A four-dimensional model of Asian American stereotypes

Abstract

The content of the Asian American model minority stereotype is important for understanding how Asian American individuals are perceived. Existing theories about stereotype content may not capture the unique historical and cultural context that could affect perceptions of Asian American individuals. We have identified a more differentiated underlying structure with four dimensions—warmth, competence, self-centeredness, and submissiveness—that differ in their rated typicality and desirability for Asian and White Americans. We then developed the 16-item Asian American Stereotypes Scale to measure perceptions of Asian Americans on these four dimensions. Ratings on the different dimensions predict unique variance in attitudes toward Asian Americans and other minority groups, contact with Asians or Asian Americans, perceptions of size of the Asian American population, and system justification. The four-dimensional model and the Asian American Stereotypes Scale allow us to predict and examine the unique impacts of Asian American stereotypes in a way that differs from more general models.

Keywords

Asian American stereotypes Asian American Stereotypes Scale (AAS)four-dimensional model model minority stereotype

What is the model minority stereotype of Asian Americans? In social psychology, Asian Americans have been touted as a primary example of the high-competence, low-warmth quadrant of the stereotype content model (Fiske et al., 2002). But is there more to the stereotype than competence and unsociability?

This question is important because stereotype content has implications for how individuals are perceived and treated. The stereotype content model predicts that groups perceived as high in competence but low in warmth elicit respect, envy, and dislike from members of other groups (e.g., Fiske, 2018; Fiske et al., 1999). Based on this model, Lin et al. (2005) developed a Scale of Anti-Asian American Stereotypes to capture perceptions of Asian Americans as overly competent and unsociable, which in turn were found to predict attitudes and behaviors toward Asian American individuals.

However, the stereotype content model does not address unique aspects of the perception and treatment of Asian Americans that may differ from other high-competence, low-warmth groups. The Scale of Anti-Asian American Stereotypes, though it is specific to Asian American stereotypes, focuses on negatively valenced content and thus does not speak to the effects of positive or neutral stereotypes of Asian Americans. Additionally, neither the stereotype content model nor the Scale of Anti-Asian American Stereotypes lend themselves directly to predictions about how an individual’s conformity or lack thereof to their group’s stereotypes influences others’ perceptions of them.

Within the literature on racial attitudes and racial bias, the stereotype content model is the main theory that addresses the content of stereotypes as opposed to processes of social categorization, ingroup favoritism, and outgroup derogation (see e.g., Dovidio & Gaertner, 2010, p. 1085; Yzerbyt & Demoulin, 2010, pp. 1034–1035). Although competing general stereotype content models have been developed (e.g., Koch et al., 2016; Leach et al., 2007), they are less widely known than the stereotype content model and do not capture the nuances that a group-specific model might.

A recent line of research by Zou and Cheryan (2017) identified two dimensions of negative racial and ethnic stereotypes: inferiority and cultural foreignness. But like the Scale of Anti-Asian American Stereotypes, this model focuses on negatively valenced stereotypes—Asian Americans are seen as relatively “superior” to African Americans and Latinos but not as superior to Whites—and it does not directly lend itself to predictions of how stereotypical and counterstereotypical individuals might be perceived or treated. Ho and Jackson (2001) developed a four-factor scale specific to Asian American stereotypes, but they did so in the process of developing a separate scale of evaluative attitudes, and they did not further validate the stereotype scale.

Relatedly, although some literature on intergroup contact addresses the relationship between typicality of outgroup members and effectiveness of contact (e.g., Stathi et al., 2011), that literature does not address what makes an individual typical of any particular group or whether different kinds of atypicality have different consequences. Thus, the racial prejudice literature does not lend itself to specific predictions about how individuals who are typical or atypical of a racial group (and specifically of Asian Americans) on particular sets of traits will be perceived and treated.

By contrast, two theories from the gender stereotyping and discrimination literature—role congruity theory (Eagly & Karau, 2002) and the status incongruity hypothesis (Rudman et al., 2012)—explicitly predict that stereotype-conforming and stereotype-nonconforming individuals will be perceived and treated differently. Both of these theories involve prescriptive stereotypes, traits that are perceived as desirable for members of a particular group, as well as the descriptive stereotypes generally considered in the stereotype content model literature.

Role congruity theory further predicts that people will tend to infer stereotype-consistent characteristics when perceiving men and women, and that women who violate gender stereotypes (for example, by taking on a leadership role) will be evaluated negatively (Eagly & Karau, 2002). Thus, role congruity theory predicts two different types of prejudice toward female leaders, depending on their conformity or nonconformity to gender stereotypes:

(a) less favorable evaluation of women’s (than men’s) potential for leadership because leadership ability is more stereotypical of men than women and (b) less favorable evaluation of the actual leadership behavior of women than men because such behavior is perceived as less desirable in women than men. (Eagly & Karau, 2002, p. 576)

The applicability of role congruity theory to racial and ethnic stereotypes was further demonstrated in Koenig and Eagly’s (2014) studies of current and projected occupational roles associated with a variety of groups, including racial and ethnic groups, and stereotypes of those groups. Importantly for our purposes, Koenig and Eagly (2014) advocate using role congruity theory to explain “unique stereotypes” associated with specific immigrant groups (p. 388). But so far, little research has explored what those unique ethnic stereotypes might entail.

The status incongruity hypothesis (Rudman et al., 2012) proposes that prescriptive and proscriptive (undesirable) traits are tied to groups’ status in society: Specifically, traits that are prescriptive for men (“agency” traits) and traits that are proscriptive for women (“dominance” traits) are associated with high status, while traits that are proscriptive for men (“weakness” traits) are associated with low status. When men and women behave in ways that support the status quo, they are perceived as possessing their prescriptive traits, but when they behave in ways that challenge the status quo, they are perceived as and punished for possessing their proscriptive traits. The status incongruity hypothesis thus predicts that individual women who conform to traditional feminine stereotypes reinforce the status quo in which women hold lower status than men, and individual women who break traditional feminine stereotypes face backlash, as people attempt to preserve the status quo.

Importantly, the implications of the status incongruity hypothesis for racial minorities and specifically Asian Americans are (a) that prescriptive and proscriptive traits might exist for racial groups as well as for genders and (b) that violating status quo-reinforcing stereotypes might result in backlash for Asian American individuals in similar ways as it does for women. Indeed, Phelan and Rudman (2010), Rosette et al. (2016), and Sy et al. (2010) demonstrated that Asian Americans can face prejudice when violating racial stereotypes or occupying roles that are incongruent with racial stereotypes. Berdahl and Min (2012) showed that Asian Americans face a dominance penalty in the workplace: using competence items from the stereotype content model and warmth and dominance items from the Interpersonal Adjective Scale (Wiggins et al., 1988), these researchers found that East Asians are rated as more competent, less warm, and less dominant than Whites (descriptively); that dominance is rated as less desirable for Asians than for Whites; and that people dislike Asian coworkers who show dominance compared to nondominant Asian and White coworkers as well as dominant White coworkers.

What is not yet developed in the literature, however, is a model for the content of Asian American stereotypes that can be used to systematically test the effects of various types of stereotype violations on attitudes and behavior toward Asian American individuals. Phelan and Rudman (2010) and Sy et al. (2010) did not attempt to tie their stereotyped characteristics of interest to specific stereotype dimensions. Berdahl and Min (2012) came close to doing this by focusing on warmth, competence, and dominance as dimensions, but they only demonstrated that dominance as measured by the Interpersonal Adjective Scale is a descriptive and proscriptive stereotype for Asians. There might be other dimensions of Asian American descriptive and prescriptive/proscriptive stereotypes that are rewarded or punished in an individual who displays them (cf. Phelan & Rudman, 2010; Rosette et al., 2016; Sy et al., 2010) and that might be identified through a systematic examination of stereotyped traits. If these additional dimensions or prescriptive/proscriptive elements exist, identifying these dimensions or elements might make it possible to differentially predict treatment of Asian American individuals who violate the stereotype in different ways. For example, the dominance penalty that Berdahl and Min (2012) found against Asians in the workplace might differ from the perceptions and behaviors triggered by an Asian American individual who demonstrates high sociability or who does not demonstrate high academic competence. Thus, our goal was to develop a model to organize research on prejudice against stereotype-congruent and stereotype-incongruent Asian American individuals.

Moreover, our research goes beyond the existing stereotype content research because we posit that negative dimensions of stereotype content exist as something other than the opposite of warmth or competence. We build on the idea suggested by Zou and Cheryan’s (2017) research that negative dimensions outside the stereotype content model framework exist, as well as the possibility that Rudman et al.’s (2012) proscriptive dimensions of dominance and weakness might be something other than excessive agency and lack of agency, respectively. Like Koch et al. (2016), we took precautions to avoid predisposing our data to fall into the two dimensions of warmth and competence, in our case by starting with a large, varied pool of traits.

