Trista Shi, and Roselyn Thomas, for their assistance with the data
collection. Finally, I thank the editors of Comparative Education Review and three anonymous reviewers
for their constructive criticism. This study was funded by a Graduate Dissertation Fellowship from
Stanford Universitys Research Institute for Comparative Studies in Race and Ethnicity, a Ford Foundation Diversity Dissertation Fellowship, and a Public Policy Institute of California Dissertation Fellowship
during various stages of the research. Data collection was funded by a National Science Foundation
Dissertation Improvement Grant (grant no. SES-0802645) and a grant from the Stanford University
Sociology Research Opportunity Fund. All errors are my own.
Received December 1, 2011; revised June 24, 2012; accepted October 9, 2012; electronically published May 13, 2013
Comparative Education Review, vol. 57, no. 3.
2013 by the Comparative and International Education Society. All rights reserved.
0010-4086/2013/5703-0003$10.00
Comparative Education Review
369
SAMSON
population will soon reflect this demographic shift. White residents are already a minority in four states: California, Texas, New Mexico, and Hawaii.
In the midst of such demographic flux, might a drop in group position affect
how whites (as the dominant group) evaluate applicants in a meritocratic
contest over scarce and desired educational opportunities?
The current Asian American student plurality at the University of California system,1 the top tier of public higher education in the state, provides
a nascent social context epitomizing a drop in group position among whites,
as the rise in Asian American enrollments over the past decade came largely
at the cost of white student enrollments. Group position theory, a framework
tested primarily with probability sample survey data (Bobo 1983, 1999; Bobo
and Tuan 2006; Charles 2006), posits a course of action among those who
occupy the dominant group position: they will adopt attitudes and behaviors
that allow them to preserve as much advantage and privilege for themselves
as they can within an institution where fairness and meritocracy are valued.
This article tests whether such a response can be detected in an applicantevaluations study using a controlled laboratory experiment.
This whiteAsian American group dynamic presents two puzzles for the
long-standing study of double standards in evaluations. Prior research has
largely focused on disparities in standards among groups possessing different
stereotypical attributes that can be ranked hierarchically (Biernat et al. 1991,
2003; Biernat and Manis 1994; Biernat and Kobrynowicz 1997). Such studies
typically compared the historically dominant group (i.e., whites) with a subordinate group (i.e., blacks). The first puzzle then is that unlike previous
studies on shifting standards of evaluation, both whites and Asian Americans
can be perceived to possess comparably high academic competence or ability.
Some studies have even ceded higher perceived competence to Asian Americans rather than whites (Wong et al. 1998; Fiske et al. 2002), making Asian
Americans prime competitors to whites. This leads to the second puzzle:
Does racial group positioning still matter in a nation increasingly viewed as
postracial? The real world phenomenon of competing interests in prestigious university enrollments between whites and Asian Americans allows the
use of credible cues signifying a change in group position. The present study
addresses these two puzzles using laboratory experiments that have previously
identified the effects of making group identifiers such as race or gender
salient.2
1
While the term majority is a more widely understood concept, we use the term plurality here
rather than majority to indicate a demographic context in which a group has become the most numerous
or predominant group among three or more groups, but without its percentage having surpassed 50
percent. This studys use of the term plurality is consistent with its use in popular media, e.g., Asian
plurality (Chandler 2008). Also the majority/plurality distinction is implied, though not defined, in
some scholarly studies (e.g., Blacks or Hispanics have become the population majority, or at least
plurality [McClain and Karnig 1990, 539]).
2
Biernat et al. (1991, 2003); Biernat and Manis (1994); Biernat and Kobrynowicz (1997); Biernat
and Fuegen (2001).
370
August 2013
371
SAMSON
and some sociohistorical research (Almaguer 1994), the group position paradigm has not involved any laboratory-based experiments exploring double
standards in application or resume evaluations.
Utilizing a resume evaluation experiment that has revealed a number of
outcome and standards disparities between ethno-racial groups (Pager 2003;
Bertrand and Mullainathan 2004; Correll et al. 2007), the present study draws
historically from Karabel and conceptually from Blumers group position
theory to test whether a change in group position would result in a change
of actual standards of evaluation by which a members of a group are evaluated.
