INTRODUCTION
In the domain of higher learning, the principle of meritocracy and indicators such as grade point average and standardized test scores used to measure an individual's merit, sit at the center of a competitive process in which the ultimate reward is the scarce and valuable resource of a high quality college education (Alon and Tienda, Reference Alon and Tienda2007). Historically, the content and the importance of particular indicators of merit have changed over time (Tsay et al., Reference Tsay, Lamont, Abbott and Guetzkow2003). In a context of persistent racial inequality, the commitment to meritocracy as a guiding principle remains a significant challenge at the dawn of the twenty-first century (Castilla Reference Castilla2008). Research has also revealed that one reason for changes in the indicators used to measure individual merit is the function of merit as a gate keeping mechanism to preserve the status of the wealthy and the powerful, particularly that of White elites, against the incursion of ethnoracial outgroups, principally Jews in the early half of the twentieth century (Karabel Reference Karabel2005).
Since university admissions criteria in the twentieth century have already been implicated as part of processes of racialized group inclusion/exclusion, it is useful to analyze attitudes about public university admissions criteria as public policy attitudes, subject to the same biases, values, and mechanisms described in the scholarly debates on racialized politics and public opinion (Sears et al., Reference Sears, Sidanius and Bobo2000). Viewed through this lens, attitudes toward American meritocracy, as measured by the importance attributed to public university admissions criteria, should be subject to the same concepts, ideologies, and attitudes that shape policy attitudes in other domains: educational equity, social welfare, and housing policies, etc. (Bobo Reference Bobo1983; Bobo and Smith, Reference Bobo, Smith, Danziger, Sandefur and Weinberg1994; Bobo and Zubrinsky, Reference Bobo and Zubrinsky1996; Charles Reference Charles2006; Huddy and Sears, Reference Huddy and Sears1995; Krysan Reference Krysan2000; Sears et al., Reference Sears, Sidanius and Bobo2000; Sniderman and Carmines, Reference Sniderman and Carmines1997; Sniderman and Piazza, Reference Sniderman and Piazza1993).
The present study aims to address the following questions: As the White populace in the United States moves toward numerical minority status by 2042, the political consequences of which have already been observed for minorities living in minority-majority localities (Barreto et al., Reference Barreto, Segura and Woods2004), how might Whites respond to impending threat of losing their dominant ethnoracial group position? In particular, how will Whites react in the domain of higher education, especially at selective, elite institutions, where Asians are increasingly prominent and other non-Whites are maintaining or capturing a larger share of enrollments? I rely upon survey-based experiments (Schuman and Bobo, Reference Schuman and Bobo1988) to assess the role played by group threat, a central dimension of group position theory (Blumer Reference Blumer1958; Bobo Reference Bobo1999), in shaping the importance grade point average should have as a factor for public university admissions.Footnote 2 In the celebrated “postracial” era marked by the election of Barack Obama as President of the United States, racial group threat should have little bearing on White commitment to academic meritocracy or the importance of merit-based academic criteria. Alternatively, race might still matter if Whites continue to react to racial and ethnic threat to their dominant group position by redefining merit, as they did in response to Jewish threat during the early twentieth century. However, unlike the historical case of Jewish threat, such a reaction today may be tempered by multiple group threat from different directions, from groups stereotypically viewed as either superior or inferior to Whites on various dimensions (Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002).
THEORY AND HYPOTHESES
Beliefs about inequality have often been tied to the existence of a “dominant ideology” in the United States that reproduces itself by legitimating the actions of individuals who have succeeded in a social mobility structure (Huber and Form, Reference Huber and Form1973; Kluegel and Smith, Reference Kluegel and Smith1983). Kluegel and Smith described this dominant ideology as consisting of three components: a belief in the widespread availability of economic opportunity, a belief in the individual locus of control over personal success and failure in this open economic system, and a belief that the system is therefore fair and equitable. Kluegel and Smith, drawing from their data analysis, state, “Individualism is a central aspect of the American cultural pattern (the dominant ideology) and is held to a major extent across social strata. There is not much variation, even along such powerful structural divisions as race or status, in perceptions of the importance of individual causes of achievement” (p. 93). Recent research in the stratification beliefs tradition point to the continuing importance of individual-centered accounts for stratification outcomes, though with some group variation (Hunt Reference Hunt1996, Reference Hunt2004, Reference Hunt2007).
Social psychologists and political psychologists studying racialized politics have situated these ideologies of inequality within contemporary conflicts over race-related policies as well as tensions emerging from group competition over scarce and valued resources. These works analyze the ideational mechanisms that influence inter-group attitudes, policy preferences, and behaviors by offering explanations for individual attitudes in concrete settings (e.g., conflicts over school desegregation busing plans, local electoral campaigns, etc.) and introducing explanatory variables (e.g., prejudice, value commitments, and psychological constructs) beyond traditional demographic and social background indicators. These theoretical approaches have largely fallen into two camps: group-centered and values-centered approaches. By studying White commitment to meritocracy in a multi-racial context where Whites' dominant group position is potentially being threatened by different groups, this paper provides an opportunity to revisit the racialized politics debate by focusing on two of its theoretical exemplars: group position theory and principled politics.