Our research also goes beyond the existing stereotypes and prejudice literature because the literature to date does not generally focus on Asian Americans. Yet Asian Americans occupy a unique position in the history of intergroup relations in the United States that has led to their current model minority image (Wu, 2014): Asian Americans were initially stereotyped as unassimilable foreigners, that is, “definitively not-white” (p. 2) and then were rebranded as “definitively not-black” in response to the Civil Rights Movement (p. 149). As a result, one can expect both positive and negative traits to be associated with Asian Americans (reflecting not-Black and not-White historical characterizations, respectively), and one can expect these stereotyped traits to reinforce a unique status distinct from those of White and Black Americans. Theories like the stereotype content model, Koch et al.’s (2016) ABC model, and other general stereotype models designed to apply broadly and parsimoniously across a large number of groups and contexts do not address the more specific aspects of Asian American stereotypes that follow from American history. Drawing on a theory from the gender literature (role congruity theory) that emphasizes social and cultural context as a basis for stereotypes gives us a stronger basis for exploring the historically situated stereotypes of Asian Americans and how those stereotypes affect individuals who do or do not behave in ways that American society has come to expect.

To this end, we conducted two studies on the content of Asian American stereotypes. In Study 1, participants rated the typicality or desirability of a large pool of traits for Asian or White Americans. We used the trait ratings to identify descriptive, prescriptive, and proscriptive traits and examine the dimensionality of the descriptive traits. Using the results of Study 1, we developed a four-dimensional scale of Asian American stereotype content (abbreviated AAS, for Asian American Stereotypes Scale), and in Study 2, we tested the construct and predictive validity of our scale.

Study 1

Method

Participants

Participants were undergraduates from a large Midwestern university who completed the study online for 1 extra credit point.¹ In the first wave of data collection, over the course of a semester, 471 students participated (128 male, 336 female; age 18–37, M = 20.1, SD = 2.1; 68% White, 24% Asian, 5% Black, 4% Latino). Because we planned to conduct factor analysis with the Asian descriptive trait ratings, we collected data over the following semester from an additional 306 participants (76 male, 212 female, two other; age 18–35, M = 20.4, SD = 2.6; 63% White, 25% Asian, 4% Black, 5% Latino, 1% Native American) in the Asian-typical condition only (see Procedure section) to meet rule-of-thumb requirements for exploratory factor analysis (N of at least 250–300), stopping data collection at the end of the semester.

Materials

The main portion of this study consisted of a list of 103 traits compiled from the literature on the stereotype content model; agency and communion dimensions of group stereotypes (e.g., Abele et al., 2008); Rudman et al.’s (2012) male and female prescriptive and proscriptive traits; and trait versions of some of the items from the Scale of Anti-Asian American Stereotypes (Lin et al., 2005). A full list of traits is included in the supplemental material. Participants were asked to rate how common or typical each trait is for Asian Americans, White Americans, high-status Americans, or low-status Americans, or how desirable each trait is for Asian Americans or White Americans. For example, in the Asian-typical condition, participants were instructed, “Please indicate how common or typical it is for an Asian person in American society to possess each of the following characteristics.” Traits were rated on a 9-point scale (1 = not at all typical [or not at all desirable], 9 = very typical [or very desirable]). Complete instructions for all conditions are included in the supplemental material.

Participants also completed the 22-item Scale of Anti-Asian American Stereotypes (Lin et al., 2005) and three questions about interpersonal contact with Asians or Asian Americans, and were asked to report their age, gender, and race/ethnicity. (Neither the Scale of Anti-Asian American Stereotypes nor the contact questions were analyzed for purposes of this study.)

Procedure

After consenting to participate in the study, participants were randomly assigned to one of six conditions: Asian desirable, White desirable, Asian typical, White typical, high-status typical, and low-status typical. Thus, to minimize social desirability responding and in keeping with Rudman et al.’s (2012) and Koenig and Eagly’s (2014) key studies, our comparisons between Whites and Asians and high and low status were between subjects (see also Dovidio et al., 1996, for an overview of measures and methods used in stereotype trait assessment studies, most of which involve between-subject comparisons). Each condition corresponded to an instruction set according to which participants were to rate the 103 traits. The traits were presented on two pages, with page order counterbalanced, and the traits on each page were presented in random order. The additional 306 participants recruited for the exploratory factor analysis (EFA) portion of the study were all assigned to the Asian-typical condition.

Following the trait ratings, participants completed the Scale of Anti-Asian American Stereotypes, again on two pages, with page order counterbalanced and items presented in random order within each page. Participants then completed the questions about contact with Asian Americans, followed by the age, gender, and race questions. Finally, they read a debriefing paragraph before exiting the survey.

Using the data from the participants who rated how desirable the traits are for Asian Americans and the participants who rated how desirable they are for White Americans, we determined Asian and White prescriptive and proscriptive traits using the method described by Rudman et al. (2012): prescriptive traits were those for which the mean desirability rating for the target group was greater than or equal to 6.0 (on our 9-point scale), and the mean desirability rating for the target group was greater than the mean desirability rating for the other group by a standardized difference score of at least d = 0.40. Standardized between-groups difference scores were calculated by using the Hedges’s g formula $(d = \frac{M_{W h i t e} - M_{A s i a n}}{s_{p}}, s_{p} = \frac{\sqrt{n_{1} s_{1}^{2} + n_{2} s_{2}^{2}}}{n_{1} + n_{2} - 2})$ , with positive d scores indicating higher ratings for Whites and negative d scores indicating higher ratings for Asians. Thus, a White prescriptive trait was one for which the mean White desirability rating was at least 6.0, and the difference between White and Asian desirability ratings was d ⩾ 0.40; an Asian prescriptive trait was one for which the mean Asian desirability rating was at least 6.0, and the difference between White and Asian desirability ratings was d ⩽ −0.40. Proscriptive traits were those for which the mean desirability rating for the target group was less than or equal to 4.0, and the mean desirability rating for the target group was less than the mean desirability rating for the other group by at least d = 0.40. Using the same formula, a White proscriptive trait was thus a trait for which the mean White desirability rating was 4.0 or lower, and the difference between White and Asian desirability ratings was d ⩽ −0.40; an Asian proscriptive trait was one for which the mean Asian desirability rating was 4.0 or lower, and the difference between White and Asian desirability ratings was d ⩾ 0.40.

Descriptive scores were calculated as the standardized difference score between White and Asian mean typicality ratings or between high- and low-status mean typicality ratings, and these mean differences were tested for significance using independent-sample t tests. Positive descriptive d scores indicate that the trait was rated higher in typicality for Whites than for Asians or higher in typicality for high-status than for low-status people. Negative descriptive d scores indicate that the trait was rated higher in typicality for Asians than for Whites or higher in typicality for low-status than for high-status people.

Finally, combining the data from the 79 participants originally assigned to the Asian-typical condition and the 306 later participants, we performed an exploratory factor analysis (EFA) of the trait descriptiveness ratings for Asian Americans. Figure 1 illustrates our procedure.

Figure 1.

Schematic of Study 1 procedure.

All statistical analyses were conducted using R statistical software (Version 3.3.0). EFA was conducted using the nFactors package. The optimal number of factors was determined using parallel analysis, and factor solutions were derived using oblimin rotation.

Results

Prescriptive and proscriptive traits

Most of the traits that met Rudman et al.’s criteria for prescriptive and proscriptive traits were prescriptive for Whites. As shown in Table 1, the White prescriptive traits in descending order of d scores were popular, outgoing, persuasive, active, assertive, energetic, fun-loving, social, confident, friendly, enthusiastic, self-confident, likeable, independent, affectionate, self-reliant, successful, open, loyal, and competent. Difference scores ranged from d = 0.96 to d = 0.40. Based on independent-sample t tests, all d scores ⩾ 0.40 or ⩽ −0.40 in this study correspond to statistically significant differences between the two groups being compared. A post hoc power analysis using the smallest group size (n = 76) indicated that an effect size of d = ±0.40 could be detected with power = .69, and that at a power of .80 we could detect an effect size in this sample of d = ±0.46 or larger.

Table 1.

Study 1: White prescriptive and proscriptive traits and Asian proscriptive traits.