The central hypotheses are these:
H1A:
When whites as the dominant group are primed with Asian American plurality, they will lower the standard of evaluation by which whites are judged,
and
H1B:
H2A:
When whites as the dominant group are primed with Asian American plurality, they will increase the number of positive outcome recommendations
for whites, and
H2B:
August 2013
1987; Espenshade and Radford 2009), which may mask academic performance differences between Asian American subgroups.3 Additionally, the
present studys university admissions evaluation task involves nonwhite applicants (Asian Americans) who are typically not among the target population
for higher education affirmative action policies and therefore unlikely to
expose evaluators to affirmative actionrelated pressures (Norton et al. 2004).
Finally, none of the aforementioned studies randomly assign participants to
different social (demographic) contexts under which the application evaluation is being conducted, the difference being whether the dominant group
retains or has lost its plurality share of scarce and desired opportunities.
Laboratory Experiment
Some might question the characterization of Asian Americans as a subordinate group, given
decreasing whiteAsian American inequality on some aggregate socioeconomic indicators such as income
and educational attainment (Sakamoto et al. 2009; Kim and Sakamoto 2010). Since the present study
explicitly draws on the social psychological approach of group position theory, see Blumer (1958) and
Bobo (1999) for social psychological arguments against reducing the social construction of dominant
and subordinate groups and the hierarchical connection between them to relative rankings on a handful
of structural socioeconomic indicators.
4
Some might argue that solely evaluating an applicants academic record without including extracurricular activities and the personal essay threatens the external validity of the present study. However,
the University of California for a period of time used this particular college admissions procedure,
separating academic and nonacademic portions of an applicants portfolio for separate evaluation; the
current design was specifically modeled on this example.
Comparative Education Review
373
SAMSON
told that the two independently scored sections would later be combined to
produce a comprehensive review score for the applicant. Apart from the outsourcing of research on student input in the admissions process, this cover
story closely resembles the cover story used in another laboratory-based experiment (Hodson et al. 2002) tasking undergraduate students to evaluate
university applications in order to test the theory of aversive racism (Dovidio
and Gaertner 2000; Hodson et al. 2004).5
Setting the University Context
Unlike the University of California system with its Asian American student plurality, the present
studys sample was drawn from an undergraduate population where Asian American students represented almost a quarter of total undergraduate enrollments. Minority students (i.e., Asian American,
African American, Hispanic, and Native American) together constituted slightly less than half the total
undergraduate population. This difference necessitated the outsourcing component of the cover story,
rather than simply telling participants that they were admission committee members at their own
institution where no Asian American plurality existed.
374
August 2013
Fig. 1.Control and experimental prompts in laboratory-based study 1 counterbalanced the trend
lines colors for white and Asian American enrollments to minimize the possible influence that the
colors might have on the dependent variables.
375
SAMSON
After reviewing these university web pages, the participant read a fabricated
admissions application that contained a second experimental manipulation: race
of applicant. The admissions applications were identical in all respects, except
for the race of applicant. A race of applicant check box on the application
portfolio indicated either white or Caucasian or Asian American. The applicant was also listed as a US citizen who spoke English at home.
Academic information provided on the fabricated application suggested
a well-qualified applicant. Because percentile rank served as a dependent
variable, the percentile ranks corresponding to the scores listed on the application, though provided below, were not included with the application.
The application first listed ACT scores of 31 for English (94th percentile for
students from the state of California), 28 for Math (85th percentile), 30 for
Reading (89th percentile), 27 for Science (90th percentile), and 29 for Composite (92nd percentile). The application next listed SAT I/Reasoning scores
of 670 for Verbal/Critical Reading, 590 Math, and 650 Writing. According
to College Board data, the hypothetical applicants composite SAT score
(1,910 p 670 590 650) corresponds to the 89th percentile among
college-bound seniors nationwide in 2007. SAT II/Subject Test scores were
listed for Literature (660, 70th percentile) US History (700, 75th percentile),
and Math IC (590, 42nd percentile). The application next listed an academic
honor: US History Award, eleventh grade. Finally, the application listed the
students senior year course selection: AP English II, AP Calculus AB, AP
French, Physics, and US History II. Neither GPA nor class rank (another
dependent variable) were included on the application. Taken together, the
hypothetical applicant presented respectable academic qualifications, scoring
generally in the top 12 percent or higher according to most of the applicants
test scores, and enrolled in challenging, advanced placement courses. Evaluators took note of the applicants good qualifications, describing the applicant as intelligent and capable, with ratings of 4.81 (intelligence) and 5.0
(capability) on a scale from 0 (not at all) to 7 (extremely).