Group Position Theory
Herbert Blumer (Reference Blumer1958) proposed a theory that situates ethnoracial group hierarchy, and the attitudes of the dominant racial group, at the center of intergroup attitudes and relations. Blumer sought to focus attention on how individuals of a racial in-group collectively and publicly characterize other racial groups in toto or members of such groups, thereby defining the individual's own racial group in opposition to characterizations of the outgroups. The social process by which group identities are collectively defined vis-à-vis other groups, Blumer argues, positions groups in relation to each other. Scholars have since used group position theory to explain the connection between group threat and contemporary racial attitudes (Almaguer Reference Almaguer1994; Quillian Reference Quillian1995; Smith Reference Smith1981). In the domain of education, group position theory partially explains White opposition to policies such as busing, affirmative action, and bilingual education (Bobo Reference Bobo1983, Reference Bobo, Sears, Sidanius and Bobo2000; Huddy and Sears, Reference Huddy and Sears1995). Group position theory holds that perceived group competition over scarce and valued resources, such as education, can trigger hostility towards educational policies that threaten the greater share of this valued resource held by the dominant group by virtue of its position in the group hierarchy. To the extent that Asian over-representation in a university system is perceived as a group threat to Whites' dominant group position in the domain of higher education, and Asian college success is stereotypically associated with high academic performance, I hypothesize that:
H1: Whites who are exposed to an Asian over-representation context as an indicator of Asian group threat will decrease the importance that grade point average should have compared to other Whites.
Perceived zero-sum ethnoracial group competition has already been shown to influence White attitudes towards affirmative action (Bobo Reference Bobo, Sears, Sidanius and Bobo2000). As an indicator of group threat, perceived zero-sum Asian group competition should parallel the effects of the experimentally manipulated Asian over-representation group threat cue, once again drawing upon the stereotypically high academic performance of Asians (Brand Reference Brand1987) to provoke a negative reaction against academic criteria.
H2: As their perceptions of zero-sum Asian group competition increases, Whites will decrease the importance that grade point average should have.
However, since Blacks and Hispanics are not stereotypically associated with high academic performance, group position theory would predict a different outcome than that resulting from Asian group threat. Because Whites stereotypically perceive Blacks and Hispanics to be intellectually inferior to Whites (Bobo and Johnson, Reference Bobo, Johnson, Bobo, Oliver, Johnson and Valenzuela2000), White respondents might increase the importance that grade point average should have, precisely to give Whites an advantage in university admissions in response to the perceived group threat that Blacks and Hispanics represent.
H3: As their perceptions of zero-sum Black or Hispanic group competition increases, Whites will increase the importance that grade point average should have.
Principled Politics
The principled politics approach eschews group-based accounts for White opposition to race-based policies (Sniderman and Carmines, Reference Sniderman and Carmines1997; Sniderman and Piazza, Reference Sniderman and Piazza1993). Adherents to this perspective argued that beliefs about government action—whether or not government should intervene in a given set of issues—rather than fundamentally group-centered theories (such as group position), explains White opposition to race-conscious policies and programs. Additionally, the opposition-as-politics advocates forward explanations that highlight a commitment to the abstract value of individualism, equality, and fairness in order to explain White opposition to controversial policies such as affirmative action.
However, empirical research exploring the relationship between individualism and policy attitudes shows that individualism has no effect on policies related to race, but does influence race neutral policy attitudes such as funding for food stamps and unemployment assistance (Kinder and Mendelberg, Reference Kinder, Mendelberg, Sears, Sidanius and Bobo2000). Drawing upon the work of Mary Jackman (Jackman Reference Jackman1994; Jackman and Muha, Reference Jackman and Muha1984), Kinder and Mendelberg reason that if Whites draw upon individualism to oppose race-related policies, they are committed to a notion of “political individualism,” in which “government, under political individualism, should be limited, its purpose confined to enabling individuals' wants to be satisfied, individuals' interests to be pursued, and individuals' rights to be protected” (p. 59). If individualism is instead measured by items that capture the belief in self-reliance (Gamson and Modigliani, Reference Gamson, Modigliani, Braungart and Braungart1987) or that hard work will be rewarded, rather than protecting the rights of individuals, this work-ethic form of individualism, according to Kinder and Mendelberg, would not be related to race-related policy attitudes, as the empirical research they review suggests.
In the case of university admissions criteria, where merit is closely tied to notions of individual effort, the abstract value of individualism may yet play a significant role in explaining attitudes towards a public policy like public university admissions criteria, despite its inability to explain attitudes towards welfare, affirmative action, etc. The trope of individual effort and dependence on “the same standard” (Feagin and O'Brien, Reference Feagin and O'Brien2003, p. 105) of merit evaluation cuts through much of the discourse around how Whites believe scarce opportunities should be allocated (see also Bonilla-Silva Reference Bonilla-Silva2001, Reference Bonilla-Silva2003). Therefore, because the concept of individual merit usually refers to rewarding an individual's demonstrated effort and ability, a respondent's commitment to individualism should vary positively with their rating of an individual-oriented admissions criterion such as grade point average.
H4: As Whites' belief in individualism increases, the importance that grade point average should have as an admissions criterion, increases.
Redefinition Redux (Corollary Hypotheses)
The “Big Three” private universities (Harvard, Yale, and Princeton) redefined admissions merit criteria during the first third of the twentieth century partially in response to the “Jewish invasion” (Karabel Reference Karabel2005). They sought to avoid the fate that befell Columbia, where the “proportion of Jews … had reached perhaps forty percent” (p. 87), the same share of the University of California (UC) undergraduate population that Asian Americans hold in the twenty-first century. Believing that Jewish students were primarily more successful in academic endeavors than their Anglo Saxon counterparts, administrators redefined merit to de-emphasize academic proficiency as the sole admissions criteria and highlighted nonacademic factors such as athletic prowess, leadership, and personal character thought to favor the Anglo-Saxon elite.
The large presence of Asian Americans at the University of California, and not just perceived Asian competition, could prompt White Californians to more forcefully reconsider the importance of academic merit in a repeat of history. I therefore propose corollary formulations of Hypotheses 2–4 above, to take into account possible interactions between the hypothesized effects of attitudes and the Asian over-representation context. If information that Asians represented a substantial share of the UC student body confirmed perceptions of Asian group competition, this might cause Whites to further decrease the importance of grade point average as a function of perceived Asian group competition compared to the absence of confirmatory information.