	M _White-p	M _Asian-p	t(df)	p	d	M _White-d	M _Asian-d	t(df)	p	d	M _high	M _low	t(df)	p	d
Popular	6.86	5.02	6.07 (157)	< .001	0.96	6.65	4.91	6.48 (155)	< .001	1.03	6.83	3.74	11.35 (146)	< .001	1.83
Outgoing	7.47	5.92	5.59 (136)	< .001	0.88	6.49	4.71	6.89 (133)	< .001	1.09	6.75	4.88	6.40 (151)	< .001	1.03
Persuasive	7.00	5.72	5.00 (143)	< .001	0.79	6.13	4.75	5.32 (147)	< .001	0.84	7.61	4.59	11.30 (137)	< .001	1.81
Active	7.51	6.41	4.80 (151)	< .001	0.75	6.50	5.39	4.54 (155)	< .001	0.72	6.69	4.64	7.44 (151)	< .001	1.20
Assertive	6.77	5.49	4.42 (157)	< .001	0.70	6.28	4.62	6.80 (143)	< .001	1.08	7.17	4.38	10.72 (142)	< .001	1.72
Energetic	7.35	6.38	4.33 (158)	< .001	0.68	6.28	5.06	4.70 (134)	< .001	0.75	6.32	4.55	6.45 (151)	< .001	1.04
Fun-loving	7.20	6.10	4.25 (158)	< .001	0.67	6.32	5.32	4.17 (140)	< .001	0.66	5.51	5.26	0.86 (151)	.389	0.14
Social	7.43	6.40	4.00 (148)	< .001	0.63	6.85	4.91	7.47 (136)	< .001	1.18	7.11	5.33	6.60 (140)	< .001	1.06
Confident	7.67	6.76	3.84 (157)	< .001	0.60	6.81	5.56	4.91 (147)	< .001	0.78	7.75	4.40	12.93 (136)	< .001	2.07
Friendly	7.78	6.90	3.71 (157)	< .001	0.59	6.23	6.14	0.42 (125)	.679	0.07	5.76	5.51	0.90 (146)	.368	0.14
Enthusiastic	7.32	6.44	3.51 (151)	.001	0.55	6.28	5.44	3.29 (139)	.001	0.52	6.53	4.54	7.43 (151)	< .001	1.19
Self-confident	7.66	6.88	3.39 (144)	.001	0.53	6.69	5.62	4.62 (155)	< .001	0.73	7.52	4.16	12.96 (133)	< .001	2.10
Likeable	7.52	6.64	3.32 (157)	.001	0.52	6.15	6.11	0.19 (139)	.848	0.03	6.21	5.05	4.26 (151)	< .001	0.68
Independent	7.61	6.80	3.30 (140)	.001	0.52	6.82	6.32	1.95 (148)	.053	0.31	7.49	5.23	7.60 (129)	< .001	1.22
Affectionate	6.82	5.97	3.24 (157)	.001	0.51	5.95	4.92	4.21 (144)	< .001	0.67	5.27	5.42	−0.56 (151)	.574	−0.09
Self-reliant	7.44	6.72	3.02 (157)	.003	0.48	6.28	6.47	−0.75 (155)	.454	−0.12	6.88	5.40	4.58 (135)	< .001	0.73
Successful	8.16	7.54	2.99 (148)	.003	0.47	6.77	7.09	−1.46 (155)	.147	−0.23	8.00	3.83	15.25 (129)	< .001	2.44
Open	6.71	5.89	2.75 (153)	.007	0.43	5.95	4.42	5.47 (145)	< .001	0.87	5.64	4.95	2.29 (151)	.024	0.37
Loyal	7.43	6.70	2.74 (157)	.007	0.43	5.83	6.14	−1.26 (155)	.209	−0.20	5.27	5.45	−0.60 (151)	.550	−0.10
Competent	7.75	7.10	2.57 (140)	.011	0.40	6.36	6.90	−2.38 (154)	.018	−0.38	7.28	4.74	8.58 (145)	< .001	1.37
Superstitious	3.35	4.16	−2.67 (142)	.009	−0.42	5.22	4.70	1.91 (147)	.058	0.30	3.93	4.99	−3.33 (150)	.001	−0.54
Shy	3.65	4.62	−3.20 (147)	.002	−0.50	4.42	6.16	−6.46 (154)	< .001	−1.03	3.36	4.99	−5.74 (150)	< .001	−0.92
Rebellious	3.90	3.01	3.31 (157)	.001	0.52	5.18	3.28	7.40 (154)	< .001	1.18	4.45	5.32	−3.04 (150)	.003	−0.49
Aggressive	3.92	3.08	2.60 (147)	.010	0.41	5.79	3.62	7.85 (155)	< .001	1.24	5.79	4.65	3.75 (150)	< .001	0.60

Two traits met the criteria for White proscriptive traits: superstitious (M_White = 3.35, M_Asian = 4.16, d = −0.42) and shy (M_White = 3.65, M_Asian = 4.62, d = −0.50); and two traits met the criteria for Asian proscriptive traits: rebellious (M_White = 3.90, M_Asian = 3.01, d = 0.52) and aggressive (M_White = 3.92, M_Asian = 3.08, d = 0.41). No traits met the criteria for Asian prescriptive traits.

Both White proscriptive traits were rated as more descriptive of low-status people than of high-status people: superstitious: M_high = 3.93, M_low = 4.99, t(150) = −3.33, p = .001; shy: M_high = 3.36, M_low = 4.99, t(150) = −5.74, p < .001. Of the Asian proscriptive traits, rebellious was rated as more descriptive of low-status people, M_high = 4.45, M_low = 5.32, t(150) = −3.04, p = .003, but aggressive was rated as more descriptive of high-status people, M_high = 5.79, M_low = 4.65, t(150) = 3.75, p < .001. Most of the White prescriptive traits were rated as significantly more descriptive of high-status people than of low-status people, with the exception of fun-loving, friendly, affectionate, and loyal, which did not significantly differ in typicality ratings between high and low status. All of the prescriptive/proscriptive traits, along with their descriptive and high- and low-status rating information, are displayed in Table 1.

Descriptive traits for Asian and White Americans

Based on the data from the original sample, 43 traits had White – Asian typicality d scores above 0.40, indicating that they were considered more descriptive of Whites than of Asians (or less descriptive of Asians than of Whites). Thirteen traits had typicality d scores below −0.40, indicating that they were considered more descriptive of Asians than of Whites (or less descriptive of Whites than of Asians).

As shown in Table 2, the 10 traits with the largest standardized differences favoring Whites were attention-seeking, dominating, boastful, greedy, power-seeking, egotistical, arrogant, conceited, and aggressive. All of these traits except aggressive had mean typicality ratings ⩾ 6.0 for Whites, indicating that they were considered descriptive of White Americans. None of these traits had mean typicality ratings ⩾ 6.0 for Asians, and in fact seven of the 10 had typicality ratings ⩽ 4.0, indicating that these traits were considered atypical for Asian Americans.

Table 2.

Study 1: White and Asian descriptive traits.

	M _White-p	M _Asian-p	t(df)	p	d	M _White-d	M _Asian-d	t(df)	p	d	M _high	M _low	t(df)	p	d
Attention-seeking	4.24	3.61	1.97 (158)	.050	0.31	6.46	3.89	9.90 (155)	< .001	1.79	6.11	4.60	4.86 (151)	< .001	0.78
Dominating	4.78	4.11	2.04 (157)	.043	0.32	6.36	3.82	9.90 (155)	< .001	1.57	6.78	3.65	10.85 (150)	< .001	1.75
Boastful	3.78	3.71	0.17 (158)	.853	0.03	6.27	3.87	9.74 (154)	< .001	1.55	6.00	4.14	6.13 (151)	< .001	0.98
Greedy	2.99	3.00	−0.04 (157)	.966	−0.01	6.19	3.90	8.82 (155)	< .001	1.40	5.97	4.05	6.25 (150)	< .001	1.01
Power-seeking	5.77	5.28	1.42 (157)	.158	0.22	7.13	4.97	8.31 (138)	< .001	1.31	7.27	3.99	11.22 (151)	< .001	1.80
Egotistical	3.71	3.32	1.20 (153)	.234	0.19	6.29	4.27	8.17 (155)	< .001	1.30	6.04	3.96	6.82 (151)	< .001	1.10
Arrogant	3.30	2.98	1.02 (157)	.307	0.16	6.15	3.91	8.15 (155)	< .001	1.29	5.96	3.91	6.54 (151)	< .001	1.05
Self-centered	3.52	3.12	1.24 (149)	.218	0.20	6.32	4.09	8.14 (155)	< .001	1.29	6.08	4.06	6.81 (151)	< .001	1.09
Conceited	3.64	3.08	1.73 (156)	.086	0.27	6.05	3.97	7.96 (155)	< .001	1.26	5.29	4.01	4.15 (151)	< .001	0.67
Aggressive	3.92	3.08	2.60 (147)	.010	0.41	5.79	3.62	7.85 (155)	< .001	1.24	5.79	4.65	3.75 (150)	< .001	0.60
Polite	7.53	7.34	0.83 (157)	.408	0.13	6.00	6.71	−2.99 (154)	.003	−0.48	5.75	5.27	1.68 (151)	.095	0.27
Humble	7.03	6.54	1.68 (156)	.096	0.26	5.26	6.08	−3.23 (155)	.002	−0.51	4.68	5.77	−3.58 (151)	< .001	−0.56
Striving	7.28	7.06	0.87 (157)	.386	0.14	6.23	6.97	−3.41 (155)	.001	−0.54	7.36	4.72	8.89 (117)	< .001	1.41
Analytical	6.58	7.11	−1.86 (151)	.064	−0.29	6.09	6.99	−3.58 (155)	< .001	−0.57	6.56	4.23	8.42 (143)	< .001	1.35
Efficient	7.57	7.28	1.20 (157)	.233	0.19	5.99	6.92	−4.04 (155)	< .001	−0.64	7.25	4.65	9.58 (125)	< .001	1.52
Hard-working	7.90	7.59	1.33 (156)	.186	0.21	6.33	7.37	−4.50 (155)	< .001	−0.71	7.47	5.41	6.40 (135)	< .001	1.02
Intelligent	7.85	7.86	−0.06 (157)	.948	−0.01	6.35	7.53	−4.85 (155)	< .001	−0.77	7.53	4.47	11.09 (135)	< .001	1.77
Self-disciplined	7.56	7.06	2.08 (142)	.039	0.33	5.94	7.15	−5.26 (155)	< .001	−0.83	6.80	4.82	6.22 (146)	< .001	1.00
Reserved	4.69	5.48	−2.81 (158)	.006	−0.44	4.97	6.42	−5.69 (147)	< .001	−0.90	4.37	5.13	−2.53 (151)	.012	−0.41
Shy	3.65	4.62	−3.20 (147)	.002	−0.50	4.42	6.16	−6.46 (154)	< .001	−1.03	3.36	4.99	−5.74 (150)	< .001	−0.92

