An Investigation of Positive and Negative Contact As Predictors of Intergroup Attitudes in the United States,Hong Kong,and Thailand

Abstract

Contact researchers have overlooked (a) the mechanisms that explain the association between negative contact and prejudice, (b) the effects of positive and negative contact on outcomes beyond prejudice, and (c) the importance of testing contact effects cross-culturally. In the present article, we addressed these gaps in the literature by drawing on data from White Americans (N = 207; Study 1), Hong Kong Chinese (N = 145; Study 2), and Buddhist Thais (N = 161; Study 3). Specifically, we examined positive and negative contact as predictors of old-fashioned and modern prejudice toward, and negative metaperceptions about, Black Americans, Mainland Chinese, and Muslim Thais, respectively. We also tested intergroup anxiety as a mediator of the associations between positive and negative contact, and all intergroup outcomes. Across three studies, positive contact predicted reduced intergroup anxiety, prejudice, and negative metaperceptions, while negative contact predicted increased intergroup anxiety, prejudice, and negative metaperceptions. Negative contact, however, was the more consistent predictor of intergroup attitudes. Intergroup anxiety emerged as a robust mediator of the relationships between both types of contact and all intergroup outcomes. We thus present the first test of a model of positive and negative contact that holds across both Western and non-Western contexts.

Keywords

contact hypothesis positive contact negative contact prejudice metaperception

Positive intergroup contact is an important factor in improving intergroup relations (see Pettigrew, Tropp, Wagner, & Christ, 2011, for a review). Despite the enthusiasm with which researchers have tested the association between positive contact and intergroup attitudes in the past six decades, a number of questions remain unexplored. First, although research suggests a strong association between negative contact and increased prejudice (Barlow et al., 2012), exactly how negative contact works to increase prejudice is still unknown. Second, contact researchers narrowly focus on prejudice as a dependent variable (Dixon, Durrheim, & Tredoux, 2005). While prejudice reduction is important, a constricted focus means that the effect of contact on other relevant intergroup outcomes may be overlooked. Finally, the bulk of contact studies focus on Western samples (Pettigrew & Tropp, 2006), leading to uncertainty about the generalizability of contact effects. In the present article, we extend the literature on intergroup contact by drawing on data from the United States, Hong Kong, and Thailand. In these contexts, we test intergroup anxiety as a mediator of the relationships between both types of contact and intergroup attitudes. We also introduce negative metaperceptions as a novel variable dependent on both positive and negative contact.

Negative Intergroup Contact

Gordon Allport’s (1954) contact hypothesis is one of the most influential theoretical frameworks that explain the relationship between contact and intergroup attitudes. Decades of research indicate that positive contact promotes prejudice reduction through the reduction of intergroup anxiety (see Pettigrew & Tropp, 2006, 2008, for meta-analyses). Specifically, positive contact reduces anxiety that individuals experience in response to outgroups, leading them to adopt more favorable attitudes toward outgroup members.

While positive contact has received substantive empirical attention, negative contact has remained largely unexplored. In their meta-analysis, Pettigrew and Tropp (2006) found that less than 5% of 713 independent samples examined negative contact. The exclusion of negative contact is not surprising considering that the contact hypothesis was developed based on the assumption that most types of contact did not reduce prejudice. As such, Allport (1954) encouraged his successors to focus on the types of contact that result in prejudice reduction.

Nevertheless, in a seminal review, Pettigrew (2008) argued that negative contact deserves to become the central focus of contact research due to its potential to disrupt the beneficial effects of positive contact. Consistent with this argument, Barlow et al. (2012) found that for White Americans, positive contact was a robust predictor of prejudice toward Black Americans when negative contact was not taken into account. When negative contact was included in the analyses, however, positive contact was a comparatively weak predictor of reduced prejudice. Negative contact, by contrast, emerged as a more robust and consistent predictor of prejudice. These findings demonstrated that negative contact can increase prejudice and weaken the benefits of positive contact.

Given that it is a new area of study, little is known about the mechanisms that might explain why negative contact is linked to intergroup negativity. The present article, therefore, seeks to test intergroup anxiety as a mediator of the association between negative contact and prejudice. The proposed mediational path is intuitive—theoretically, negative contact should confirm or enhance intergroup anxieties around contact and consequently increase prejudice (Barlow et al., 2012; Paolini, Harwood, & Rubin, 2010; Pettigrew, 2008). This path, however, has never been empirically tested. As such, we are the first to investigate whether negative contact predicts increased prejudice through increased intergroup anxiety.

Prejudice Problematic: Negative Metaperceptions as a Novel Dependent Variable

In addition to elucidating how negative contact works to increase prejudice, we examine positive and negative contact as predictors of an outcome other than prejudice. Dixon and colleagues (2005) recently highlighted “the restrictions of prejudice problematic” phenomenon, positing that the vast majority of contact research uses scores on prejudice scales as sole indicators of the failure or success of contact. The authors argued that indices of prejudice alone may not reflect individuals’ willingness to engage in collective action, or many other important intergroup attitudes and behaviors. In line with this argument, critics of contact research have called for studies examining intergroup outcomes beyond prejudice (Dixon, Levine, Reicher, & Durrheim, 2012; Jackman & Crane, 1986). In this article, we answer this call by including a novel measure of negative metaperceptions and model it as dependent on both positive and negative contact.