H2c: A context of Asian over-representation will further increase the hypothesized negative effect of Whites' perceptions of zero-sum Asian group competition on the importance that grade point average should have.
However, on the Black and Hispanic ballots, exposure to the experimental Asian threat frame could force White respondents to consider admissions factors in a cognitive environment where two ethnoracial outgroups have been made salient as competitors for the remaining university spots unoccupied by Asians. While perceived Black or Hispanic group competition might relate to an increased importance for grade point average in the absence of any information about the UC's current makeup, information about Asian over-representation might temper this increase.
H3c: A context of Asian over-representation will moderate to non-significance the hypothesized positive effect of Whites' perceptions of zero-sum Black or Hispanic group competition on the importance that grade point average should have.
Prior research indicates that Asians are stereotyped as a competent group worthy of envy (Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002; Lin et al., Reference Lin, Kwan, Cheung and Fiske2005). Because Asians as a model minority are perceived as a hard working group, the positive effect of individualist beliefs on the importance of grade point average may change in a context where demonstrated Asian success potentially limits the number of remaining scarce and desirable educational opportunities. That is, individualist Whites might no longer view grade point average so favorably if another hard working group had attained success that potentially challenged Whites' dominant group position.
H4c: A context of Asian over-representation will moderate to non-significance the hypothesized positive effect of Whites' individualist beliefs on the importance that grade point average should have as an admissions criterion.
DATA AND METHOD
The survey data come from the 2007 Golden Bear Omnibus (GBO) conducted by the University of California Survey Research Center. The GBO data were collected through a random-digit dialed, computer assisted telephone interview, with a total sample size of 993 respondents. The response rate (ratio of interviews to eligible households) was 16%.Footnote 3 Regression models were estimated using post-stratification weights to minimize the effect of possible household nonresponse bias (Groves Reference Groves2006; Keeter et al., Reference Keeter, Kennedy, Dimock, Best and Craighill2006). A more in-depth discussion about the response rate and post-stratification weighting can be found in Appendix A.
Survey respondents are adults, age twenty-one and over, residing in California. Interviews were conducted in English and Spanish, according to the respondent's preference. As an analysis of the dominant group's perspective, the current study selects for White (non-Hispanic) respondents, who make up 606 of the sample's respondents. Due to missing data on a number of independent variables, the valid N prior to multiple imputation of missing data was 440 White respondents. T-tests showed significantly different responses on the dependent variable between cases with and without missing values, indicating that the data were not missing at random. In order to address this issue, multiple imputations were conducted on the missing values (see Appendix A for more information). The results presented below are based on the sample with missing values addressed through multiple imputations. For comparative purposes, Appendix A also contains the complete cases models, with incomplete cases handled through listwise deletion. Recovery of cases through multiple imputations yielded a valid N of 599 White respondents.
Independent Variables (Group Competition Threat)
I employ a survey-based experiment with a split-ballot design (Schuman and Bobo, Reference Schuman and Bobo1988) to explore beliefs about zero-sum group competition in a multiracial context (Bobo and Hutchings, Reference Bobo and Hutchings1996). Respondents are randomly assigned to one of three ballots that first ask respondents to consider competitive group threat from one of the following ethnoracial groups: Blacks, Hispanics, or Asians. I constructed a scale relying upon these beliefs about zero-sum group competition as my first indicator of group threat (Bobo Reference Bobo, Sears, Sidanius and Bobo2000; Bobo and Hutchings, Reference Bobo and Hutchings1996). The scale is an average of four items (see Appendix B) measuring beliefs about zero-sum group competition for jobs, political influence, quality housing and good neighborhoods, and economic position. Responses to the items are measured on a four-point scale ranging from strongly disagree to strongly agree, with no option for a neutral midpoint. The scale ranges from 1 to 4 and reliability alpha varies by racialized target group: .79 for Blacks, .72 for Hispanics, and .80 for Asians. The survey instrument presented these perceived group competition questions prior to presenting the dependent variable item, which for some respondents also included a second experimental manipulation (see Figure 1 for question order and order of exposure to experimental ballots).
Fig. 1. Question Order and Order of Exposure to Experimental Ballots
Independent Variable (Asian Plurality Threat) and Dependent Variable
White respondents have already considered the potential group competition posed by one of three outgroups (Blacks, Hispanics, and Asians) when some of the respondents randomly encounter a second experimental manipulation that serves as an additional indicator of group threat: half of the survey respondents are exposed to an experimental frame that specifies the actual proportional share of Asians at the University of California. Under the control frame, the dependent variable (the importance of UC admission factors) is introduced using the following contextual narrative:
In the tax-payer supported University of California system, high school grade point average and standardized test scores should be the main factors in student admissions. Some people think that leadership and community service should be given substantial weight as well.
On a scale of 0–10 where 0 means not important and 10 means extremely important, please rate each of the following factors for University of California admission decisions. You may choose any number between 0 and 10 to answer.
First, on the 0 to 10 scale, how unimportant or important should GRADE POINT AVERAGE be as a factor for UC admission?Footnote 4
Because prior research has indicated that the population share of a racialized outgroup can serve as an indicator of group threat (Blalock Reference Blalock1967; Quillian Reference Quillian1995, Reference Quillian1996; Taylor Reference Taylor1998), including a more recent social psychological study that used a high percentage of Asians as an indicator of group threat (Maddux et al., Reference Maddux, Galinsky, Cuddy and Polifroni2008), I deploy an experimental frame that utilizes a high proportion of Asians to make Asian group threat salient.Footnote 5 In this situation, half of all survey respondents received the same contextual narrative above (see Figure 1), which is immediately preceded and framed by the following experimental cue:
Under current admissions procedures in the University of California system, Asians make up almost 40% of the student body (or 2 out of every 5 students) while they are only 12% of the California population.