Application Review
Throughout the review of the universitys web pages, the reading of the
admissions application, and the subsequent evaluation of the application, the
participant filled out a web-based survey containing questions about the university, the applicant, and a number of admissions-related items. To control
for possible cross-gender evaluation effects, the participant was matched with
same-gender application portfolios. The survey also included task engagement and manipulation checks.
376
August 2013
Dependent Variables
The admissions outcome variable was dichotomous and binary coded, with
1 indicating a recommendation for admission and 0 indicating no recommendation for admission.
Study Debriefing and Suspicion Probe
377
SAMSON
Sample Characteristics
378
August 2013
conditions is not significantly different from a random distribution; no condition was more likely than another to elicit suspicion.
The sample of 128 participants varied along some demographic and social
characteristics. Sixty-four percent of the sample was female, with post hoc
chi-square analysis revealing a random distribution of participants gender
across experimental conditions. The average age was 19.6 years (SD p 1.27),
with 47 percent identifying as freshmen, 25 percent as sophomores, 10 percent as juniors, 15 percent as seniors, and 2 percent as graduate students.
Participants also hailed from a wide variety of campus majors: 13 percent
from the humanities, 21 percent from the social sciences, 39 percent in
sciences or engineering, 15 percent were undeclared, and the remainder
identified as double majors across multiple disciplinary categories.
Finally, using a controlled laboratory experiment among college students
to identify differences in admission standards might raise reasonable concerns
about external validity. However, previous research identifying a theoretical
mechanism using a laboratory experiment that placed college students in
the role of employers found results similar to those produced by an audit
study of actual employers (Correll et al. 2007). While the present study is
not meant to identify evaluator biases in college admissions processes per se
but to more broadly determine if a change in group position might cause
individuals in general to alter their standards of evaluation, it does propose
a new theoretical mechanism that could produce a previously undocumented
bias in evaluations that skilled evaluators should at the least be trained to
avoid, even though we are not arguing that this mechanism is producing the
already observed inequality in qualifications between white and Asian American students (Bunzel and Au 1987; Espenshade and Radford 2009).9
Results
379
SAMSON
TABLE 1
Analysis of Variance for Minimum Class Rank and Standardized Test Percentile
(Pooled Sample)
Partial SS
Class rank (N p 128):
Model
Race of applicant
School demographics
Race of applicant # school
demographics
Race of evaluator
Race of evaluator # race of
applicant
Race of evaluator # school
demographics
Race of evaluator x race of
applicant # school demographics
Residual
Total
R2
Standardized test percentile
(N p 128):
Model
Race of applicant
School demographics
Race of applicant # school
demographics
Race of evaluator
Race of evaluator # race of
applicant
Race of evaluator # school
demographics
Race of evaluator # race of
applicant # school
demographics
Residual
Total
R2
df
MS
Prob 1 F
3,195.34
325.13
2.00
7
1
1
456.48
325.13
2.00
1.87
1.33
.01
.08
.25
.93
871.53
128.00
1
1
871.53
128.00
3.57
.53
.06
.47
318.78
318.78
1.31
.26
603.78
603.78
2.48
.12
1
120
127
946.13
243.79
255.51
3.88
.05
1,479.13
648.00
.03
7
1
1
211.30
648.00
.03
.75
2.31
0
.63
.13
.99
603.78
18.00
1
1
603.78
18.00
2.15
.06
.15
.80
40.50
40.50
.14
.70
11.28
11.28
.04
.84
1
120
127
157.53
281.04
277.20
.56
.46
946.13
29,254.88
32,450.22
.0985
157.53
33,724.75
35,203.88
.042
Note.The data source is a controlled laboratory study conducted by the author. Partial SS: partial sum of squares.
