Less biased yet more defensive: The impact of control processes.

Educational and training programs designed to reduce racial bias often focus on increasing people's awareness of psychological sources of their biases. However, when people learn about their biases, they often respond defensively, which can undermine the effectiveness of antibias interventions and the success of prejudice regulation. Using process (Quad) modeling, we provide one of the first investigations of the relationships between (a) controlled and automatic cognitive processes that underpin performance on the Implicit Association Test and (b) defensive reactions to unflattering implicit racial bias feedback. In two correlational samples (one preregistered; N = 8,000) and one experiment in which the provision of bias feedback was manipulated (N = 547), we find racially biased associations and some control over these associations among White people. Nonetheless, more defensiveness to bias feedback consistently predicted weaker ability to control biased associations. We also find correlational evidence that lower levels of biased associations predict more defensiveness, but did not replicate this observation in the experimental study. These results are critical for theories of implicit attitudes, models of prejudice regulation, and strategies for antibias interventions. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Estimating the Reliability and Stability of Cognitive Processes Contributing to Responses on the Implicit Association Test.

Implicit measures were initially assumed to assess stable individual differences, but other perspectives posit that they reflect context-dependent processes. This pre-registered research investigates whether the processes contributing to responses on the race Implicit Association Test are temporally stable and reliably measured using multinomial processing tree modeling. We applied two models-the Quad model and the Process Dissociation Procedure-to six datasets (N = 2,036), each collected over two occasions, examined the within-measurement reliability and between-measurement stability of model parameters, and meta-analyzed the results. Parameters reflecting accuracy-oriented processes demonstrate adequate stability and reliability, which suggests these processes are relatively stable within individuals. Parameters reflecting evaluative associations demonstrate poor stability but modest reliability, which suggests that associations are either context-dependent or stable but noisily measured. These findings suggest that processes contributing to racial bias on implicit measures differ in temporal stability, which has practical implications for predicting behavior using the Implicit Association Test.

Real-life outgroup exposure, self-reported outgroup contact and the other-race effect.

People are better at recognizing faces from their own racial or ethnic group compared with faces from other racial or ethnic groups, known as the other-'race' effect (ORE). Several theories of the ORE assume that memory for other-race faces is impaired because people have less contact with members of other racial or ethnic groups, resulting in lower visual expertise. The present research investigates contact theories of the ORE, using self-report contact measures and objective measures of potential outgroup exposure (estimated from participants' residential location and from GPS tracking). Across six studies (total N = 2660), we observed that White American and White German participants displayed better memory for White faces compared with Black or Middle Eastern faces, whereas Black American participants displayed similarly equal or better memory for White compared with Black faces. We did not observe any relations between the ORE and objective measures of potential outgroup exposure. Only in Studies 2a and 2b, we observed very small correlations (rs = -.08 to .06) between 4 out of 30 contact measures and the ORE. We discuss methodological limitations and implications for theories of the ORE.

A recognition advantage for members of higher-status racial groups.

The other-race effect (ORE) is a recognition memory advantage afforded to one's racial ingroup versus outgroup. The motivational relevance of the ingroup-because of relationships, belonging and self-esteem-is central to many theoretical explanations for the ORE. However, to date, the motivational relevance of outgroups has received considerably less attention in the ORE literature. Across six experiments, Black, White, Asian and Latinx American participants consistently demonstrated better recognition memory for the faces of relatively higher-status racial/ethnic group members than those of lower-status groups. This higher-status recognition advantage even appeared to override the ORE, such that participants better recognized members of higher-status outgroups-but not an outgroup of equivalent status-compared to members of their own ingroup. However, across a variety of self-reported perceived status measures, status differences between the high- and low-status groups generally did not moderate the documented recognition advantage. These findings provide initial evidence for the potential role of group status in the ORE and in recognition memory more broadly, but future work is needed to rule out alternative explanations.

The contributions of positive outgroup and negative ingroup evaluation to implicit bias favoring outgroups.