This framing cue on the randomly assigned experimental survey ballot therefore cues the high percentage of a racial outgroup, an indicator of group threat utilized in previous social scientific studies.
Independent Variables (Individualism)
To measure individualism, I construct a scale by averaging the responses of six individualism items originating from the 1986 National Election Studies survey, reverse-coding a number of items as necessary (see Appendix B). These items were specifically designed to tap into “economic individualism,” or “the belief that people should get ahead on their own through hard work” (Feldman Reference Feldman1988, p. 419), rather than “political individualism” or the protection of individual rights from the government (Kinder and Mendelberg, Reference Kinder, Mendelberg, Sears, Sidanius and Bobo2000). Cronbach alpha reliability is .73 for the individualism scale and the scale ranges from 1 to 4.
Controls
To isolate the effects of group threat, I control for a survey instrument-related factor as well as individual-level characteristics in the estimation models. In terms of survey-related factors, I controlled for an earlier set of questions on the survey omnibus regarding race and crime. The data collected for this study were part of a larger collection of data for multiple studies, with separate study modules rotated randomly among respondents. Because at least one module would always be omitted at random, no respondents were exposed to all available modules. One module in particular that preceded the module for the present study in terms of data collection, a study on race, crime, and punitive attitudes, may have affected the results of the current study by already exposing respondents to questions pertaining to race prior to the present module. To control for the effects of earlier exposure to questions on race and crime, a dummy-coded control variable was introduced into all the models with the value 1 representing respondents who were exposed to the race/crime module, and 0 for those who skipped the race/crime module.
For individual-level demographic and social background characteristics, I introduce controls for age, education, income, gender, and nativity. The age variable is calculated by subtracting the respondent's year of birth from the year 2007, the year in which the data were collected. Income is a continuous variable imputed using the midpoint value of twenty-four ordered categories for household income (e.g., respondents in the $35,000–$50,000 category were assigned a midpoint income of $42,500). Education is a scale variable representing highest grade of school (or year of college) completed, ranging from 0 to 18 (with possession of an undergraduate degree coded as 16). The gender variable is a dummy variable coded 0 for male, 1 for female. Nativity is coded as a dummy variable, with 1 indicating respondents born outside of the United States.
I also introduce a social psychological factor to control for another concept in the principled politics tradition: political ideology. Political ideology is important in shaping individuals' attitudes about the role that government intervention should play in ensuring equal opportunity (Feldman and Huddy, Reference Feldman and Huddy2005; Sniderman and Carmines, Reference Sniderman and Carmines1997; Sniderman and Piazza, Reference Sniderman and Piazza1993). Because public university admissions are a matter of public policy, beliefs about the proper role of government in regards to equal opportunity should bear on attitudes toward public university admissions criteria. To control for political conservatism, I construct a scale from four different questions that ask respondents to identify themselves politically as liberal, conservative, or moderate (or haven't thought about it). Respondents who identified as either liberal or conservative were asked a further question about strength of ideology (e.g., strong Liberal or not a very strong Liberal). Those who identify as moderates, who haven't thought about it, or who claimed neither labels were then asked if they considered themselves more like liberals, more like conservatives, or neither. Those who responded with neither were given a midpoint value (4) on a seven-point political conservatism scale constructed for this study, while more like liberal and more like conservatives were each assigned a value of 1 point away from this midpoint value.
Interactions with Social Psychological Attitudes
Corollary Hypotheses H2c, H3c, and H4c predict possible changes in the impact of perceived group competition and individualism in a context where White dominant group position has been threatened. To control for these possible changes, I interact perceived group competition and individualism with the experimental Asian over-representation ballot and include these interaction terms in the estimation models.
Method
This study uses weighted Ordinary Least Squares regression models, using post-stratification weights, to estimate respondents' ratings of the importance that grade point average should have as a public university admissions criterion. Separate models were estimated for each of the racial target group ballots. For example, a model was estimated for GPA among those who were exposed to Blacks as the target group for group threat questions, a separate model for Hispanics as the target group, and a separate model for respondents exposed to Asians as the target group on the group-related social psychology items.
RESULTS
Table 1 provides descriptive characteristics of the sample, weighted sample means for the social psychological scales, and dependent variable means. Neither the mean scores on the social psychological scales nor the overall mean attitudes on the dependent variable vary by group under consideration. Interestingly, the mean difference on the dependent variables between control and experimental ballots indicate an increase in the importance that grade point average should have on the Asian group competition ballot and a decrease in such importance on the Black group competition ballot, contrary to the hypotheses. Neither difference is statistically significant. Possible explanations for these seemingly contradictory findings are explored below.
Table 1. Weighted Sample Characteristics and Descriptive Statistics (Non-Hispanic Whites)
Data Source: University of California Survey Research Center Golden Bear Omnibus (2007)
Notes: Standard errors of weighted means in parentheses; some data missing for age; multiple imputation used to address missing values on social psychological indicators.
a Sample means
Perceived Asian Group Competition in a Context of Asian Over-representation
Table 2 presents the results of the OLS regression models estimating the importance that grade point average should have as a factor for University of California admissions when only Asians have been primed as a group threat. Initial models suggest, at first glance, that the data do not support the predicted hypotheses regarding Asian threat (H1 and H2). The main effect of the Asian plurality cue and the main effect of beliefs about Asian zero-sum competition are not significant and have opposite signs than theoretically predicted. However, including the interaction between Asian group competition and the Asian over-representation cue reveals that the data do in fact support Hypothesis H1 and move in the direction specified in Hypothesis H2.
Table 2. Multiple Linear Regression of the Importance of GPA as a University Admissions Criterion (Asian Group Competition Primed)
Data Source: University of California Survey Research Center Golden Bear Omnibus (2007)
Standard errors in parentheses; multiple imputation used to address missing values (31 cases recovered); see Appendix A for complete cases model.