Negative metaperceptions refer to an individual’s beliefs that their group is viewed negatively by outgroup members (Vorauer, Main, & O’Connell, 1998). Common metaperceptions held by majority group members include thinking that minorities perceive them to be privileged and prejudiced (Vorauer, Hunter, Main, & Roy, 2000). Negative metaperceptions are important to consider, as they predict whether majorities engage in behaviors that maintain their power and sense of position (Plant & Butz, 2006; Vorauer, 2003). For instance, majorities tend to exhibit a wide range of discriminatory behaviors toward outgroup members (e.g., reduced eye contact and increased hostile reactions to criticism) when they believe that the outgroup evaluates them negatively (see Vorauer, 2006, for a review). Hence, identifying factors that alter negative metaperceptions will have implications for efforts aimed at improving intergroup relations.

In this article, we argue that contact should play a critical role in influencing metaperceptions. This argument stems from two independent lines of research. First, contact is a robust predictor of intergroup anxiety (see Pettigrew & Tropp, 2006, 2008). Positive contact allows individuals to gain critical insights about outgroup norms, values, and behavioral scripts, eliminating the novelty attached to intergroup exchange and reducing intergroup anxiety (Stephan & Stephan, 1985). In contrast, negative contact prompts individuals to focus on the differences between their own and the outgroup’s perspectives and worldviews, and thus elevate intergroup anxiety (Paolini et al., 2010). Second, negative metaperceptions vary as a function of intergroup anxiety (Finchilescu, 2005; Laher & Finchilescu, 2010; Plant, 2004). In particular, the more anxiety individuals experience in intergroup contexts, the more they hold beliefs that the outgroup evaluates them negatively. Given that contact is reliably associated with intergroup anxiety and that intergroup anxiety is reliably associated with negative metaperceptions, we propose that contact should predict intergroup anxiety, which, in turn, should predict negative metaperceptions (cf. Ames, 2004; Vorauer, 2006). Specifically, reduced anxiety resulting from positive contact should subsequently decrease one’s concerns about being evaluated negatively. Conversely, increased anxiety caused by negative contact should further elevate evaluative concerns.

Intergroup Contact Research in Non-Western Nations

Contact research has been extended in new directions in recent decades. For instance, Pettigrew and Tropp’s (2006) meta-analysis revealed that contact is reliably associated with prejudice reduction, irrespective of whether the target outgroup is a racial, religious, or regional minority. Unfortunately, contact research in non-Western nations is still extremely scarce. In fact, Pettigrew and Tropp found that only 10% of the studies sampled were conducted in non-Western nations (with 5% in Africa, 3% in Israel, 1% in Latin America, and 1% in Asia). The lack of contact research beyond Western borders is problematic, with many scholars implicitly assuming that contact effects are universal in the absence of empirical confirmation. Thus, our final goal is to make a contribution by testing the proposed model in both Western (i.e., the United States) and non-Western contexts (i.e., Hong Kong and Thailand).

To our knowledge, there have been no contact studies conducted in Thailand, and only one in Hong Kong. The latter, conducted by King, Winter, and Webster (2009), revealed that positive contact heterosexuals had with transgendered people predicted reduced transprejudice, offering some support to the application of the contact hypothesis in Hong Kong (and Asia more generally). However, this study did not test several aspects of our proposed model. Like much past contact research, it did not consider the roles of negative contact and intergroup anxiety, or examine the effect of contact on negative metaperceptions.

In the present article, we consider White Americans’ attitudes toward Black Americans (Study 1), Hong Kong Chinese people’s attitudes toward Mainland Chinese (Study 2), and Buddhist Thais’ attitudes toward Muslim Thais (Study 3). In line with the literature reviewed above, we do not expect cultural differences a priori. In all contexts, we hypothesize that positive contact will predict reduced prejudice and negative metaperceptions, and conversely that negative contact will predict increased prejudice and negative metaperceptions. We also predict that negative contact will emerge as a more consistent predictor of intergroup attitudes than positive contact. Finally, we hypothesize that intergroup anxiety will mediate both positive and negative contact effects.

Study 1

Study 1 focused on the intergroup dynamic of White and Black Americans, who make up 78.1% and 13.1% of the U.S. population, respectively (United States Census Bureau, 2011). Black Americans presently face severe disadvantage in comparison with White Americans in terms of education, housing, and health (Harris, 2010). Research indicates that prejudice is a substantive causal factor, with most Black Americans experiencing discrimination at the hands of White Americans (Dovidio & Gaertner, 2000; Gallup Organization, 2015). White Americans are deeply concerned that Black Americans view them as prejudiced (Bergsieker, Shelton, & Richeson, 2010). Thus, contemporary White–Black relations are characterized by anxiety, prejudice, and negative metaperceptions, and as such is an ideal context in which to first test our model.

Method

Participants and measures

A community sample of 207 White Americans (49% female; M_age = 25.00, SD = 8.04) was recruited from an online scientific survey pool (www.socialsci.com). Positive contact and negative contact were measured via single items adapted from Barlow and colleagues (2012). Intergroup anxiety was assessed via seven items adapted from Stephan and Stephan (1985). Old-fashioned prejudice was measured using five items adapted from Walker (1994). Modern prejudice was assessed with a five-item scale adapted from McConahay (1986). Six items adapted from Vorauer and colleagues (1998) were used to assess the extent to which participants believed that White Americans are viewed negatively by Black Americans (i.e., negative metaperceptions). The appendix shows all the survey scales included in this study.¹

Results

To avoid any sample confounds, age and gender were included as covariates in all analyses. Hierarchical regressions were first conducted to determine whether positive contact directly predicted the dependent variables without controlling for negative contact. Covariates were entered at Step 1, followed by positive contact at Step 2. Further hierarchical regressions were then performed to test the direct effects of positive and negative contact, and to determine whether the effects of positive contact are weakened after including negative contact in the model. Covariates were entered at Step 1, followed by positive and negative contact at Step 2. We then tested whether the associations between positive and negative contact and the dependent variables were mediated by intergroup anxiety using Hayes and Preacher’s (2014) bootstrapping procedures. A total of 5,000 bootstrap samples were created to estimate bias-corrected standard errors and 95% percentile confidence intervals for the indirect effects. Descriptive statistics and bivariate correlations can be found in Table 1.