MS: mean sum of squares. Total: total sum of squares.
August 2013
TABLE 2
Analysis of Variance for Minimum Class Rank and Standardized Test Percentile
by Race of Evaluator
Partial SS
df
2,085.05
.02
268.14
3
1
1
1,816.89
13,312.06
15,397.11
.14
MS
Prob 1 F
695.02
.02
268.14
3.13
0
1.21
.0320*
.993
.276
1
60
63
1,816.89
221.87
244.40
8.19
.0058**
982.30
643.89
337.64
3
1
1
327.43
643.89
337.64
1.23
2.42
1.27
.306
.125
.264
.77
15,942.81
16,925.11
.06
1
60
63
.77
265.71
268.65
.957
1,201.56
506.25
6.25
3
1
1
400.52
506.25
6.25
2.33
2.95
.04
.083
.091
.849
689.06
10,294.88
11,496.44
.10
1
60
63
689.06
171.58
182.48
4.02
.0496*
259.56
182.25
5.06
3
1
1
86.52
182.25
5.06
.22
.47
.01
.881
.497
.910
72.25
23,429.88
23,689.44
.01
1
60
63
72.25
390.50
376.02
.19
.669
Note.The data source is a controlled laboratory study conducted by the author. Partial SS: partial sum of squares.
MS: mean sum of squares. Total: total sum of squares.
* P ! .05.
** P ! .01.
381
Fig. 2.Minimum class rank standard by applicant race and university demographics (white evaluators)
demographics (see table 2). As figure 3 displays, white evaluators raised the
minimum standardized test score percentile (t p 2.01, df p 30, P ! .05,
one-tailed) for Asian American applicants in an Asian American plurality (M
p 74.06, SD p 9.35) compared to when such applicants applied to a white
plurality university (M p 66.88, SD p 10.78). No other group comparisons
Asian evaluators on class rank or test percentile or white evaluators on class
rank for Asian American applicants or test percentile for white applicants
were significantly different (means and t-statistics are available in app. B).
That white evaluators manipulated only class rank for white applicants and
test percentile for Asian American applicants will be addressed in the discussion section.
Comparing minimum academic standards between applicants, holding
plurality constant, also provided notable results. In an Asian American plurality condition, white evaluators lowered the class rank standard (t p 1.926,
P ! .05, Satterthwaites dfp 22.98) for the white applicant (M p 32.56, SD
p 19.55) compared to the Asian American applicant (M p 21.88, SD p
10.49). Also, white evaluators raised the test percentile standard for the Asian
American applicant (M p 74.06, SD p 9.35) compared to the white applicant
(M p 61.86, SD p 14.24). These results were consistent with the hypotheses.
However in a white plurality, white evaluators showed favoritism toward the
Asian American out-group by setting a lower class rank admission standard
(M p 28.43, SD p 16.60) for Asian Americans than the standard set (M p
17.81, SD p 10.91) for the in-group (t p 2.14, P ! .05). This finding was
unexpected and has implications for social identity theory (Tajfel and Turner
1979, 1986), as detailed in the Discussion section.
Analyses of admissions recommendations did not support hypothesis 2
Fig. 3.Minimum test percentile standard by applicant race and university demographics (white
evaluators).
(see fig. 4); they did however yield a notable, unanticipated finding. White
evaluators denied admissions recommendations to three out of 16 white
applicants in a white plurality, but did not deny admission to any of the white
applicants in an Asian American plurality; this difference was not significant
using Fishers Exact test for small sample sizes (however, see note 8). Nevertheless, the lack of denials of white applicants in an Asian American plurality
was still revealing (Asymmetric Sympathy, fig. 4). While white evaluators demonstrated sympathy for white applicants in an Asian American plurality, this
sympathy did not extend to Asian American applicants in a white plurality
(P p .051, Fishers Exact test). When Asian American applicants were at a
racial disadvantage in a white plurality university, four white evaluators still
denied admission recommendations to Asian American applicants.