*** p < 0.001,
** p < 0.01,
* p < 0.05, +p < 0.10 (one-tailed tests)
The Asian over-representation cue's anticipated effect was suppressed because the perceived group competition coefficients bear opposite signs depending on exposure to the Asian over-representation cue. When they are not exposed to Asian over-representation, White respondents predictably decrease the importance that grade point average should have as their belief in zero-sum group competition with Asians increases, though the effect is not significantly significant. After controlling for the interaction between perceived group competition and Asian over-representation, exposure to the Asian over-representation cue does cause White respondents to decrease the importance that grade point average should have, providing support for Hypothesis H1. When White respondents are primed with the Asian over-representation cue, the effect of Asian group competition becomes positive and significant (p < 0.05), disguising the main negative effect of the Asian over-representation cue. The data do not support corollary Hypothesis H2c.
A second notable finding from the Asian group competition ballot in Table 2 is individualism's lack of effect on the importance that grade point average should have. This result runs counter to individualist principles. The significant interaction term would seem to support Hypothesis H4, but the positive effect is not significant (B = 0.365, p = 0.17). Individualism's effect is indeed moderated to non-significance, as predicted in the corollary Hypothesis H4c. However, this initial effect is negative, not the positive effect derived from the principled politics perspective. The data therefore do not support Hypotheses H4 or H4c.
Figures 2 and 3 illustrate the different effects of perceived Asian group competition as a function of exposure to the Asian over-representation ballot, based on the coefficients in the final model and using group means for the control variables. Because there is a second interaction effect with individualism, Figure 2 illustrates the effects of the perceived Asian group competition interaction for those who “somewhat disagree” with an individualist work ethic (approximately 9% of the sample report individualism attitudes that range from strongly disagree (1) to somewhat disagree (2) responses). As Figure 2 reveals, White respondents that express low perceptions of Asian group competition (about 75% of respondents) are most affected by the Asian over-representation cue as hypothesized. Among those who somewhat agree with perceived Asian group competition (about 8% of respondents), priming Asian over-representation produces the opposite effect. Asian over-representation causes White respondents to more strongly endorse meritocracy when they feel that Asians pose a group threat and they recognize that Asians are already over-represented. This unanticipated and contradictory finding is even more closely revealed in Figure 3, when we set individualism at its mean value (2.82) for the sample. Because most of the sample professes low levels of perceived Asian group competition, Figure 3 indicates those who have an average level of belief in the value of individual work ethic are largely unmoved by the introduction of the Asian over-representation prime. Figure 3 also shows that Asian over-representation prompts a greater endorsement of grade point average for public university admissions among Whites who somewhat agree with the value of individual hard work and are thinking solely about Asians as a group competitor. The results of the next group competition ballot, however, further complicate the findings.
Fig. 2. Perceived Zero-Sum Group Competition with Asians
Fig. 3. Perceived Zero-Sum Group Competition with Asians
Perceived Black Group Competition in a Context of Asian Over-representation
White attitude towards grade point average differs substantially once Whites are primed with potential Black group competition rather than solely Asian group threat (see Table 3). First, it is worth noting that perceived Black group competition is positively associated with the importance that grade point average should have as a factor for public university admissions when respondents are not primed by the Asian over-representation cue. This provides evidence in support of Hypothesis H3 and reveals that the effect of perceived zero-sum group competition varies depending on which group is under consideration. However, supporting corollary hypothesis H3c, the interaction term shifts the sign of the Black competition coefficient, moderating the Black group competition slope in the context of Asian threat to statistical non-significance. The impulse to increase grade point average among Whites that perceive zero-sum Black group competition is muted when these Whites are informed of Asian over-representation.
Table 3. Multiple Linear Regression of the Importance of GPA as a University Admissions Criterion (Black Group Competition Primed)
Data Source: University of California Survey Research Center Golden Bear Omnibus (2007)
Standard errors in parentheses; multiple imputation used to address missing values (35 cases recovered); see Appendix A for complete cases model.
*** p < 0.001,
** p < 0.01,
* p < 0.05, +p < 0.10 (one-tailed tests)
Whites also react differently to the Asian plurality threat cue when considering Black group competition, indicating a positive and significant effect on the importance that grade point average should have. Whites experience the countervailing forces of a decreased emphasis on grade point average in response to Asian threat and the increased emphasis on grade point average in response to Black competition. Despite the countervailing forces, perceived Black competition is sufficiently strong enough to pull White attitude towards emphasizing the importance that grade point average should have.
To more clearly illustrate the differential effect of racial group priming, Figure 4 uses Table 3's final model coefficients, setting covariates at group means, to show the effects of perceived Black group competition both with and without the priming Asian over-representation. As hypothesized in Hypothesis H3, higher levels of perceived Black group competition is related to higher levels of importance attributed to grade point average as a public university admissions criterion for Whites who have not been primed with Asian over-representation.Footnote 6 For the slightly less than a third of the White sample who strongly disagree that there is group competition between Blacks and Whites, the Asian over-representation cue also causes these Whites to increase the importance of grade point average, mirroring the hypothesized effect of Black group competition without the Asian over-representation priming (Hypothesis H3). This finding suggests that White concerns over Black group threat reconfigure the meaning of the Asian over-representation cue. However, among the third of the sample that somewhat disagrees to somewhat agrees with perceived Black group competition, exposure to the Asian over-representation cue counteracts their inclination to increase the importance they attribute to grade point average, contrary to the comparable result for Asian group competition in the context of Asian over-representation found in Table 2.