The 10 traits with the largest standardized differences favoring Asians were shy, reserved, self-disciplined, intelligent, hard-working, efficient, analytical, striving, humble, and polite (see Table 2). All 10 had typicality ratings ⩾ 6.0 for Asians, indicating that they were considered descriptive of Asian Americans. Half (polite, striving, analytical, hard-working, intelligent) had typicality ratings ⩾ 6.0 for Whites, indicating that some Asian descriptive traits were considered descriptive of White Americans as well.

Exploratory factor analysis

Parallel analysis indicated four factors with eigenvalues greater than what would be expected by chance (eigenvalues: 24.77, 11.78, 6.55, 5.02).

Using oblimin rotation, we identified four correlated factors that roughly correspond to (a) warmth (e.g., warm, cheerful, enthusiastic, kind, supportive); (b) competence or skill (e.g., intelligent, skillful, hard-working, determined); (c) self-centeredness or dominance (e.g., self-centered, greedy, arrogant, aggressive); and (d) submissiveness or passivity (shy, submissive, passive). Factor loadings for all traits are included in the supplemental material. As shown in Table 3, the four factors correlated with each other between r = −.04 (competence and self-centeredness) and r = .39 (warmth and competence). Several traits loaded ⩾ .50 on both the warmth and competence factors: friendly, good-natured, sincere, trustworthy, loyal, polite, likeable, moral, and honest; but cross-loadings between other pairs of factors were less prevalent.

Table 3.

Study 1: Factor correlations.

	Warmth	Competence	Self-centeredness	Submissiveness
Warmth	1.00
Competence	.39	1.00
Self-centeredness	−.04	.14	1.00
Submissiveness	.03	.03	−.03	1.00

We also examined three- and two-factor solutions. A three-factor solution produced essentially the same warmth, competence, and self-centeredness factors, but traits like shy, submissive, and passive did not load strongly (loadings < .30) on any of those factors. A two-factor solution combined the warmth and competence traits into a single factor that was weakly positively correlated with a self-centeredness factor. Based on these results, the four-factor solution appeared to be the best fit for trait ratings of Asian Americans in terms of both fitting with existing theory (e.g., Rudman et al.’s [2012] prescriptive and prospective trait clusters; three of Ho and Jackson’s [2001] Asian stereotype subscales) and accounting for most of the traits in our data.

Finally, examining the overlap between the four factors and the Asian and White descriptive and proscriptive traits we identified, we noticed that the top 10 White descriptive traits and the Asian proscriptive traits loaded onto the self-centeredness factor. The top 10 Asian descriptive traits loaded onto the competence factor and, to a lesser extent, the warmth and submissiveness factors. (Loading patterns were less clear for the White prescriptive and proscriptive traits.)

Discussion

Our lists of prescriptive and proscriptive traits indicate that, unlike men and women, Asian Americans are not ascribed a prescriptive stereotype, that is, a set of traits that they should possess. Both Asian and White Americans, however, were ascribed proscriptive traits. Interestingly, this pair of findings suggests that Americans have a clear idea of what Asian Americans should not be, but not of what they should be.

Consistent with Rudman et al.’s (2012) findings, the prescriptive and proscriptive traits for White Americans (the dominant group in our study) tended to be associated with high status and low status, respectively. However, unlike women’s proscriptive traits, Asian Americans’ proscriptive traits were not consistently associated with high status. Nonetheless, the Asian proscriptive traits loaded onto a single self-centeredness factor, suggesting a common theme.

Asian trait typicality ratings loaded onto four modestly correlated factors, with cross-loadings between the two positively valenced factors (warmth and competence). The modest correlations and relative lack of cross-loadings between the competence-related trait ratings and the self-centeredness- and submissiveness-related trait ratings suggest that the latter two dimensions are not merely varieties of competence. (A similar conclusion can be drawn for the warmth factor and the self-centeredness and submissiveness factors.) Additionally, the White descriptive traits tended to load onto the self-centeredness factor, while the Asian descriptive traits tended to load onto the competence factor, further supporting the idea of a distinction between competence and self-centeredness.

Using the results of Study 1, we created a 16-item scale, the Asian American Stereotypes Scale or AAS, to capture ratings on the four dimensions we identified. We first identified the traits with the highest loadings on each factor. Then, we dropped any traits with a cross-loading ⩾ .40 on any other factor. To narrow down the list of traits for Factor 2 to a similar number as the other factors, we dropped the remaining traits from the initial list with the highest cross-loadings on other factors. Our final scale included five traits for Factor 1 (warmth): cheerful, sympathetic, fun-loving, warm, and enthusiastic; four traits for Factor 2 (competence): intelligent, motivated, well-educated, and successful; four traits for Factor 3 (self-centeredness): self-centered, conceited, arrogant, and irritable; and three traits for Factor 4 (submissiveness): shy, passive, and submissive.

Study 2

Study 2 aimed to validate the four-factor structure of Asian American stereotypes and validate the Asian American Stereotypes Scale.

First, to test the four-factor structure of the stereotypes and the construct validity of the scale, we used confirmatory factor analysis to confirm that the items we selected loaded onto the four factors of warmth, competence, self-centeredness, and submissiveness that we identified in Study 1. We hypothesized that a four-factor model would show reasonable model fit, with items loading onto their appropriate factors.

Second, we examined the discriminant validity of our scale against the other scale of Asian American stereotypes that we included in our study, the Scale of Anti-Asian American Stereotypes (Lin et al., 2005), by examining the correlations between our subscales and those of the Scale of Anti-Asian American Stereotypes.

Finally, we examined the predictive validity of the Asian American Stereotypes Scale by testing relationships between ratings of Asian Americans on the four subscales and various scales and constructs found in the literature to be associated with racial prejudice. We hypothesized that our subscales would, individually and in combination, predict responses to various racial attitude measures, including the Racial Resentment Scale (Kinder & Sanders, 1996) and feeling thermometers for racial groups and other groups that people associate with non-Whites (e.g., immigrants). Importantly, associations between individual subscales of our scale and racial prejudice measures would suggest roles for different dimensions of Asian American stereotypes in maintaining or justifying prejudice against Asian Americans and other groups, and indicate whether the same or different stereotype dimensions are likely to drive prejudice against Asian Americans compared to other groups.

We expected our measure to predict feeling thermometer ratings of its target group, Asians. The Warmth and Competence subscales, which capture positive attributes, were hypothesized to predict higher Asian feeling thermometer ratings, and the Self-Centeredness and Submissiveness subscales, which capture negative attributes, were hypothesized to predict lower Asian feeling thermometer ratings. Additionally, we expected contact with Asians or Asian Americans to be associated with more positive attitudes toward Asian Americans; thus, we expected intergroup contact to be associated with higher warmth and competence ratings and lower self-centeredness and submissiveness ratings.