People sometimes prefer groups to which they do not belong (outgroups) over their own groups (ingroups). Many long-standing theoretical perspectives assume that this outgroup favorability bias primarily reflects negative ingroup evaluations rather than positive outgroup evaluations. To examine the contributions of negative ingroup versus positive outgroup evaluations to outgroup bias, we examined participants' data (total n > 879,000) from Implicit Association Tests [A. G. Greenwald, D. E. McGhee, J. L. K. Schwartz, J. Pers. Soc. Psychol. 74, 1464-1480 (1998)] measuring intergroup attitudes across four social domains in exploratory and preregistered confirmatory analyses. Process modeling [F. R. Conrey, J. W. Sherman, B. Gawronski, K. Hugenberg, C. J. Groom, J. Pers. Soc. Psychol. 89, 469-487 (2005)] was applied to the responses of participants who demonstrated implicit outgroup bias to separately estimate the contributions of negative ingroup and positive outgroup evaluations. The outgroup biases of lower-status group members (i.e., Asian, Black, gay and lesbian, and older people) consistently reflected greater contributions of positive outgroup evaluations than negative ingroup evaluations. In contrast, the outgroup biases of higher-status group members (i.e., White, straight, and younger people) reflected a more varied pattern of evaluations. We replicated this pattern of results using explicitly measured intergroup evaluations. Taking these data together, the present research demonstrates a positive-negative asymmetry effect of outgroup bias, primarily among members of lower-status groups.

The role of recoding in implicit social cognition: Investigating the scope and interpretation of the ReAL model for the implicit association test.

The ReAL model is a multinomial processing tree model that quantifies the contribution of three qualitatively distinct processes-recoding, associations, and accuracy-to responses on the implicit association test (IAT), but has only been validated on a modified version of the IAT procedure. The initial goal of the present research was to validate an abbreviated version of the ReAL model (i.e., the Brief ReAL model) on the standard IAT procedure. Two experiments replicated previous validity evidence for the ReAL model on the modified IAT procedure, but did not produce valid parameter estimates for the Brief ReAL model on the standard IAT procedure. A third, pre-registered experiment systematically manipulated all of the task procedures that vary between the standard and modified IAT procedures-response deadline, number of trials, trial constraints-to determine the conditions under which the Brief ReAL model can be validly applied to the IAT. The Brief ReAL model estimated theoretically-interpretable parameters only under a narrow range of IAT conditions, but the ReAL model generally estimated theoretically-interpretable parameters under most IAT conditions. We discuss the application of these findings to implicit social cognition research, and their implications to social cognitive theory.

Moderators of intergroup evaluation in disadvantaged groups: A comprehensive test of predictions from system justification theory.

We examined hypotheses proposed by System Justification Theory (SJT; Jost, Banaji, & Nosek, 2004) regarding intergroup evaluation in disadvantaged groups, using large samples of online participants (total N = 715,721), spanning 8 intergroup domains and 14 nations. Using a meta-analytic approach, we tested these hypotheses at the individual level (as SJT is generally articulated), as well as at the social group level. Consistent with SJT, individual-level analyses revealed that disadvantaged groups demonstrated outgroup favoritism on Implicit Association Tests (IATs; i.e., implicit measures), but demonstrated ingroup favoritism or no intergroup preference on self-report (i.e., explicit) measures. Additionally, these average effects were characterized by high heterogeneity, and follow-up exploratory analyses revealed that intergroup evaluation in disadvantaged groups was moderated by the intergroup domain: Whereas some disadvantaged groups consistently displayed outgroup favoritism (e.g., age, weight), others consistently displayed ingroup favoritism (e.g., sexual orientation, religion), and yet others displayed diverging patterns on implicit and explicit measures (e.g., race, ethnicity). Consistent with SJT, intergroup evaluation on all measures was moderated by self-reported conservatism. Furthermore, the magnitude of these relationships depended on the level of analysis, with small effects emerging at the individual level and medium-sized effects emerging at the social group level. Social group-level analyses also indicated that intergroup evaluation in disadvantaged groups was moderated by stigma. Overall, these findings support and extend the predictions of SJT, but the relatively complex patterns of intergroup evaluation in disadvantaged groups identified here illustrate a need for further theory development and more theory-driven research in this domain. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Reducing implicit racial preferences: III. A process-level examination of changes in implicit preferences.