Table 1.

Descriptive Information and Correlations Among Focal Variables (Study 1).

Focal variables	M	SD	1	2	3	4	5	6	7	8
1. Age	25.00	8.04	—
2. Gender (1 male, 2 female)	1.51	0.50	.06	—
3. Positive contact	3.93	0.64	.05	.03	—
4. Negative contact	2.20	0.64	−.15*	.01	−.35***	—
5. Intergroup anxiety	1.45	0.63	−.07	−.10	−.30***	.35***	(.92)
6. Old-fashioned prejudice	1.27	0.46	.04	−.12	−.28***	.29***	.62***	(.74)
7. Modern prejudice	2.58	0.67	−.06	−.18**	−.29***	.30***	.51***	.47***	(.66)
8. Negative metaperceptions	2.99	0.58	.07	.11	−.17*	.19**	.35***	.29***	.33***	(.72)

Note. Internal reliabilities appear in parentheses.

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

Without the inclusion of negative contact in the analyses, positive contact predicted decreased old-fashioned prejudice, β = −.28, t(203) = −4.14, p < .001, modern prejudice, β = −.29, t(203) = −4.36, p < .001, and negative metaperceptions, β = −.17, t(203) = −2.50, p = .013. However, after negative contact was included in the analyses, positive contact predicted decreased old-fashioned prejudice, β = −.20, t(202) = −4.14, p = .005, and modern prejudice, β = −.21, t(202) = −3.25, p = .003, but not negative metaperceptions, β = −.12, t(202) = −1.60, p = .112. Negative contact, by contrast, predicted increased old-fashioned prejudice, β = .23, t(202) = 3.27, p < .001, modern prejudice, β = .23, t(202) = 3.25, p < .001, and negative metaperceptions, β = .16, t(202) = 2.20, p = .029.

The associations between positive contact and old-fashioned prejudice (B = −.08, B_SE = .04, 95% confidence intervals [CI] = [−.18, −.02]), as well as modern prejudice (B = −.09, B_SE = .04, 95% CI = [−.18, −.03]), were mediated by reduced intergroup anxiety. Although there was no direct association between positive contact and negative metaperceptions, positive contact predicted decreased negative metaperceptions indirectly through reduced intergroup anxiety (B = −.06, B_SE = .03, 95% CI = [−.13, −.02]). The associations between negative contact and old-fashioned prejudice (B = .12, B_SE = .04, 95% CI = [.05, .22]), modern prejudice (B = .12, B_SE = .04, 95% CI = [.05, .23]), and negative metaperceptions (B = .08, B_SE = .03, 95% CI = [.03, .17]) were mediated by increased intergroup anxiety.

Discussion

In line with prior research (e.g., Pettigrew & Tropp, 2006, 2008), results demonstrated that when negative contact was not considered, positive contact was a robust predictor of intergroup attitudes. The inclusion of negative contact, however, substantially weakened the effect of positive contact on our outcome variables. While the amount of positive contact participants had with Black Americans predicted decreased old-fashioned and modern prejudice, it was not directly related to negative metaperceptions. Negative contact, by contrast, appeared to be a more stable predictor of intergroup attitudes. The more negative contact participants reported, the higher old-fashioned prejudice, modern prejudice, and negative metaperceptions they expressed. These results were consistent with Barlow et al. (2012), who suggested that negative contact has the potential to weaken the benefits of positive contact, and is a more consistent predictor of intergroup attitudes than positive contact. Additionally, in line with our theorizing, intergroup anxiety mediated all relationships, with positive contact predicting improved attitudes through reduced anxiety and negative contact predicting increased intergroup negativity through increased anxiety.

Study 2

In Study 2, we turned our attention to the conflict between Hong Kong and Mainland Chinese residing in Hong Kong. Hong Kong is an island that was initially governed by China, but was occupied by the British in 1841 after they claimed victories over China in the First Opium War. Hong Kong remained a colony of the United Kingdom until its sovereignty was returned to Mainland China in 1997. Since then, Hong Kong has been governed by Mainland China under the policy of “one country, two systems,” where Hong Kong retains a high degree of political autonomy. In recent years, tensions between Hong Kong and Mainland Chinese have emerged. Hong Kong has seen an influx of investors and immigrants from Mainland China as a result of easier access to the country (Community Legal Information Centre, 2009). This has increased the daily cost of living as well as overall property prices, with Hong Kong Chinese struggling to adapt to the change (Time World, 2012). Consequently, prejudice toward Mainlanders is now prevalent among Hong Kong Chinese (Cheuk-Lam, 2013). Negative metaperceptions are also widespread, as Mainlanders tend to believe that Hong Kong Chinese are privileged and prejudiced while Hong Kong Chinese are aware of and concerned about these beliefs (Cheuk-Lam, 2013).