Fig. 4. Perceived Zero-Sum Group Competition with Blacks
Perceived Hispanic Group Competition in a Context of Asian Over-representation
The Hispanic group competition models (Table 4) provide a couple of interesting results. First, while not a predictor variable, it is the only group for which political ideology factors into the importance that White Californians believe grade point average should have as a UC admissions criterion. Moreover, unlike the Asian and Black group competition ballots, there is no significant effect from the indicator of beliefs about Hispanic zero-sum group competition. This anomalous result is likely because strong stereotypes about Hispanic academic performance as a group have not yet fully crystallized. However, similar to the Black group competition ballot, the Asian over-representation cue does have a significantly positive effect on the dependent variable in the final model. As with Blacks, when White respondents are thinking about Hispanic group competition, the reality of Asian over-representation causes Whites to increase the importance of grade point average, overcoming their inclination to decrease the importance of grade point average when they are thinking about Asians alone. Figure 5 illustrates this effect.
Table 4. Multiple Linear Regression of the Importance of GPA as a University Admissions Criterion (Hispanic Group Competition Primed)
Data Source: University of California Survey Research Center Golden Bear Omnibus (2007)
Standard errors in parentheses; Multiple Imputation used to address missing values (42 cases recovered); see Appendix A for complete cases model.
*** p < 0.001,
** p < 0.01,
* p < 0.05, +p < 0.10 (one-tailed tests)
Fig. 5. Perceived Zero-Sum Group Competition with Hispanics
DISCUSSION
Taken together, the results reveal that the meritocratic standard, in this case as measured by grade point average as a public university admissions criterion, is affected by the threat that ethnoracial outgroups pose to Whites. Moreover, among White Californians who perceive low levels of Black group competition, Black threat is more influential than Asian threat, as their response to the Asian cue runs opposite to those of Whites who are thinking solely about Asians. This Black over Asian hierarchy confirms findings from an earlier survey-based experiment on White racial attitudes (Schuman and Bobo, Reference Schuman and Bobo1988). For some Whites in the twenty-first century, “the Black image in the White mind,” the object of study by the late historian George M. Fredrickson (Reference Fredrickson1987), may still hold a unique and particularly threatening significance compared to the image of other non-White ethnoracial groups.
If grade point average is understood simply as an indicator of an individual's work ethic or average academic achievement over a period of three or more years of high school coursework, the importance that grade point average should have as an admissions criteria should not vary based on the racial makeup of the university or perceived group competition. The survey-based experiments reveal, however, that commitment to conventional academic meritocracy does vary in response to large Asian enrollments at a prestigious, public university and perceived group competition from Asians and Blacks.
The debate on a race-neutral, values-based approach versus various group interest-based approaches to explaining racialized politics has at times been heated (Sears et al., Reference Sears, Sidanius and Bobo2000; Sniderman and Carmines, Reference Sniderman and Carmines1997). To be sure, while the data did not support the influence of individualism for the importance of grade point average as predicted in Hypothesis H4, individualism may yet play a role on other admissions criteria, albeit a role that might differ depending on which ethnoracial group Whites are considering.Footnote 7 Also, another important factor in the principled politics approach, though used as a control in the present study, political ideology did influence the importance attributed to grade point average when Whites were thinking about Hispanic group competition (with marginal significance on the Asian group competition ballot).
An alternative to the group threat line of interpretation is that Whites are motivated by a notion of fairness based on distributional outcomes, and the over-representation of Asians at the university violates this notion of distribution relative to Asians' representation in the state. If fairness is understood as procedural fairness, in which procedures are set and universally applicable to all, this idea of procedural fairness crumbles in the face of malleability in the importance that grade point average should have in response to the survey's experimental cue and the various social psychological indicators that differ by race of target group. If fairness is understood, rather, as an outcome-based fairness, then one might argue that the large over-representation of Asians is seen as unfair according to a distributional standard, prompting a White response to alter the admissions criteria to restore a distributional outcome in which the number of Asians at a public university is substantially decreased. However, it is worth noting, that the effect of the experimental cue in the survey data has a different valence between the Black/Hispanic and Asian reference group ballots. Also, the perceived Black group competition indicator in the control condition is positive and significant. These findings suggest that a fairness-as-outcome-distribution argument is not applicable under these circumstances.
The experimental design also produced unanticipated results, clearly indicating the real need to embed attitudinal research in a multiracial context, and yielding potential opportunities for future research to theoretically account for the unsuspected findings. Further examination of the data provided some insight into the nature of the surprising results.
The first unanticipated finding was the positive effect of Asian group competition on the importance that GPA should have for UC admission when Asian over-representation has been primed. A closer look at the components of the group threat variable indicated that most of this unforeseen effect is driven by those who somewhat or strongly agree with perceived economic competition against Asians (Many Asians have been trying to get ahead economically at the expense of people like me). Closer scrutiny of the open-ended follow-up question (Would you please tell me why you feel this way?) revealed a possible explanation for the unexpected result. Consider the following remarks from survey respondents, differentiated by their perceptions of Asian economic competition and their rating of GPA's importance as an admissions criterion:
Some interactions with Asians have been positive, but some have been extremely rude, pushy, and they own everything.
— Respondent somewhat agrees (3) with perceived Asian economic competition and rates GPA as extremely important (10) under Asian over-representation
I've heard that they get low loans to start business and they get help because they are Oriental and their kids can get grants where if you're White you can't.
— Respondent somewhat agrees (3) with perceived Asian economic competition and rates GPA as almost extremely important (9) under Asian over-representation
Because they want to always blame their racial background, you do what you need to do to get ahead.
— Respondent strongly agrees (4) with perceived Asian economic competition and rates GPA as extremely important (10) under Asian over-representation
I think sometimes, that they, and I'm not against them having such programs, but I, being born here, should have the same rights. In my neighborhood, an Asian man bought up land on three corners of the street. He was able to do this because of government assistance.
— Respondent strongly agrees (4) with perceived Asian economic competition and rates GPA as extremely important (10) under Asian over-representation
Yeah, because they come over here, and go on welfare, and go for it. We brought them over here, but that doesn't mean they don't have the right to work.