Implicit bias change was initially assumed to reflect changes in associations, but subsequent research demonstrated that implicit bias change can also reflect changes in control-oriented processes that constrain the expression of associations. The present research examines the process-level effects of 17 different implicit bias-reduction interventions and one sham intervention by analyzing data from more than 20,000 participants who completed an intervention condition or a baseline control condition followed by a race Implicit Association Test (IAT). To identify the processes influenced by each intervention, we applied the Quadruple process model to participants' IAT responses then meta-analyzed parameter estimates according to a taxonomy of interventions based on shared features. Interventions that relied on counterstereotypic exemplars or strategies to override biases influenced both associations and control-oriented processes, whereas interventions that relied on evaluative conditioning influenced only control-oriented processes. In contrast, interventions that focused on egalitarian values, perspective taking, or emotion had no reliable influence on any of the processes examined. When interventions did change associations, they were much more likely to reduce positive White associations than negative Black associations. The present research extends upon traditional dual-process perspectives by identifying robust intervention effects on response biases. These findings connect features of interventions with changes in the processes underlying implicit bias. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

A Process Dissociation Model of Implicit Rapid Revision in Response to Diagnostic Revelations.

Previous research has demonstrated that implicit evaluations can be reversed with exposure to a single impression-inconsistent behavior. But what exactly is changing when perceivers encounter diagnostic revelations about someone? One possibility is that rapid changes are occurring in the extent to which perceivers view the person positively or negatively. Another possibility is that they override the expression of initial evaluations through control-oriented processes. We conducted three studies (one preregistered) that used multinomial process trees to distinguish between these possibilities. We find consistent support across two different implicit measures that diagnostic behaviors result in rapid changes in evaluative processes. We obtained only inconsistent evidence for effects on more control-oriented processes. These findings thus help to reveal the cognitive processes underlying rapid implicit revision. Implications for theoretical perspectives on implicit attitudes are discussed.

The impact of instruction- and experience-based evaluative learning on IAT performance: a Quad model perspective.

Learning procedures such as mere exposure, evaluative conditioning, and approach/avoidance training have been used to establish evaluative responses as measured by the Implicit Association Test (IAT). In this paper, we used the Quad model to disentangle the processes driving IAT responses instantiated by these evaluative learning procedures. Half of the participants experienced one of these three procedures whereas the other half only received instructions about how the procedure would work. Across three experiments (total n = 4231), we examined the extent to which instruction-based versus experience-based evaluative learning impacted Quad estimates of the Activation of evaluative information in IAT responses. Relative to a control condition, both instruction- and experience-based evaluative learning procedures influenced Activation. Moreover, and contrary to what prevailing models of implicit evaluations would predict, in no instance did experience-based procedures influence (positive or negative) Activation more strongly than instruction-based procedures. This was true for analyses which combined procedures and also when testing all three procedures individually. Implications for the processes that mediate evaluative learning effects and the conditions under which those processes operate are discussed.

Retrieval cues fail to influence contextualized evaluations.

Initial evaluations generalise to new contexts, whereas counter-attitudinal evaluations are context-specific. Counter-attitudinal information may not change evaluations in new contexts because perceivers fail to retrieve counter-attitudinal cue-evaluation associations from memory outside the counter-attitudinal learning context. The current work examines whether an additional, counter-attitudinal retrieval cue can enhance the generalizability of counter-attitudinal evaluations. In four experiments, participants learned positive information about a target person, Bob, in one context, and then learned negative information about Bob in a different context. While learning the negative information, participants wore a wristband as a retrieval cue for counter-attitudinal Bob-negative associations. Participants then made speeded as well as deliberate evaluations of Bob while wearing or not wearing the wristband. Internal meta-analysis failed to find a reliable effect of the counter-attitudinal retrieval cue on speeded or deliberate evaluations, whereas the context cues influenced speeded and deliberate evaluations. Counter to predictions, counter-attitudinal retrieval cues did not disrupt the generalisation of first-learned evaluations or the context-specificity of second-learned evaluations (Experiments 2-4), but the counter-attitudinal retrieval cue did influence evaluations in the absence of context cues (Experiment 1). The current work provides initial evidence that additional counter-attitudinal retrieval cues fail to disrupt the renewal and generalizability of first-learned evaluations.

Establishing construct validity evidence for regional measures of explicit and implicit racial bias.