Method

Participants and measures

A web link to an online survey was posted on two major online discussion forums in Hong Kong (www.discuss.com.hk and www.uwants.com). A total of 145 Hong Kong Chinese completed the survey (41% female; M_age = 25.92, SD = 8.74). The measures used in this study were identical to those in Study 1, except that the target outgroup was changed from “Black Americans” to “Mainlanders in Hong Kong,” and that the target ingroup was changed from “White Americans” to “Hong Kong people.” The questionnaire was also translated into Traditional Chinese language.

Results and Discussion

A small percentage of participants (<5%) did not complete all the dependent variable measures. Thus, listwise deletion was used to deal with missing data. An identical set of analyses used in Study 1 was performed to test the hypotheses. Scale reliabilities, descriptive statistics, and intercorrelations can be seen in Table 2.

Table 2.

Descriptive Information and Correlations Among Focal Variables (Study 2).

Focal variables	M	SD	1	2	3	4	5	6	7	8
1. Age	25.92	8.74	—
2. Gender (1 male, 2 female)	1.59	0.49	.03	—
3. Positive contact	3.28	0.90	.12	−.19*	—
4. Negative contact	2.69	0.73	.07	.03	−.44***	—
5. Intergroup anxiety	2.34	0.79	−.09	−.04	−.40***	.51***	(.94)
6. Old-fashioned prejudice	2.59	0.71	−.08	−.02	−.39***	.44***	.58***	(.79)
7. Modern prejudice	3.68	0.61	−.14	−.02	−.21**	.34***	.36***	.60***	(.70)
9. Negative metaperceptions	2.95	0.56	−.09	−.18*	−.19*	.28***	.52***	.37***	.20*	(.69)

Note. Internal reliabilities appear in parentheses.

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

When negative contact was not included in the analyses, positive contact reliably predicted decreased old-fashioned prejudice, β = −.41, t(141) = −5.12, p < .001, modern prejudice, β = −.21, t(141) = −42.51, p = .013, and negative metaperceptions, β = −.23, t(140) = −2.73, p = .007. After including negative contact in the analyses, positive contact directly predicted decreased old-fashioned prejudice, β = −.25, t(140) = 4.08, p = .004, but not modern prejudice, β = −.05, t(140) = −0.05, p = .550, or negative metaperceptions, β = −.12, t(139) = −1.26, p = .210. In contrast, negative contact emerged as a reliable predictor of increased old-fashioned prejudice, β = .34, t(140) = 4.08, p < .001, modern prejudice, β = .33, t(140) = 3.77, p < .001, and negative metaperceptions, β = .24, t(139) = 2.64, p = .009. As in Study 1, the associations between and old-fashioned prejudice (B = −.07, B_SE = .03, 95% CI = [−.15, −.02]), modern prejudice (B = −.03, B_SE = .02, 95% CI = [−.09, −.01]), and negative metaperceptions (B = −.07, B_SE = .03, 95% CI = [−.13, −.02]) were mediated by reduced intergroup anxiety. The associations between negative contact and old-fashioned prejudice (B = .18, B_SE = .05, 95% CI = [.09, .28]), modern prejudice (B = .08, B_SE = .04, 95% CI = [.01, .17]), and negative metaperceptions (B = .16, B_SE = .04, 95% CI = [.09, .25]) were mediated by increased intergroup anxiety.

In sum, Study 2 replicated the main findings of Study 1 in a very different intergroup setting. Positive contact was a reliable predictor of improved intergroup attitudes via reduced intergroup anxiety when negative contact was not included in the analyses. However, the inclusion of negative contact weakened the relationship between positive contact and intergroup attitudes. While positive contact Hong Kong Chinese had with Mainland Chinese was directly associated with lower levels of old-fashioned prejudice, it was not directly associated with modern prejudice or negative metaperceptions. In Hong Kong, as in the United States, negative contact appeared to be a more consistent predictor of intergroup attitudes than positive contact. Furthermore, intergroup anxiety explained the associations of both positive and negative contact with all outcome variables. As this pattern of results mirrored those of Study 1, Study 2 provides the first evidence that the effects of negative contact on prejudice may be largely cross-culturally invariant.

Study 3

Like the United States, Hong Kong is a nation that has been substantively affected by the British. Thus, it is important to turn to another intergroup setting to properly test our model cross-culturally. In Study 3, we aimed to replicate our results in Thailand, a non-Western nation that has seen relatively little Western influence. Thailand is one of four Asian countries (alongside Nepal, Bhutan, and Turkey) that have never been colonized by any European nations. Approximately, the Thai population consists of 94% Buddhists and 5% Muslims (Central Intelligence Agency, 2015). Muslim Thais have suffered many decades of social and economic marginalization. For instance, the four Southern provinces (in which Muslim Thais primarily live) are among the least developed provinces in Thailand in terms of education, poverty incidence and unemployment rates (Chongkittavorn, 2004; Croissant, 2007). Many Buddhist Thais today refer to Muslim Thais as “khaek,” a term with derogatory overtones that Muslim Thais resent (Horstmann, 2011). Muslim Thais, in contrast, tend to hold beliefs that Buddhist Thais are privileged and prejudiced, and Buddhist Thais are often concerned about these negative beliefs (Croissant, 2007; Horstmann, 2011).

Method

Participants and measures

A total of 161 Buddhist Thais (74% female; M_age = 26.54, SD = 7.59) were recruited via a major online discussion forum in Thailand (www.pantip.com). All measures were matched to those used in Studies 1 and 2. The target ingroup and outgroup were changed to “Buddhist Thais” and “Muslim Thais,” respectively. The questionnaire was also translated into Thai.

Results and Discussion

A small percentage of participants (<5%) did not complete all the dependent variable measures. Listwise deletion was thus used to deal with missing data. An identical set of primary analyses used in Studies 1 and 2 was performed to test our hypotheses. Scale reliabilities, descriptive statistics, and intercorrelations can be seen in Table 3.