We also hypothesized that our negative subscales (self-centeredness and submissiveness) would predict higher estimates of the proportion of Asians in the U.S. population (which we expected to be associated with higher racial bias). This prediction was based on Lin et al.’s (2005) finding that high scores on the Scale of Anti-Asian American Stereotypes’ Social subscale predicted larger overestimates of the percentage of Asian students on a college campus. Because the population of the United States is much larger than the student body at any university and thus cannot be easily estimated from individuals’ everyday experiences, we expected that people would rely at least in part on stereotypes and prejudice to make their estimates, and that stereotyping of Asian Americans on one or both of the negatively valenced dimensions of the Asian American Stereotypes Scale would thus predict overestimates of the proportion of Asians in the U.S. population.

Our hypotheses about the relationship between our measure and measures of attitudes toward other groups were more exploratory and nondirectional. Because attitudes toward different minority groups tend to show some correlation (e.g., Brandt & Crawford, 2019), our subscales could predict feeling thermometer ratings in the same way for other groups as for Asians, though these relationships may not be as strong as the relationship with Asian feeling thermometer ratings. On the other hand, given the history of Asian American stereotypes and their relationship to anti-Black racism (see e.g., Wu, 2014), feeling thermometer ratings for some groups and racial resentment scores might show the opposite pattern as the Asian feeling thermometer: the positive AAS subscales might predict lower feeling thermometer ratings or higher racial resentment, and the negative subscales might predict higher feeling thermometer ratings or lower racial resentment.

We also predicted that our scale would be associated with system justification (Kay & Jost, 2003) based on Rudman et al.’s (2012) theory that prescriptive and proscriptive gender stereotypes serve to preserve and justify the status quo. If Asian American stereotypes also have a status quo-supporting function, then ratings of Asian Americans on the Competence, Self-Centeredness, and/or Submissiveness subscales should be related to raters’ level of system justification. However, the direction of these relationships was not obvious: rating Asian Americans as high in competence and low in self-centeredness and submissiveness could justify their relatively high status compared to other minority groups, but rating Asian Americans as high in either self-centeredness or submissiveness could justify their lower status compared to Whites.

Method

Participants

Participants were 405 Amazon Mechanical Turk (MTurk) workers (216 male, 188 female, one other; age 18–80, M = 32.3, SD = 10.2; 288 White, 65 Black, 33 Hispanic/Latino, 34 Asian, 11 Native American, three Native Hawaiian/Pacific Islander, three other) who participated in the study for $0.75 in June 2018. Because we collected our data during a time frame that was subsequently flagged as having potential bot-related data quality problems (Chmielewski & Kucker, 2019), we checked the reliability of scales and ran our key factor analysis both with the full dataset and excluding the data from 32 participants who shared a common latitude and longitude.²

Materials

The focus of this study was the Asian American Stereotypes Scale (AAS) described before. Responses were given on a 9-point scale (1 = not at all typical, 9 = very typical). Participants also responded to the Scale of Anti-Asian American Stereotypes, as in Study 1. Additionally, they were given the four-item Racial Resentment Scale (Feldman & Huddy, 2005; Kinder & Sanders, 1996) to be answered on a 5-point scale (1 = disagree strongly, 3 = neither agree nor disagree, 5 = agree strongly); feeling thermometers (100-point slider: 1 = most negative, 100 = most positive) for White people, Black people, Latinos, Asians, legal immigrants, and undocumented immigrants; an open-ended question asking participants to estimate the percentage of Asians in the U.S. population; and the eight-item General System Justification Scale (Kay & Jost, 2003) to be answered on a 9-point scale (1 = strongly disagree, 9 = strongly agree). All items that did not come from an existing scale (i.e., the Asian American Stereotypes Scale, feeling thermometers, and estimate of the percentage of Asians in the U.S. population) are included in the supplemental material.

Participants were also asked four questions about direct interpersonal contact (e.g., “How frequently do you interact with Asians or Asian Americans in daily life?”) and four questions about extended contact (e.g., “How many of your friends have friends who are Asian or Asian American?”) with Asians or Asian Americans. These items are included in the supplemental material.

At the end of the study, participants were asked to report their age, gender, and race/ethnicity. Additionally, because we expected responses to racial attitude questions to be associated with political ideology and party identification, we asked participants to report their political ideology on a 7-point scale (1 = extremely liberal, 7 = extremely conservative); their party identification (Republican, Democrat, independent, or other); for Republicans and Democrats, whether they identify strongly or weakly with their party; and for independents and others, whether they consider themselves closer to the Republican Party or the Democratic Party.

Procedure

After consenting to participate in the study, all participants were given our 16-item Asian American Stereotypes Scale, with traits presented in random order. They then completed the Scale of Anti-Asian American Stereotypes, the Racial Resentment Scale, feeling thermometers, Asian population estimate item, contact with Asian Americans items, and System Justification Scale. The order of all measures besides the Asian American Stereotypes Scale was counterbalanced across participants, and items within each scale were presented in random order. After completing all of the scales, participants completed the demographics section (age, gender, race/ethnicity, political ideology, and party identification), and were then presented with a debriefing paragraph and paid.

Confirmatory factor analyses (CFA) were conducted for both the Asian American Stereotypes Scale and the Scale of Anti-Asian American Stereotypes.³ These analyses were performed in Mplus (Version 8.1) using oblique rotation, reference variables, and full-information maximum likelihood estimation. Additionally, we conducted a two-group CFA for our scale to test its structural invariance for White and non-White perceivers.⁴

To determine predictive validity, we first created composite scores for racial resentment (α = .77 for the full dataset; α = .81 dropping responses from the common location); system justification (α = .83 for the full dataset; α = .84 dropping responses from the common location); direct contact with Asians or Asian Americans (α = .89 for the full dataset; α = .88 dropping responses from the common location); and extended contact with Asians or Asian Americans (α = .88 for the full dataset; α = .86 dropping responses from the common location), by averaging responses to the items in each scale (after reverse-scoring racial resentment and system justification items as appropriate). Similarly, we created composite scores for each subscale of our scale and the Scale of Anti-Asian American Stereotypes. Next, we tested correlations between each subscale of both our scale and the Scale of Anti-Asian American Stereotypes⁵ and each of the variables we sought to predict: racial resentment, each feeling thermometer rating, estimate of the Asian population in the US, direct contact, indirect contact, and system justification. Finally, we used multiple regression to test the predictive validity of each subscale score for each criterion variable when controlling for age, gender, race/ethnicity, political ideology, and Scale of Anti-Asian American Stereotypes subscale scores.

Results

Construct validity

A four-factor oblique CFA model showed acceptable fit with the data from the Asian American Stereotypes Scale: SRMR = .04, RMSEA = .07, CFI = .95.⁶ Factor loadings and correlations are presented in Figure 2. The four subscales had fair to good internal consistency (warmth: α = .87; competence: α = .89; self-centeredness: α = .87; submissiveness: α = .77. Dropping common location: warmth: α = .86; competence: α = .90; self-centeredness: α = .86; submissiveness: α = .75). Factor scores were slightly to moderately correlated with each other, ranging from r = .14 (competence and self-centeredness) to r = .59 (self-centeredness and submissiveness; see Table 4; see also Table 5 for correlations between subscale composite scores). To test for configural invariance across perceivers (i.e., whether the same four-factor model fits the data from White and non-White perceivers), we ran separate CFAs for White and non-White participants. These separate CFAs produced somewhat poorer fit statistics (White: SRMR = .05, RMSEA = .07, CFI = .94; non-White: SRMR = .06, RMSEA = .10, CFI = .90) than the model including all participants, but this might be expected from the smaller sample size of each group (n_White = 288, n_non-White = 121). To test for metric invariance (i.e., whether the scale items have similar relationships to their respective factors for White and non-White perceivers), we ran multigroup CFAs with factor loadings constrained to equality across groups (SRMR = .06, RMSEA = .08, CFI = .92), and with unconstrained factor loadings (SRMR = .06, RMSEA = .08, CFI = .92). The constrained and unconstrained models did not differ significantly in fit: χ²_constrained = 483.13 (df = 220), χ²_{unconstrained} = 469.29 (df = 208), ∆χ² = 13.84 (df = 12), p = .311. Thus, the factor structure and factor loadings for our scale appear reasonably invariant for White and non-White perceivers. Reliability for each of the subscales was fair to good for both White and non-White perceivers, ranging from α = .75 (non-White, submissiveness) to α = .90 (White, competence).

Figure 2.

Confirmatory factor analysis model for the Asian American Stereotypes Scale in Study 2.

Table 4.

Study 2: AAS factor correlations.

	Warmth	Competence	Self-centeredness	Submissiveness
Warmth	1.00
Competence	.41***	1.00
Self-centeredness	.31***	.14*	1.00
Submissiveness	.34***	.44***	.59***	1.00

Note. AAS = Asian American Stereotypes Scale.

p < .05. **p < .01. ***p < .001.

Table 5.