Discussion
Operating in a society where dominant American ideologies give widespread legitimacy to fairness and meritocracy, it is not surprising that the
results of this study only confirm some of the hypotheses. A change in group
position does prompt members of the dominant group to selectively change
the standard of evaluation in order to slightly lower the bar for co-ethnics
on one criterion, while slightly raising the bar on a second criterion to obstruct a successful decision outcome for members of the ascending out-group.
The fact that white evaluators do not strongly defend their groups position by making biased admissions recommendations (i.e., denying many,
rather than a few, Asian American applicants) does not weaken group position
theorys predictions. Instead, it attests to the importance of organizational
Comparative Education Review
383
SAMSON
August 2013
385
SAMSON
also thinking about potential perceived group competition from either blacks,
Hispanics, or Asians.
The reason why white evaluators only manipulated class rank for white
applicants and only test percentile score for Asian American applicants requires theoretical justification. Why did certain academic criteria attach themselves to particular groups? The nature of the comparison criteria themselves
provides a possible explanation. Stephen Garcia and Avishalom Tor (2007)
argue that task performance, as measured by test scores, should be differentiated from scale scores, such as rank, which measure a general, underlying
skill. Because Asian American academic achievement is stereotypically attributed to test-taking proficiency (Goyette and Xie 1999), white evaluators utilizing this Asian American stereotype would reasonably manipulate standardized test percentile, which encompasses test-taking performance (e.g.,
answering multiple choice questions under timed pressure) rather than solely
measuring academic knowledge. On the other hand, the high academic
achievement of whites is not stereotypically attached to well-developed testtaking proficiency; thus, manipulating a particular task performance criterion
would have little bearing for white applicants. Instead, altering the class rank
standard, as a more general measure of underlying academic skills, would
prove more beneficial for a group (whites) whose academic success stereotypically is not partially attributed to test-taking task proficiency.
Some of the results of the current study are consistent with aspects of
social identity theory (Tajfel and Turner 1986) though some cannot be easily
assimilated into this long-standing approach to group categorization and
intergroup bias/behavior. Consistent with early formulations of social identity
theory (Tajfel and Turner 1979, 1986; Turner 1982), whites display in-group
favoritism when positive self-esteem has been threatened (the decline in white
enrollments in the Asian American plurality). However, social identity theory
not only proposed the pursuit of positive self-esteem as a behavioral motivation, but also the maintenance and enhancement of self-esteem (Tajfel and
Turner 1979, 40). The control condition in the present study, with a white
plurality and Asian American enrollments falling, would have been an optimal
condition under which white evaluators could have acted to maintain or
enhance positive self-esteem by favoring the dominant in-group. This was not
the case. Recall, however, that white evaluators favored Asian American applicants over white applicants on class rank in the white plurality. Yet despite
this Asian American out-group favoritism in minimum class rank standard,
white evaluators showed neither in-group nor out-group favoritism when making an admission recommendation in a white plurality condition (three white
applicants denied, four Asian American applicants denied). Thus, white ingroup favoritism is seen only in a setting where whites move downward in
group position, potentially qualifying the scope (in terms of social/institutional context) under which in-group bias operates ( Jost 2001). The possible
386
August 2013
387
SAMSON
benefiting from post-Civil Rights gains, appears to push against a socially and
historically crystallized sense of racial alienation.
Previous research outside the group position tradition can also explain
Asian evaluators lack of either an altered standard in evaluation or a bias in
admission recommendations. Asians might be unwilling to show bias on academic dimensions that partially constitute their own group stereotype, and
possibly, their own sense of selves (Kao 2000).
The present study has its share of limitations. First, the results are based
on a small sample size, which can potentially obscure differences that might
be statistically significant if the cell sizes were larger. Second, while the study
was based on random assignment to one of four experimental conditions,
the sample itself is not randomly selected nor is it representative of the United
States population as a whole. Rather, the study drew on students enrolled at
a highly selective university, students whose academic and nonacademic
achievements are generally higher than might typically be found in the US
population. While random assignment to experimental conditions allows us
to identify causal social psychological processes, such mechanisms may not
necessarily be found among student evaluators at less selective universities
or geographic areas in the United States where a change in white group
position is not a concern. Third, the current study was not designed to yield
a comparative assessment of competing theories accounting for differences
in group-based standards. While its results do point to insufficient theoretical
coverage in some of these alternative accounts (i.e., shifting standards model
and social identity theory), additional studies, designed specifically to identify
the scope conditions under which possibly competing explanations are relevant, will be necessary to determine which process is ultimately at work in
a given evaluation scenario.