Table 3.

Descriptive Information and Correlations Among Focal Variables (Study 3).

Focal variables	M	SD	1	2	3	4	5	6	7	8
1. Age	26.54	7.28	—
2. Gender (1 male, 2 female)	1.75	0.44	.12	—
3. Positive contact	3.28	1.10	.09	.06	—
4. Negative contact	2.06	0.99	.04	.03	.00	—
5. Intergroup anxiety	1.83	0.70	−.04	−.08	−.32***	.34***	(.91)
6. Old-fashioned prejudice	2.30	0.67	−.06	.06	−.38***	.29***	.71***	(.76)
7. Modern prejudice	3.24	0.66	−.07	.04	−.27***	.20**	.41***	.53***	(.70)
9. Negative metaperceptions	2.48	0.74	−.20*	−.04	−.28***	.36***	.56***	.53***	.50***	(.87)

Note. Internal reliabilities appear in parentheses.

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

When negative contact was not included in the analyses, positive contact predicted decreased old-fashioned prejudice, β = −.38, t(157) = −5.19, p < .001, modern prejudice, β = −.27, t(157) = −3.46, p < .001, and negative metaperceptions, β = −.27, t(145) = −3.39, p < .001. After including negative contact in the analyses, positive contact remained a reliable predictor of decreased old-fashioned prejudice, β = −.38, t(156) = −5.41, p < .001, modern prejudice, β = −.26, t(156) = −3.50, p < .001, and negative metaperceptions, β = −.28, t(144) = −3.92, p < .001. Negative contact also reliably predicted increased old-fashioned prejudice, β = .29, t(156) = 4.15, p < .001, modern prejudice, β = .20, t(156) = 2.69, p = .008, and negative metaperceptions, β = .37, t(144) = 5.06, p < .001. As in Studies 1 and 2, the associations between positive contact and old-fashioned prejudice (B = −.12, B_SE = .03, 95% CI = [−.19, −.07]), modern prejudice (B = −.06, B_SE = .02, 95% CI = [−.11, −.03]), and negative metaperceptions (B = −.10, B_SE = .03, 95% CI = [−.17, −.06]), were mediated by reduced anxiety. The associations between negative contact and old-fashioned prejudice (B = .15, B_SE = .04, 95% CI = [.07, .23]), modern prejudice (B = .08, B_SE = .03, 95% CI = [.04, .14]), and negative metaperceptions (B = .12, B_SE = .04, 95% CI = [.05, .22]), were mediated by increased anxiety.

In sum, Study 3 revealed similar patterns of results to Studies 1 and 2. Positive contact reliably predicted improved intergroup attitudes when negative contact was not taken into account. Interestingly, in Thailand, when negative contact was considered, positive contact was still directly associated with lower levels of old-fashioned prejudice, modern prejudice, and negative metaperceptions. Consistent with Studies 1 and 2, however, negative contact was directly related to higher levels of all three dependent variables. In addition, Study 3 also illustrated that intergroup anxiety is an important mediator of the associations between both types of contact and the outcome variables.

General Discussion

The present article investigated three important aspects of intergroup contact. In response to calls to explain how negative contact increases intergroup negativity (Barlow et al., 2012; Pettigrew, 2008), we examined positive and negative contact as predictors of intergroup attitudes, and intergroup anxiety as a core mediator. In response to Dixon et al.’s (2005) appeal for examining outcomes beyond prejudice, we introduced negative metaperceptions as a novel variable dependent on positive and negative contact. Finally, given the almost non-existent contact research in Thailand and Hong Kong (Pettigrew & Tropp, 2006), we tested our model in these nations to determine if our proposed contact effects were generalizable to non-Western populations.

The current findings allow us to draw several conclusions that advance our understanding of intergroup contact. First, negative contact was a consistent predictor of prejudice and attitudes beyond prejudice (i.e., negative metaperceptions). We argued that negative metaperceptions should be inherently tied to intergroup contact. What we think the outgroup thinks about our own group should be shaped by the interactions that we have with them (Plant & Butz, 2006; Vorauer, 2003, 2006). In line with our argument, negative contact was directly associated with increased negative metaperceptions of, as well as old-fashioned and modern prejudice toward Black Americans (Study 1), Mainland Chinese (Study 2), and Muslim Thais (Study 3). Second, intergroup anxiety helped explain why negative contact increases intergroup antipathy. In particular, all three studies revealed that negative contact had indirect effects on increased intergroup negativity via increased anxiety about interacting with outgroup members. Our results thus provide the first evidence that intergroup anxiety is a critical mediator of the negative contact-intergroup attitudes relationship.

Third, although past research indicated that positive contact directly predicts improved intergroup attitudes (see Pettigrew & Tropp, 2006), our data revealed mixed support for this notion. We found that positive contact was directly associated with reduced prejudice and negative metaperceptions in all studies when negative contact was not taken into account. But when it was considered, positive contact became a less reliable predictor of intergroup attitudes. In particular, when accounting for negative contact, positive contact was unrelated to negative metaperceptions of Black Americans in Study 1, and modern prejudice toward, and negative metaperceptions about, Mainland Chinese in Study 2. Our results, therefore, challenge the long-standing assumption that positive contact directly reduces intergroup antipathy. Instead, our findings are in line with contact theorists who have posited that negative contact can weaken the beneficial effects of positive contact, and that positive contact may be a less stable predictor of intergroup attitudes (Barlow et al., 2012). They further highlight the importance of considering negative contact when examining the benefits of positive contact.