Study 2: AAS and SAAAS subscale score correlations.

	Warmth	Competence	Self-centered	Submissive	SAAAS_S	SAAAS_C
Warmth	1.00
Competence	.37***	1.00
Self-centeredness	.26***	.11*	1.00
Submissiveness	.31***	.39***	.48***	1.00
SAAAS_S	.12*	.22***	.49***	.54***	1.00
SAAAS_C	.15**	.26***	.56***	.47***	.76***	1.00

Note. AAS = Asian American Stereotypes Scale, SAAAS = Scale of Anti-Asian American Stereotypes, SAAAS_S = social subscale, SAAAS_C = competence subscale.

p < .05. **p < .01. ***p < .001.

For comparison, we ran a CFA (with all participants) on the two-factor Scale of Anti-Asian American Stereotypes. With all items in the model, the fit was poor: SRMR = .10, RMSEA = .11, CFI = .77. Excluding the reverse-scored items produced a model with comparable fit to our scale: SRMR = .04, RMSEA = .07, CFI = .94. Each subscale had good internal consistency without the reverse-scored items (Social subscale: α = .92, Competence subscale: α = .91; dropping duplicate location: Social subscale: α = .91, Competence subscale: α = .90). Unlike the subscales of our scale, however, the subscales of the Scale of Anti-Asian American Stereotypes were highly correlated with each other (r = .76 for the composite scores; r = .84 for the factor scores).⁷

Discriminant validity

Correlations among the composite scores for the four subscales of the Asian American Stereotypes Scale and the two subscales of the Scale of Anti-Asian American Stereotypes indicated that the scales are related but not redundant (see Table 5). The highest correlation across scales was r = .56, between our Self-Centeredness subscale and the Scale of Anti-Asian American Stereotypes’ Competence subscale. For comparison, the Scale of Anti-Asian American Stereotypes’ subscales were more highly correlated with each other (r = .76) than with any of the AAS subscales, again indicating that our subscales are distinct from the Scale of Anti-Asian American Stereotypes.⁸

Predictive validity

Simple correlations between our subscales and the criterion variables suggest that each subscale has some predictive validity as an individual difference measure (see Table 6).

Table 6.

Study 2: Predictive validity: Correlations.

	Warmth	Competence	Self-centered	Submissive	SAAAS_S	SAAAS_C
Direct contact	.38***	.06	.40***	.26***	.26***	.33***
Extended contact	.33***	−.02	.45***	.29***	.33***	.37***
Racial resentment	.07	.07	.09⁺	.09⁺	.33***	.30***
System justification	.35***	.11*	.27***	.19***	.44***	.37***
Asian population estimate	.28***	.02	.35***	.29***	.34***	.33***
FT-Asian	.16**	.21***	−.09⁺	.00	−.13*	−.06
FT-White	.15**	.18***	.07	.09⁺	.14**	.12*
FT-Black	.16**	.10*	.02	.01	−.10⁺	−.07
FT-Latino	.18***	.13*	−.00	−.01	−.13*	−.10*
FT-legal immigrants	.08	.15**	.0	.07	−.05	.00
FT-undocumented immigrants	.24***	.03	.20***	.15**	−.03	−.04

Note. FT = feeling thermometer. SAAAS_S and SAAAS_C are the Social and Competence subscales, respectively, of the Scale of Anti-Asian American Stereotypes.

p < .10. *p < .05. **p < .01. ***p < .001.

As predicted, our Warmth and Competence subscales were positively correlated with feeling thermometer ratings of Asians. Our Self-Centeredness subscale was marginally negatively correlated with these ratings, but our Submissiveness subscale was not.

Contact with Asians (both direct and extended) was positively correlated with warmth ratings of Asians (as one would predict from intergroup contact theory), but it was also positively correlated with self-centeredness and submissiveness ratings.

Every subscale except the Competence subscale was positively correlated with estimates of the Asian population in the US. The highest correlation was with our Self-Centeredness subscale (r = .35).

As with the Asian feeling thermometer ratings, our Warmth subscale was positively correlated with feeling thermometer ratings of all groups except legal immigrants; our Competence subscale was positively correlated with ratings of all groups except undocumented immigrants. Although it was not related to feeling thermometer ratings for Asians, Blacks, or Latinos, our Submissiveness subscale was marginally positively correlated with White feeling thermometer ratings. And in contrast to Asian feeling thermometer ratings, our Self-Centeredness and Submissiveness subscales were significantly positively correlated with feeling thermometer ratings of undocumented immigrants—perhaps people who feel that Asians are either too dominant or too submissive prefer immigrants they perceive as less privileged than Asians.

Our Self-Centeredness and Submissiveness subscales were marginally positively correlated with racial resentment, and all four subscales were positively correlated with system justification.

The Scale of Anti-Asian American Stereotypes’ subscales were also correlated with most of these variables, but not all AAS and Scale of Anti-Asian American Stereotypes’ subscales correlated with the dependent variables in the same direction.

Multiple regression analyses

In multiple regression,⁹ controlling for age, gender, race (White vs. non-White, with White as the reference group), ideology, Scale of Anti-Asian American Stereotypes scores, and the other AAS subscales, the AAS subscales—especially the Warmth and Competence subscales—appeared to predict most of the criterion variables (see Tables 7 and 8).

Table 7.

Study 2: Predictive validity: Regression (criterion variables excluding feeling thermometers).

	Direct contact	Extended contact	Asian population estimation	System justification	Racial resentment
Age	−0.01 (0.00)	−0.01 (0.00)⁺	−0.06 (0.10)	0.00 (0.01)	0.01 (0.00)
Gender	−0.11 (0.10)	−0.04 (0.10)	1.91 (2.00)	−0.21 (0.14)	−0.07 (0.09)
Non-White	0.40 (0.11)***	0.45 (0.11)***	3.59 (2.27)	0.05 (0.15)	−0.16 (0.10)⁺
Ideology	0.05 (0.03)	0.03 (0.03)	1.17 (0.61)⁺	0.38 (0.04)***	0.29 (0.03)***
SAAAS-S	−0.07 (0.08)	0.02 (0.08)	3.07 (1.54)*	0.46 (0.10)***	0.09 (0.07)
SAAAS-C	0.21 (0.08)**	0.19 (0.08)*	2.21 (1.61)	0.05 (0.11)	0.18 (0.07)*
Warmth	0.20 (0.04)***	0.20 (0.04)***	3.30 (0.74)***	0.26 (0.05)***	−0.02 (0.03)
Competence	−0.09 (0.04)*	−0.17 (0.04)***	−2.53 (0.77)**	−0.01 (0.05)	0.01 (0.03)
Self-centeredness	0.08 (0.03)*	0.09 (0.03)**	0.95 (0.67)	0.00 (0.05)	−0.06 (0.03)*
Submissiveness	0.02 (0.04)	0.05 (0.04)	0.87 (0.74)	−0.16 (0.05)**	−0.06 (0.03)⁺
Intercept	1.33 (0.38)***	1.30 (0.37)***	−14.45 (7.66)⁺	1.118 (0.516)*	1.17 (0.33)***
N	358	358	364	364	370
R ²	.31	.36	.25	.45	.37

Note. SAAAS_S and SAAAS_C are the Social and Competence subscales, respectively, of the Scale of Anti-Asian American Stereotypes.

p < .10. *p < .05. **p < .01. ***p < .001.

Table 8.

Study 2: Predictive validity: Regression (feeling thermometers).

	FT-Asian	FT-White	FT-Black	FT-Latino	FT-legal immigrants	FT-undoc. immigrants
Age	0.07 (0.11)	0.14 (0.12)	0.01 (0.12)	−0.09 (0.11)	0.05 (0.11)	−0.08 (0.14)
Gender	0.28 (2.16)	−1.12 (2.46)	5.26 (2.39)*	3.13 (2.31)	1.72 (2.19)	−0.61 (2.91)
Non-White	6.93 (2.44)**	−7.59 (2.80)**	5.48 (2.71)*	6.52 (2.63)*	5.83 (2.49)*	8.26 (3.29)*
Ideology	−2.21 (0.66)***	3.40 (0.75)***	−2.97 (0.73)***	−2.61 (0.70)***	−3.07 (0.67)***	−6.11 (0.88)***
SAAAS-S	−0.84 (1.66)	2.06 (1.89)	0.63 (1.83)	0.92 (1.79)	−0.05 (1.68)	1.09 (2.24)
SAAAS-C	0.12 (1.74)	−0.41 (2.00)	−2.34 (1.91)	−3.34 (1.85)⁺	0.61 (1.77)	−4.80 (2.33)*
Warmth	1.93 (0.80)*	0.81 (0.90)	2.59 (0.89)**	2.84 (0.84)***	0.71 (0.82)	5.40 (1.08)***
Competence	2.68 (0.85)**	2.09 (0.94)*	1.23 (0.91)	1.43 (0.88)	1.48 (0.84)⁺	−1.77 (1.11)
Self-centeredness	−1.31 (0.74)⁺	0.14 (0.83)	1.00 (0.81)	0.62 (0.78)	−0.10 (0.75)	3.16 (1.00)**
Submissiveness	−0.27 (0.80)	−0.86 (0.92)	−0.51 (0.89)	−0.74 (0.86)	0.36 (0.81)	2.08 (1.08)⁺
Intercept	59.14 (8.22)***	37.71 (9.35)***	55.32 (9.05)***	63.76 (8.79)***	63.86 (8.34)***	50.54 (11.07)***
N	354	359	352	354	359	357
R ²	.13	.13	.12	.12	.09	.24

Note. FT = feeling thermometer. SAAAS_S and SAAAS_C are the Social and Competence subscales, respectively, of the Scale of Anti-Asian American Stereotypes.

p < .10. *p < .05. **p < .01. ***p < .001.