— Respondent somewhat agrees (3) with perceived Asian economic competition and rates GPA as almost extremely important (9) under Asian over-representation
Each of these remarks captures resentment towards Asians, or feeling that Asians may not be playing by the meritocratic rules that others are following. These White respondents thus likely reinforce traditional academic meritocracy as a counter-reaction to the violations of meritocracy they see coming from Asians, a violation which has allowed Asians to capture the plurality of enrollments at the UC system as indicated by the experimental cue. When faced with the presumed illegitimate fruits of Asian success, these White respondents answer with a call for greater legitimacy in the process of university admissions, bolstering the importance of grade point average as a conventional meritocratic criterion.
The second unanticipated finding was that perceived Black group competition did not have a statistically significant effect on the importance of GPA after Asian over-representation had been primed. The null effect can be attributed to two phenomena. First, the percentage of those who attributed a score of 9 or 10 to the importance of GPA and who agreed with perceived Black competition dropped from 85% of respondents in the control condition to 46% in the Asian over-representation condition. The clear success of Asian students at the UC system drastically tempered GPA support among those who saw themselves threatened by Blacks. However, their endorsement of GPA did not fall precipitously, as these respondents clustered around scores of 7 or 8 (40% of the respondents in the Asian over-representation condition compared to roughly 10% in the control condition).
The second contributing factor was the proportion of those who disagree with perceived Black competition increasing the importance they attribute to GPA after exposure to the Asian over-representation cue. For instance, only 30% of this subset of White respondents thought GPA was extremely important in the control condition while 37% of them thought the same in the experimental condition. Therefore, the null effect of Black group competition in the face of Asian over-representation encompasses both the elevation of the importance of GPA among those Whites that perceive relatively little group competition with Blacks, alongside the diminished importance of GPA among those that perceive relatively more group competition with Blacks. This suggests that those who initially perceive themselves to be less vulnerable to Black competition attempt to erect a barrier only when opportunities have truly become scarce. Meanwhile, those who already perceived themselves to be competing with Blacks and are very much willing to set the barrier high are compelled to slightly lower the barrier to give themselves a better chance to access these scarce and valued educational opportunities.
These unanticipated findings present consequential theoretical and methodological lessons. While the hypotheses initially derived from the theories of group position and principled politics did not involve interactions anticipating unusual effects in the presence of Asian over-representation, it is clear that the results sans interactions would have incorrectly indicated no main effect resulting from the over-representation cue and therefore no support for group position theory except in the case of Hispanics as the reference group for competition. However, because some White respondents changed the importance they attributed to GPA drastically depending upon the social context they are considering, what was eventually revealed as a disordinal interaction disguised the very real main effect of the experimental over-representation cue. Once the models took into account a fundamental sociological tenet—that context matters—and controlled for the intersecting slopes elicited by the different contexts, the experimental over-representation cue did produce a main effect, providing evidence largely in support of group position's predictions. Without controlling for the impact of social context as an interaction effect, one might have erroneously concluded that group threat's influence was underwhelming at best, and inconsequential at worst. However, because individuals do in fact alter their attitudes and behaviors in response to different social contexts, controlling for these sometimes contradictory adaptations eventually did bring to light the main effect of group threat on the importance of grade point average as an admissions factor.
Prior research has pointed to the importance of meritocracy for White judgments about ethnoracial outgroups (Feagin and O'Brien, Reference Feagin and O'Brien2003) and the ideology of color-blindness (Bonilla-Silva Reference Bonilla-Silva2001, Reference Bonilla-Silva2003). While these studies, based on in-depth interviews, play a valuable role in revealing the centrality of meritocracy in shaping the views of many Whites, the quantitative data analyses employed here directly reveal meritocracy's political malleability in ways that complement these qualitative studies. Not only do we have evidence that meritocracy matters, but also with the use of survey-based experiments, we have evidence that the level of importance given to meritocracy can be manipulated by experimentally making a particular group salient. Furthermore, different effects ensue when priming multiple groups simultaneously. Ultimately, this study's findings call for additional studies utilizing survey-based experiments to explore various attitudes in a multi-racial context.
CONCLUSION
The current study presents three main results. First, how Whites adjust the importance that a criterion of academic merit should have depends upon which ethnoracial outgroups Whites are considering.Footnote 8 This finding weakens the argument that White commitment to meritocracy is purely based on principle, since the importance given to particular meritocratic criteria, here grade point average, varies depending upon the outgroups under consideration and the extent of the group threat they pose to Whites. Second, unlike the early twentieth century changes to meritocratic ideology when Whites discounted the importance of intellectual criteria in order to stave off the encroachment of Jews at elite universities, White respondents are now confronted with multiple racial groups who are stereotypically viewed as either superior or inferior to Whites on certain dimensions (Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002). White response to group threat from opposite directions is a historically different phenomenon than the response to Jewish threat; the data reveal this conflict of being caught in between multiple groups, not unlike the experience of “triangulation” for Asian Americans (Kim Reference Kim1999).
The third central finding is that cognitive contexts in which White positioning as the dominant group is being seriously threatened can potentially alter the effects of some theoretical mechanisms. Some of the predicted effects of perceived group competition were found only when Whites were operating in a cognitive environment in which Asian over-representation had not been made salient. This context-based finding presents a promising opportunity for future research in a variety of social and political psychological areas of inquiry, as perceived group competition may not be the only social psychological factor altered by the introduction of a new, cognitive and demographic environment where Whites are no longer the dominant group.