Fourth, intergroup anxiety emerged as a robust mediator of the positive contact-intergroup attitudes relationship. Throughout all studies, positive contact predicted outcome variables via decreased anxiety. These results are interesting, considering that indirect effects of positive contact on some of the dependent outcomes were apparent in the absence of direct effects. This suggests that there might be multiple factors that mediate the relationship between positive contact and intergroup attitudes. Some of these mediators may act as suppressors, obscuring the direct effects of positive contact on intergroup attitudes. Although positive contact improves intergroup attitudes through reduced anxiety, this relationship might be suppressed by unmeasured mediating factors that increase intergroup negativity. Thus, the direct effects of positive contact on some intergroup outcomes would appear to be zero without taking the suppressing variables into consideration (see Rucker, Preacher, Tormala, & Petty, 2011 for a discussion). As potential suppressers have not yet been discovered, additional research is needed to specify suppressing variables that might temper the direct effects of positive contact.

Finally, data collected from White Americans (majorities in a Western nation), Hong Kong Chinese (majorities in a non-Western nation previously colonized by the British), and Buddhist Thais (majorities in a non-Western nation that has never been colonized) revealed similar patterns of associations between contact and intergroup attitudes. Specifically, positive contact reduced intergroup negativity indirectly via reduced anxiety and negative contact increased intergroup negativity both directly and indirectly via increased anxiety in all three samples. Therefore, we are the first to establish that these aspects of contact are generalizable to the Hong Kong Chinese and Thai populations. Our results also extend on Pettigrew and Tropp (2006), which found that the effects of contact are similar across multiple target groups. Here, we demonstrate that contact effects may be more robust than originally anticipated. Even when we varied two critical components (i.e., culture and target group), the pattern of results emerged in each study was remarkably similar.

This raises the question as to why contact worked so similarly across nations. Perhaps an answer can be found in the fact that, in each case, we surveyed comparatively privileged majority group members. Majorities, irrespective of culture, face similar challenges that they must contend with in intergroup settings (Devine & Vasquez, 1998; Tropp & Pettigrew, 2005). These challenges concern power and status maintenance, coexisting with anxieties about being perceived as prejudiced and being rejected by members of minority groups (Barlow, Louis, & Hewstone, 2009; Hyers & Swim, 1998; Plant, 2004; Stephan & Stephan, 1985; Vorauer et al., 1998). Given such pressures, it is perhaps unsurprising that we found evidence for parallels in the impact of contact on intergroup attitudes cross-culturally. Future studies, however, should examine whether minority groups are more sensitive to cultural particulars and respond differently to contact with majorities group members in different intergroup settings.

Limitations and Future Research Directions

Heterogeneity of target groups

We are not aware of an intergroup context that allows us to systematically compare the effects of positive and negative contact on intergroup attitudes across cultures without presenting confounds. However, we acknowledge that it is imperative to hold constant factors that have been shown to moderate contact effects. Prior research indicated the contact effects are stronger (a) among majorities compared with minorities (Tropp & Pettigrew, 2005) and (b) in neighborhoods where contact is less prevalent in comparison with neighborhoods where contact is more prevalent (Barlow, Hornsey, Thai, Sengupta, & Sibley, 2013). As such, in our studies, we chose to examine the largest and most salient conflict in each nation (i.e., White vs. Black Americans, Hong Kong vs. Mainland Chinese, and Buddhist vs. Muslim Thais). In doing so, we held constant these two established moderators. Specifically, participants consisted of only members of majority groups, preventing the possibility of a group membership status confound. In addition, the amounts of both types of contact appeared to be similar across nations (see Tables 1 -3), thereby eliminating a contact prevalence confound.

It is important to note that the current contexts did not allow us to hold target group constant. However, this should not pose a serious issue in the present studies. As Pettigrew and Tropp’s (2006) meta-analysis found that contact effects are stable across multiple different targets (perhaps among majority groups in general, as discussed above), any reported differences between studies would likely be due to the culture of the respondents rather than the heterogeneity of target groups. In the present article, we found one notable difference. When accounting for negative contact, positive contact failed to directly predict negative metaperceptions in Study 1 (the United States), and negative metaperceptions and modern prejudice in Study 2 (Hong Kong), but reliably predicted all outcome variables in Study 3 (Thailand). It is possible that there is a cultural factor in the West (and nations influenced by the West) that disproportionately biases people to attend to negative (rather than positive) information when making intergroup judgments. Nonetheless, given the design of our studies, we are unable to definitively conclude whether the observed difference was due to heterogeneity of the target groups, culture, or an interaction between the two. Thus, future research should replicate our findings cross-culturally while holding target ingroup and outgroup constant.

Common method bias

We recognize that self-report measures might give rise to concerns about common method variance. Therefore, we performed a Harman’s one-factor test, a statistical technique widely used by researchers to address this issue (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003; To, Fisher, Ashkanasy, & Rowe, 2012). Exploratory factor analyses of all variables generated 8 factors, with the first accounting for only 27% of the variance in Study 1, only 28% in Study 2, and only 29% in Study 3. As a single factor did not emerge, and the general factor did not account for the majority of the covariance among the variables, it is unlikely that common method bias was a major issue in our studies. Nevertheless, we encourage future studies to utilize more rigorous designs (e.g., experimental, longitudinal, multilevel) and more objective measurements (e.g., behavioral, archival, physiological) to avoid this issue entirely.