Higher warmth and competence ratings predicted, and lower self-centeredness ratings marginally predicted, higher feeling thermometer ratings of Asians. Higher warmth and self-centeredness ratings and lower competence ratings predicted higher levels of both direct and extended contact with Asians or Asian Americans. Higher warmth ratings and lower competence ratings predicted higher estimates of the Asian population in the US.

Higher warmth ratings predicted higher feeling thermometer ratings of Blacks, Latinos, and undocumented immigrants as well as Asians, while higher competence ratings predicted higher feeling thermometer ratings of Whites as well as Asians. Higher self-centeredness ratings predicted, and higher submissiveness ratings marginally predicted, higher feeling thermometer ratings of undocumented immigrants. Contrary to our expectations, relationships with the AAS subscales were not consistently stronger for the Asian feeling thermometer than for the other feeling thermometers.

Lower self-centeredness ratings and marginally lower submissiveness ratings were associated with higher racial resentment, as would be expected if Asian American stereotypes were used to contrast them with African Americans. Lower submissiveness ratings and higher warmth ratings predicted higher system justification, suggesting that some positive Asian stereotypes might be used to justify the status quo.

While the Scale of Anti-Asian American Stereotypes’ subscales also showed incremental predictive validity on some of the criterion variables when controlling for the AAS, the pattern of significant regression coefficients for the Scale of Anti-Asian American Stereotypes differed from that of the AAS subscales (see Tables 7 and 8).

Overall, the AAS subscales showed incremental predictive validity even when controlling for the Scale of Anti-Asian American Stereotypes, the existing scale that we included in this study.

Discussion

Results from Study 2 show that the Asian American Stereotypes Scale is psychometrically valid and distinct from the Scale of Anti-Asian American Stereotypes. The Asian American Stereotypes Scale captures the four dimensions we identified in Study 1: warmth, competence, self-centeredness, and submissiveness, with items loading onto their expected factors, and acceptable fit of the four-factor model. Scores on specific subscales are correlated with each other and with the Scale of Anti-Asian American Stereotypes’ subscales, as can be expected from measures that all capture stereotypes of Asian Americans. However, correlations between the AAS subscales and the Scale of Anti-Asian American Stereotypes’ subscales were lower than those between the Scale of Anti-Asian American Stereotypes’ subscales, providing evidence that these scales measure related but unquestionably distinct constructs.

Study 2 also shows that the four subscales of the Asian American Stereotypes Scale are related to variables that the scale should be related to (i.e., racial attitudes, intergroup contact, and system justification), though not always in the direction we initially predicted. Endorsement of positive stereotypes predicts higher feeling thermometer ratings of Asians, but endorsement of both positive (higher warmth) and negative (lower competence, higher self-centeredness) stereotypes is associated with intergroup contact and overestimates of the Asian American population in the US. Consistent with our predictions, submissiveness ratings predicted system justification, but inconsistent with our predictions, so did warmth ratings. These findings suggest that something more complex than positive and negative warmth and competence stereotypes might underlie attitudes toward Asian Americans.

In support of the idea that Americans hold similar attitudes toward all minority groups, we generally found a positive relationship between warmth ratings of Asian Americans and attitudes toward other non-White groups. However, in support of the idea that model minority stereotypes contrast Asian and African Americans (Wu, 2014, p. 149), higher self-centeredness ratings of Asian Americans were associated with lower racial resentment (i.e., the idea that African Americans are undeserving or overly demanding).

Nonetheless, none of these relationships are so strong as to suggest conceptual overlap. In other words, the Asian American Stereotypes Scale shows discriminant and predictive validity. Additionally, the AAS subscales show incremental validity in predicting these measures when controlling for each other and for the Scale of Anti-Asian American Stereotypes. It is important theoretically to note that, even though our Warmth and Competence subscales showed the greatest predictive validity, they are not equivalent to the warmth and competence dimensions of the stereotype content model; instead, they involve warmth/morality and competence/skill controlling for the more agency-related traits in our self-centeredness and submissiveness dimensions. To the extent that the self-centeredness and submissiveness dimensions predict racial attitudes, they indicate that agency-related stereotypes (separate from competence) shape or perpetuate these attitudes; to the extent that the warmth and competence dimensions predict racial attitudes, they indicate where agency-related stereotypes play a smaller role. Thus, our scale shows promise as a measure of Asian American stereotypes.

General Discussion

We have developed and validated a new, four-factor model of Asian American stereotypes drawing on theories about both gender and racial prejudice. In addition to warmth and competence dimensions similar to but distinct from those in the stereotype content model, this model identifies two new dimensions of Asian American descriptive stereotypes: self-centeredness and submissiveness. This factor structure distinguishes between traits that people consider more typical of Asian Americans than of White Americans (i.e., competence) and traits that people consider more typical of White Americans than of Asian Americans (i.e., self-centeredness). Thus, this finding splits the usual conception of agency in person perception research (for a description of agency as related to self-profitability, see e.g., Abele & Wojciszke, 2007). Furthermore, not only are Asian Americans descriptively not expected to be self-centered, but they are uniquely proscribed from certain self-centered traits, suggesting a potential mechanism through which counterstereotypical Asian American individuals might experience backlash. Thus, future research should examine whether counterstereotypical individuals elicit high self-centeredness ratings and whether those ratings predict negative attitudes or behaviors toward them.

In addition to identifying these dimensions, we developed a new scale that measures people’s perceptions of Asian Americans on these dimensions. As with any trait-rating studies or scale development endeavors, our results are limited by the range of traits we selected for our original study and for our final scale. However, we attempted to draw from a variety of existing person perception, as well as race and gender stereotype, studies to cover a reasonable scope of traits in our original pool. Our final scale attempts to balance reliability and validity with the need for a relatively short scale for use in future research. As is evidenced in Study 2, the Asian American Stereotypes Scale shows construct validity, discriminant validity, and predictive validity for a range of racial attitude-related measures, including incremental predictive validity over the Scale of Anti-Asian American Stereotypes, a contemporary scale measuring negative Asian American stereotypes.

Because of Asian Americans’ specific historical and cultural context, the dimensionality of the Asian American Stereotypes Scale may not apply to other target groups, such as African Americans or Asians who do not live in the United States. It does, however, appear to be consistent for White and non-White perceivers, suggesting that these dimensions are widely shared among American perceivers and that our scale can be used for both White and non-White participants. Future research might examine whether Asian Americans perceive their own group along these dimensions and whether the dimensions also apply to Asian American subgroups.

Additionally, because the Asian American Stereotypes Scale was developed to be able to study responses to stereotypical and counterstereotypical Asian American individuals, future research should examine whether ratings of Asian American individuals tend to mirror ratings of Asian Americans as a group, and whether ratings of counterstereotypical Asian American individuals deviate in characteristic ways from group ratings. For example, if an Asian American individual displays behaviors associated with self-centeredness, will that individual be perceived as more self-centered than a White person displaying the same behaviors? And can those ratings predict disliking of that individual, as Berdahl and Min (2012) described, as well as downstream prosocial or discriminatory behaviors toward that individual? Having the Asian American Stereotypes Scale will allow researchers to systematically measure backlash in the perception of individuals as a potential mechanism to explain observed disparities in outcomes between Asian and White Americans. More broadly, the model we have developed allows us to predict and examine the unique impacts of Asian American stereotypes in a way that more general models do not.

Supplemental Material

asian_stereotypes_supplemental_materials_revised_2020-05-21 – Supplemental material for A four-dimensional model of Asian American stereotypes

Supplemental material, asian_stereotypes_supplemental_materials_revised_2020-05-21 for A four-dimensional model of Asian American stereotypes by Wen Bu and Eugene Borgida in Group Processes & Intergroup Relations

Footnotes

Acknowledgements

The authors thank Chris Federico for his comments on an earlier draft of this paper and for his feedback on the data analytic approach.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Wen Bu

Supplemental material

Supplemental material for this article is available online.