Outside the context of education, meritocracy remains a central component in the allocation of desirable jobs and opportunities. On June 29, 2009, the U.S. Supreme Court decided in favor of White firefighters who sued the city of New Haven, Connecticut for disregarding examination scores that ranked White firefighters ahead of Black firefighters in line for promotion. At the core of this contentious issue was a commitment to the principles of meritocracy and fairness. While the lawsuit against New Haven, Connecticut was fueled by desires to emphasize fairness in the application/promotion review process, another possible explanation for these actions exist: a segment of the White population coming to terms with ethnoracial group threat and the challenge such threat poses to jobs they felt belonged to them by virtue of their test-based merit. Yet both the present study and the historical evolution of meritocracy at elite universities in the United States offer an intriguing question to consider: Would the White firefighters still have sued New Haven if Jews or Asian Americans had also taken the examination and earned scores that ranked them higher than the White firefighters?
As the White population in the United States moves towards a numerical minority, the White populace may see itself losing its share of scarce and desirable resources in some domains of social life (Eibach and Keegan, Reference Eibach and Keegan2006; Norton and Sommers, Reference Norton and Sommers2011). Group competition over scarce goods also does not solely affect Whites (Gay Reference Gay2006; Oliver and Wong, Reference Oliver and Wong2003). The consequences of social and cognitive shifts due to changing demographics in the United States should yield plenty of new findings in fields of inquiry where conventional social and political psychological theories may seem to have left questions and debates long settled.
APPENDIX A: Methodological Appendix
Response Rate
The response rate raises a legitimate question about whether or not the survey sample is representative of the state's population. While low survey response rates have traditionally been a cause for concern among social scientists, recent public opinion research indicates that low survey response rates need not necessarily imply nonresponse bias (Groves Reference Groves2006; Keeter et al., Reference Keeter, Kennedy, Dimock, Best and Craighill2006). Nevertheless, in order to adjust for possible nonresponse bias, the survey results reported in this study are based on survey regressions using post-stratification weights. The post-stratification weights created by the University of California Survey Research Center incorporates sampling weights, which address differences in the probability of being selected for the sample, and post-stratification adjustments which address differences in likelihood to have a phone, and respond to a phone interview. The sampling weights were determined based upon the number of eligible phone numbers used in the sample, the number of phone numbers within a household, the number of eligible adults living within the household, as well as the probability of using a cell phone versus a landline. In order to determine the post-stratification adjustments, the sample was divided into categories by age (18–24 years, 25–34 years, 35–44 years, 45–54 years, 55–64 years, 65 plus years), education (less than ninth grade, less than high school, high school, some college, or college grad), and combined race and gender (e.g., White male, White female, Black male, Black female, etc.). These cells, weighted with the first round of sampling weights, were then weighted again to match the state's distribution on these categories as delineated by the 2005 American Community Survey, conducted by the U.S. Census Bureau. This second level of post-stratification weights, which include the sampling weights and adjustments for California distributions across age, education, and race/gender, were utilized for the survey regression models whose results are reported in the main text.
A comparison of the weighted sample to the actual state population, based on the 2005 American Community Survey used to calculate the post-stratification weights, reveal that the weighted sample approximates the state population's characteristics on age, income, and education (see Table A1).
Table A1. Population Profile for California (White Alone) Compared to Weighted Sample Distribution of White Respondents
Missing Values and Multiple Imputation Models
A missing values analysis for White respondents using T-tests to evaluate differences in the dependent variable between cases with complete values and cases with some missing values on the social psychological independent variables revealed that values were not missing at random: significant differences in the mean of the dependent variable did indeed exist between complete cases and cases with missing values. Bivariate (T-test) results tables from the missing values analysis are available upon request.
Because the missing values are not missing at random, and therefore non-ignorable, multiple imputations on the independent variables were conducted using the ICE (Imputation by Chained Equations) module for multiple imputations in STATA (Royston Reference Royston2004), which produced five data sets with imputed values on the missing cases for the social psychological variables. The MIM (multiple imputations) module in STATA for conducting analyses on multiple imputed datasets performed the analysis on all five imputed datasets simultaneously, calculating coefficients and standard errors by utilizing within and between model variance across all five imputed datasets. The results presented in the main text are based on the combined analysis performed by the MIM module. For comparative purposes, the results presented below are based on models estimated using solely the complete cases, with listwise deletion omitting incomplete cases and yielding a valid N of 440 California respondents.
Results (Complete Cases)
The complete cases models confirm the experimental results from the multiple imputations models for grade point average (see Table A2). The Asian over-representation cue has the same effects on each of the Asian and Black ballots. On the Black target ballot, competitive group threat retains a positive effect on the importance of grade point average in the control condition. Contrary to the multiple imputation models, the positive effect of individualism is now statistically significant on the Asian ballot. However, the finding of group-salient individualism still holds.
Table A2. Multiple Linear Regression of the Importance of Grade Point Average as a University Admissions Criterion on Social Psychological Variables, Experimental Ballot, and Interactions (Complete Cases Sample)
Conclusion
The complete cases models support the general conclusions from the multiple imputations models: racial attitudes matter for White attitudes towards university admissions criteria.
APPENDIX B: Survey Items for Social Psychological Scales
Group Threat Items
(If the race of the respondent matches the race of the randomly assigned target group, respondents received the phrase “members of other groups” rather than “people like me.”)
• Many [RACE GROUP] have been trying to get ahead economically at the expense of people like me.
• More well-paying jobs for [RACE GROUP] means fewer well-paying jobs for people like me.
• The more influence [RACE GROUP] have in local politics, the less influence people like me will have in local politics.
• As more nice housing and better neighborhoods go to [RACE GROUP], the fewer nice houses and better neighborhoods there will be for people like me.
Individualism Items
• People who don't get ahead have only themselves to blame.
• Hard work offers little guarantee of success.
• If people work hard they usually get what they want.
• Most people who don't get ahead probably work as hard as those who do.
• Anyone willing to work hard has a good chance of succeeding.
• Even if people try hard they often cannot reach their goals.