Causality

Data presented here are cross-sectional. Hence, definitive conclusions that contact alters prejudice and negative metaperceptions cannot be drawn, as some reciprocal causation is also possible. Nevertheless, decades of laboratory research with manipulated contact (see Pettigrew & Tropp, 2006, 2008) together with evidence from longitudinal studies (Binder et al., 2009; Sidanius, van Laar, Levin, & Sinclair, 2004) add weight to the idea that contact is causally related to negative intergroup attitudes.

Contact measures

Some researchers may be concerned about the use of single items measuring positive and negative contact, as they could be prone to measurement errors. However, Barlow and colleagues (2012) revealed that the current items are predictively valid. As suggested by Cronbach (1961), if predictive validity is satisfactory, single-item measures must be considered to be reliable and researchers must not be discouraged from using them (also see Bergkvist & Rossiter, 2007; Rossiter, 2002). As such, the current measures were unlikely to pose a serious problem in our studies.

It should be noted that our measure tapped into contact quantity (i.e., the frequency of contact), but not contact quality (i.e., the emotive strength of contact). Thus, little is known about whether (a) negative contact quality is a more robust predictor of intergroup attitudes than positive contact quality across cultures, and (b) intergroup anxiety mediates both the effects of positive and negative contact quality. As such, an examination of positive and negative contact quality as predictors of intergroup attitudes will help future research address these empirical gaps.

Conclusion

Across three studies, we demonstrate that positive and negative contact had independent effects on both prejudice and negative metaperceptions. While positive contact predicted improved intergroup attitudes, negative contact predicted increased intergroup negativity. Negative contact, however, was the more robust and consistent predictor of intergroup attitudes. We also provide the first evidence that intergroup anxiety mediates the relationships between both types of contact and intergroup outcomes. Finally, the current findings display some support for the universality of contact effects, suggesting that when examining majority groups and their attitudes toward minorities, positive and negative contact’s patterns of prediction were largely stable irrespective of cultural context.

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by an Australian Research Council grant (DE120102068) awarded to the second author.

Notes

References

Allport

G. W.

(1954). The nature of prejudice. Reading, MA: Addison-Wesley.

Ames

(2004). Inside the mind reader’s tool kit: Projection and stereotyping in mental state inference. Journal of Personality and Social Psychology, 87, 340-353.

Barlow

F. K.

Hornsey

M. J.

Thai

Sengupta

N. K.

Sibley

C. G.

(2013). The wallpaper effect: The contact hypothesis fails for minority group members who live areas with a high proportion of majority group members. PLoS ONE, 8, e82228.

Barlow

F. K.

Louis

W. R.

Hewstone

(2009). Rejected! Cognitions of rejection and intergroup anxiety as mediators of the impact of cross-group friendships on prejudice. British Journal of Social Psychology, 48, 389-405.

Barlow

F. K.

Paolini

Pedersen

Hornsey

M. J.

Radke

H. R. M.

Harwood

Sibley

C. G.

(2012). The contact caveat: Negative contact predicts increased prejudice more than positive contact predicts reduced prejudice. Personality and Social Psychology Bulletin, 38, 1629-1643.

Bergkvist

Rossiter

J. R.

(2007). The predictive validity of multiple-item versus single-item measures of the same constructs. Journal of Marketing Research, 44, 175-184.

Bergsieker

H. B.

Shelton

J. N.

Richeson

J. A.

(2010). To be liked versus respected: Divergent goals in interracial interactions. Journal of Personality and Social Psychology, 99, 248-264.

Binder

Zagefka

Brown

Funke

Kessler

Mummendey

. . .Leyens

J. P.

(2009). Does contact reduce prejudice or does prejudice reduce contact? A longitudinal test of the contact hypothesis among majority and minority groups in three European countries. Journal of Personality and Social Psychology, 96, 843-856.

Central Intelligence Agency. (2015). The World Factbook. Retrieved from https://www.cia.gov/library/publications/the-world-factbook/geos/th.html

10.

Cheuk-Lam

P. K.

(2013). The prejudicial portrayal of immigrant families from Mainland China in Hong Kong media. In Chan

K. B.

(Ed.), International handbook of Chinese families (pp. 211-227). New York, NY: Springer.

11.

Chongkittavorn

(2004). Thailand: International terrorism and the Muslim South. In Singh

Wah

C. K.

(Eds.), Southeast Asian Affairs 2004 (pp. 267-275). Singapore: ISEAS.

12.

Community Legal Information Centre. (2009). Immigration to Hong Kong for non-residents (family members). Retrieved from http://www.hkclic.org/en/topics/immigration/for_non_residents/q1.shtml

13.

Croissant

(2007). Muslim insurgency, political violence, and democracy in Thailand. Terrorism and Political Violence, 19, 1-18.

14.

Cronbach

L. J.

(1961). Essentials of psychological testing. New York, NY: Harper & Row.

15.

Devine

P. G.

Vasquez

K. A.

(1998). The rocky road to positive intergroup relations. In Eberhardt

J. L.

Fiske

S. T.

(Eds.), Confronting racism: The problem and the response (pp. 234-262). Thousand Oaks, CA: Sage.

16.

Dixon

Durrheim

Tredoux

(2005). Beyond the optimal contact strategy: A reality check for the contact hypothesis. American Psychologist, 60, 697-711.

17.

Dixon

Levine

Reicher

Durrheim

(2012). Beyond prejudice: Are negative evaluations the problem and is getting us to like one another more the solution? Behavioral and Brain Sciences, 35, 411-425.

18.

Dovidio

J. F.

Gaertner

S. L.

(2000). Aversive racism and selection decisions: 1989 and 1999. Psychological Science, 11, 315-319.