Large-scale data collection has enabled social scientists to examine psychological constructs at broad, regional levels. However, because constructs and their measures initially operationalized at the individual level may have qualitatively and quantitatively different properties at other levels of analysis, the validity of constructs must be established when they are operationalized at new levels. To this end, the current research presents evidence of construct validity for explicit and implicit racial bias at region levels. Following classic measurement theory, we examine the substantive, structural, and external evidence of construct validity for regional biases. We do so with responses from ∼2 million Black and White North Americans collected over 13 years. Though implicit measures typically demonstrate low retest reliability at the individual level, our analyses reveal conventionally acceptable levels of retest reliability at the highest levels of regional aggregation. Additionally, whereas previous meta-analyses find relatively low explicit-implicit correlations at the individual level, the present research uncovered strong explicit-implicit correlations at regional levels. The findings have implications for how we interpret measures of racial bias at regional levels. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Reducing implicit racial preferences: II. Intervention effectiveness across time.

Implicit preferences are malleable, but does that change last? We tested 9 interventions (8 real and 1 sham) to reduce implicit racial preferences over time. In 2 studies with a total of 6,321 participants, all 9 interventions immediately reduced implicit preferences. However, none were effective after a delay of several hours to several days. We also found that these interventions did not change explicit racial preferences and were not reliably moderated by motivations to respond without prejudice. Short-term malleability in implicit preferences does not necessarily lead to long-term change, raising new questions about the flexibility and stability of implicit preferences. (PsycINFO Database Record

Intoxicated prejudice: The impact of alcohol consumption on implicitly and explicitly measured racial attitudes.

Recent research has shown that alcohol consumption can exacerbate expressions of racial bias by increasing reliance on stereotypes. However, little work has investigated how alcohol affects intergroup evaluations. The current work sought to address the issue in the context of the correspondence between implicit and explicit measures of anti-black attitudes. Participants were randomly assigned to consume an alcoholic (target BAC of 0.08%), placebo, or control beverage prior to completing implicit and explicit measures of racial attitudes. Although beverage condition did not affect prejudice levels on either measure, it did change the correlation between them. Implicitly measured attitudes significantly predicted explicit reports of prejudice and discrimination only for participants who consumed alcohol. We discuss the implications of our findings for debates regarding dissociations between implicit and explicit measures and the cultural phenomenon of intoxicated individuals attributing prejudiced statements to alcohol consumption rather than personal attitudes.

Attitudinal and non-attitudinal components of IAT performance.

The Implicit Association Test (IAT) was designed to measure automatically activated attitudinal associations, free of the influence of processes that affect their expression. Subsequent research has shown that IAT performance also is influenced by non-associative processes, but the extent to which these non-associative processes are content-specific or if they operate similarly regardless of the attitude being measured has largely gone unexamined. In the current research, participants completed pairs of IATs that varied in conceptual overlap: Tests shared a high, moderate, or low degree of overlap in the measured attitudes. The Quad model was applied to estimate the contributions of four processes to IAT performance. Evidence was found for two relatively general, non-attitudinal processes and two relatively attitude-specific processes. Implications are discussed for interpretation of IAT scores, individual differences in IAT scores, and IAT score malleability.

How malleable is categorization by race? Evidence for competitive category use in social categorization.

We contrast 3 theoretical viewpoints concerning the factors affecting social categorization by race: (a) the classical theory of social categorization highlighting the role of a priori accessibility and situational factors, (b) the classical theory augmented by a principle of competitive category use, and (c) competition between race (but not gender) and coalition with race (but not gender) encoded only as a proxy to coalition. Study 1 documents a confound that renders important portions of previous research difficult to interpret. In Studies 2 and 3, race categorization was stronger than categorization by more weakly accessible categories when situational support in terms of topic relevance was comparable across categories. A situational focus on race further increased race categorization. Race categorization was reduced in the presence of strongly cued cross-cutting coalitions. Race categorization also was depressed when situational factors promoted comparative processing of cross-cutting categories while cues to potential coalitional divisions were held constant (Study 4). Accessibility, topic relevance, and cuing cross-cutting coalitions had the same effects on gender categorization as found for race categorization (Study 5). Taken together, the results suggest that classical theories of social categorization have to be augmented by a principle of competitive category use that is not limited to a competition between race and coalition.