Conclusion
Karabel (2005); Kalev et al. (2006); Lipson (2007); Denson and Chang (2009); Berrey (2011).
388
August 2013
Because grade point averages (GPA) are difficult to compare between college
preparatory institutions, the University finds that class rank provides a better indicator of the applicants academic achievements relative to her or his classmates.
What MINUIMUM CLASS RANK would the applicant need to exceed in order
for you to recommend that she or he be admitted to the University?
The applicant would need to be in the Top XX% of their high schools senior
class.
Minimum Standardized Test Percentile
Susan Sturm and Lani Guinier (1996) provide a critical analysis of selection processes and the
assumptions they create about affirmative action.
Comparative Education Review
389
SAMSON
MINIMUM PERCENTILE RANK should the applicant score on the SAT or ACT
in order for you to recommend that she or he be admitted to the University? The
applicant would have to possess test scores better than this percentage of all other
test takers.
This candidate would need to score above the XX percentile on his or her
standardized college entrance exam.
Admissions Recommendation
Based on the applicants academic profile that you have reviewed, would you
recommend admitting this applicant?
Please keep in mind that your evaluation of this applicants academic profile
may determine whether the applicant is admitted to the top public university in
the state, and widely recognized as one of the best universities in the country. This
decision may deeply affect the applicants future employment and life chances.
Clearly and carefully select whether you would recommend admitting the applicant.
No
Yes
390
August 2013
Appendix B
TABLE B1
Minimum Academic Standards (Means) for Recommending University Admissions by Race of
Evaluators, Race of Applicant, and University Demographic Context
N
White evaluators:
White applicant:
White plurality university
Asian American plurality
university
t-statistic
df
Asian American applicant:
White plurality university:
Asian American plurality
university
t-statistic
df
Asian evaluators:
White applicant:
White plurality university
Asian American plurality
university
t-statistic
df
Aian American applicant:
White plurality university:
Asian American plurality
university
t-statistic
df
16
16
Class Rank
17.81
32.56
2.63**
23.51a
SD
Standardized
Test
Percentile
SD
10.91
67.81
16.73
19.56
61.88
1.08
30
14.24
16
28.44
16.61
66.88
10.78
16
21.88
1.34
30
10.49
74.06
2.01*
30
9.35
16
28.75
22.02
66.56
20.23
16
23.94
.71
30
15.86
63.88
.35
30
22.93
16
22.19
12.78
67.81
14.02
16
17.81
.97
30
12.78
69.38
.25
30
20.76
391
SAMSON
Appendix C
References
Almaguer, Tomas. 1994. Racial Fault Lines: The Historical Origins of White Supremacy
in California. Berkeley: University of California Press.
Aronson, Joshua, Michael J. Lustina, Catherine Good, Kelli Keough, Claude M.
Steele, and Joseph Brown. 1999. When White Men Cant Do Math: Necessary
392
August 2013
393
SAMSON
August 2013
395
SAMSON
Social Identity and Intergroup Relations, ed. Henri Tajfel. New York: Cambridge
University Press.
Uhlmann, Eric L., and Geoffrey L. Cohen. 2005. Constructed Criteria: Redefining
Merit to Justify Discrimination. Psychological Science 16 (6): 47480.
Unzueta, Miguel M., Brian S. Lowery, and Eric D. Knowles. 2008. How Believing
in Affirmative Action Quotas Protects White Mens Self-Esteem. Organizational
Behavior and Human Decision Processes 105 (1): 113.
Wang, L. Ling-Chi. 1988. Meritocracy and Diversity in Higher Education: Discrimination against Asian Americans in the Post-Bakke Era. Urban Review 20 (3):
189209.
Wong, Paul, Chienping Faith Lai, Richard Nagasawa, and Tieming Lin. 1998. Asian
Americans as a Model Minority: Self-Perceptions and Perceptions by Other Racial
Groups. Sociological Perspectives 41 (1): 95118.
396
August 2013