19.

Finchilescu

(2005). Meta-stereotypes may hinder inter-racial contact. South African Journal of Psychology, 35, 460-472.

20.

Gallup Organization. (2015). Race relations. Retrieved from http://www.gallup.com/poll/1687/Race-Relations.aspx.

21.

Harris

A. L.

(2010). The economic and educational state of Black Americans in the 21st century: Should we be optimistic or concerned? Review of Black Political Economy, 37, 241-252.

22.

Hayes

A. F.

Preacher

K. J.

(2014). Statistical mediation analysis with a multicategorical independent variable. British Journal of Mathematical and Statistical Psychology, 67, 451-470.

23.

Horstmann

(2011). Living together: The transformation of multi-religious coexistence in southern Thailand. Journal of Southeast Asian Studies, 42, 487-510.

24.

Hyers

L. L.

Swim

J. K.

(1998). A comparison of the experiences of dominant and minority group members during an intergroup encounter. Group Processes & Intergroup Relations, 1, 143-163.

25.

Jackman

M. R.

Crane

(1986). “Some of my best friends are Black . . . :” Interracial friendship and Whites’ racial attitudes. Public Opinion Quarterly, 50, 459-486.

26.

King

M. E.

Winter

Webster

(2009). Contact reduces transprejudice: A study on attitudes towards transgenderism and transgender civil rights in Hong Kong. International Journal of Sexual Health, 21, 17-34.

27.

Laher

Finchilescu

(2010). Meta-stereotypes: Intergroup anxiety in interracial contact among students. New Voices in Psychology, 6, 4-16.

28.

McConahay

J. B.

(1986). Modern racism, ambivalence, and the modern racism scale. In Dovidio

J. F.

Gaertner

S. L.

(Eds.), Prejudice, discrimination, and racism (pp. 91-125). New York, NY: Academic Press.

29.

Paolini

Harwood

Rubin

(2010). Negative intergroup contact makes group memberships salient: Explaining why intergroup conflict endures. Personality and Social Psychology Bulletin, 36, 1723-1738.

30.

Pettigrew

T. F.

(2008). Future direction for intergroup contact theory and research. International Journal of Intercultural Relations, 32, 187-199.

31.

Pettigrew

T. F.

Tropp

L. R.

(2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90, 751-783.

32.

Pettigrew

T. F.

Tropp

L. R.

(2008). How does intergroup contact reduce prejudice? Meta-analytic tests of three mediators. European Journal of Social Psychology, 38, 922-934.

33.

Pettigrew

T. F.

Tropp

L. R.

Wagner

Christ

(2011). Recent advances in intergroup contact theory. International Journal of Intercultural Relations, 35, 271-280.

34.

Plant

E. A.

(2004). Responses to interracial interactions over time. Personality and Social Psychology Bulletin, 30, 1458-1471.

35.

Plant

E. A.

Butz

D. A.

(2006). The causes and consequences of an avoidance-focus for interracial interactions. Personality and Social Psychology Bulletin, 32, 833-846.

36.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J. Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903.

37.

Rossiter

J. R.

(2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305-335.

38.

Rucker

D. D.

Preacher

K. J.

Tormala

Z. L.

Petty

R. E.

(2011). Mediation analysis in social psychology: Current practices and new recommendations. Social & Personality Psychology Compass, 5, 359-371.

39.

Sidanius

van Laar

Levin

Sinclair

(2004). Ethnic enclaves and the dynamics of social identity on the college campus: The good, the bad, and the ugly. Journal of Personality and Social Psychology, 87, 96-110.

40.

Stephan

W. G.

Stephan

C. W.

(1985). Intergroup anxiety. Journal of Social Issues, 41, 157-176.

41.

Time World. (2012). Trouble down south: Why Hong Kong and Mainland Chinese aren’t getting along. Retrieved from http://world.time.com/2012/01/24/trouble-down-south-why-hong-kong-and-mainland-chinese-arent-getting-along/.

42.

M. L.

Fisher

C. D.

Ashkanasy

N. M.

Rowe

P. A.

(2012). Within-person relationships between mood and creativity. Journal of Applied Psychology, 97, 599-612.

43.

Tropp

L. R.

Pettigrew

T. F.

(2005). Relationships between intergroup contact and prejudice among minority and majority status groups. Psychological Science, 16, 951-957.

44.

United States Census Bureau. (2011). State and county facts. Retrieved from http://quickfacts.census.gov/qfd/states/00000.html

45.

Vorauer

J. D.

(2003). Dominant group members in intergroup interaction: Safety or vulnerability in numbers? Personality and Social Psychology Bulletin, 29, 498-511.

46.

Vorauer

J. D.

(2006). An information search model of evaluative concerns in intergroup interaction. Psychological Review, 113, 862-886.

47.

Vorauer

J. D.

Hunter

A. J.

Main

K. J.

Roy

S. A.

(2000). Meta-stereotype activation: Evidence from indirect measures for specific evaluative concerns experienced by members of dominant groups in intergroup interaction. Journal of Personality and Social Psychology, 78, 690-707.

48.

Vorauer

J. D.

Main

K. J.

O’Connell

G. B.

(1998). How do individuals expect to be viewed by members of lower status groups? Content and implications of meta-stereotypes. Journal of Personality and Social Psychology, 75, 917-937.

49.

Walker

(1994). Attitudes to minorities: Survey evidence of Western Australians’ attitudes to Aborigines, Asians and women. Australian Journal of Psychology, 46, 137-143.