Notes

References

Abele

A. E.

Uchronski

Suitner

Wojciszke

(2008). Towards an operationalization of the fundamental dimensions of agency and communion: Trait content ratings in five countries considering valence and frequency of word occurrence. European Journal of Social Psychology, 38, 1202–1217. https://doi.org/10.1002/ejsp.575

Abele

A. E.

Wojciszke

(2007). Agency and communion from the perspective of self versus others. Journal of Personality and Social Psychology, 93, 751–763. https://doi.org/0.1037/0022-3514.93.5.751

Bai

G. C. C.

Yan

(2019). Does virtue lead to status? Testing the moral virtue theory of status attainment. Journal of Personality and Social Psychology. Advance online publication. https://doi.org/10.1037/pspi0000192

Berdahl

J. L.

Min

J.-A.

(2012). Prescriptive stereotypes and workplace consequences for East Asians in North America. Cultural Diversity and Ethnic Minority Psychology, 18, 141–152. https://doi.org/10.1037/a0027692

Brandt

M. J.

Crawford

J. T.

(2019). Studying a heterogeneous array of target groups can help us understand prejudice. Current Directions in Psychological Science, 28, 292–298. https://doi.org/10.1177/0963721419830382

Chmielewski

Kucker

S. C.

(2019). An MTurk crisis? Shifts in data quality and the impact on study results. Social Psychological and Personality Science. Advance online publication. https://doi.org/10.1177/1948550619875149

Dovidio

J. F.

Brigham

J. C.

Johnson

B. T.

Gaertner

S. L.

(1996). Stereotyping, prejudice, and discrimination: Another look. In Macrae

C. N.

Stangor

Hewstone

(Eds.), Stereotypes and stereotyping (pp. 276–319). Guilford Press.

Dovidio

J. F.

Gaertner

S. L.

(2010). Intergroup bias. In Fiske

S. T.

Gilbert

Lindzey

(Eds.), Handbook of social psychology (5th ed., Vol. 2, pp. 1024–1083). Wiley. https://doi.org/10.1002/9780470561119.socpsy002029

Eagly

A. H.

Karau

S. J.

(2002). Role congruity theory of prejudice toward female leaders. Psychological Review, 109, 573–598. https://doi.org/10.1037//0033-295X.109.3.573

10.

Feldman

Huddy

(2005). Racial resentment and White opposition to race-conscious programs: Principles or prejudice? American Journal of Political Science, 49, 168–183. https://doi.org/10.1111/j.0092-5853.2005.00117.x

11.

Fiske

S. T.

(2018). Stereotype content: Warmth and competence endure. Current Directions in Psychological Science, 27, 67–73. https://doi.org/10.1177/0963721417738825

12.

Fiske

S. T.

Cuddy

A. J. C.

Glick

(2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878–902. https://doi.org/0.1037//0022-3514.82.6.878

13.

Fiske

S. T.

Cuddy

A. C.

Glick

(1999). (Dis)respecting versus (dis) liking: Status and interdependence predict ambivalent stereotypes of competence and warmth. Journal of Social Issues, 55, 473–489. https://doi.org/10.1111/0022-4537.00128

14.

Jackson

J. W.

(2001). Attitude toward Asian Americans: Theory and measurement. Journal of Applied Social Psychology, 31, 1553–1581. https://doi.org/10.1111/j.1559-1816.2001.tb02742.x

15.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. https://doi.org/10.1080/10705519909540118

16.

Kan

I. P.

Drummey

A. B.

(2018). Do imposters threaten data quality? An examination of worker misrepresentation and downstream consequences in Amazon’s Mechanical Turk workforce. Computers in Human Behavior, 83, 243–253. https://doi.org/10.1016/j.chb.2018.02.005

17.

Kay

A. C.

Jost

J. T.

(2003). Complementary justice: Effects of “poor but happy” and “poor but honest” stereotype exemplars on system justification and implicit activation of the justice motive. Journal of Personality and Social Psychology, 85, 823–837. https://doi.org/10.1037/0022-3514.85.5.823

18.

Kinder

D. R.

Sanders

L. M.

(1996). Divided by color: Racial politics and democratic ideals. University of Chicago Press.

19.

Koch

Imhoff

Dotsch

Unkelbach

Alves

(2016). The ABC of stereotypes about groups: Agency/socio-economic success, conservative–progressive beliefs, and communion. Journal of Personality and Social Psychology, 110, 675–709. https://doi.org/10.1037/pspa0000046

20.

Koenig

A. M.

Eagly

A. H.

(2014). Evidence for the social role theory of stereotype content: Observations of groups’ roles shape stereotypes. Journal of Personality and Social Psychology, 107, 371–392. https://doi.org/10.1037/a0037215

21.

Leach

C. W.

Ellemers

Barreto

(2007). Group virtue: The importance of morality (vs. competence and sociability) in the positive evaluation of in-groups. Journal of Personality and Social Psychology, 93, 234–249. https://doi.org/10.1037/0022-3514.93.2.234

22.

Lin

M. H.

Kwan

V. S. Y.

Cheung

Fiske

S. T.

(2005). Stereotype content model explains prejudice for an envied outgroup: Scale of Anti-Asian American Stereotypes. Personality and Social Psychology Bulletin, 31, 34–47. https://doi.org/10.1177/0146167204271320

23.

Matherly

(2019). A panel for lemons? Positivity bias, reputation systems and data quality on MTurk. European Journal of Marketing, 53, 195–223. https://doi.org/10.1108/EJM-07-2017-0491

24.

Morton

L. C.

Layton

R. L.

Muraven

(2019). Social Modeling of Exercise Inventory: A measure of exercise-specific social influence. Journal of Applied Social Psychology, 49, 178–191. https://doi.org/10.1111/jasp.12575

25.

Muthén

L. K.

Muthén

B. O.

(1998–2018). Mplus [Computer software]. Muthén & Muthén.

26.

Phelan

J. E.

Rudman

L. A.

(2010). Reactions to ethnic deviance: The role of backlash in racial stereotype maintenance. Journal of Personality and Social Psychology, 99, 265–281. https://doi.org/10.1037/a0018304

27.

R Core Team. (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/

28.

Reid

Sparks

Jessop

D. C.

(2018). The effect of self-identity alongside perceived importance within the theory of planned behaviour. European Journal of Social Psychology, 48, 883–889. https://doi.org/10.1002/ejsp.2373

29.

Rosette

A. S.

Koval

C. Z.

Livingston

(2016). Race matters for women leaders: Intersectional effects on agentic deficiencies and penalties. The Leadership Quarterly, 27, 429–445. https://doi.org/10.1016/j.leaqua.2016.01.008

30.

Rudman

L. A.

Moss-Racusin

C. A.

Phelan

J. E.

Nauts

(2012). Status incongruity and backlash effects: Defending the gender hierarchy motivates prejudice against female leaders. Journal of Experimental Social Psychology, 48, 165–179. https://doi.org/10.1016/j.jesp.2011.10.008

31.

Stathi

Crisp

R. J.

Hogg

M. A.

(2011). Imagining intergroup contact enables member-to-group generalization. Group Dynamics: Theory, Research, and Practice, 15, 275–284. https://doi.org/10.1037/a0023752

32.

Shore

L. M.

Strauss

Shore

T. H.

Tram

Whiteley

Ikeda-Muromachi

(2010). Leadership perceptions as a function of race–occupation fit: The case of Asian Americans. Journal of Applied Psychology, 95, 902–919. https://doi.org/10.1037/a0019501

33.

Wiggins

J. S.

Trapnell

Phillips

(1988). Psychometric and geometric characteristics of the revised Interpersonal Adjective Scales (IAS-R). Multivariate Behavioral Research, 23, 517–530. https://doi.org/10.1207/s15327906mbr2304_8

34.

E. D.

(2014). The color of success: Asian Americans and the origins of the model minority. Princeton University Press.

35.

Yamamoto

Maeder

E. M.

(2019). Creating the punishment orientation questionnaire: An item response theory approach. Personality and Social Psychology Bulletin, 45, 1283–1294. https://doi.org/10.1177/0146167218818485

36.

Yzerbyt

Demoulin

(2010). Intergroup relations. In Fiske

S. T.

Gilbert

Lindzey

(Eds.), Handbook of social psychology (5th ed., Vol. 2, pp. 1024–1083). Wiley. https://doi.org/10.1002/9780470561119.socpsy002028

37.

Zou

L. X.

Cheryan

(2017). Two axes of subordination: A new model of racial position. Journal of Personality and Social Psychology, 112, 696–717. https://doi.org/10.1037/pspa0000080

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.40 MB