Psychological Moderators and Mediators of Stereotype Threat: Implications for Self-Affirmation Interventions

• In Event: Scaling Up a Social-Psychological Intervention to Combat Stereotype Threat and Narrow Achievement Gaps

Abstract

Objectives

This study addresses questions concerning potential psychological moderators and mediators of stereotype threat and self-affirmation in order to better understand treatment heterogeneity and the psychological mechanisms through which the self-affirmation intervention is working. Although many psychological processes may be involved (Cohen et al., 2009), we measured seven psychological constructs that we hypothesized as important for stereotype threat. This allowed us to investigate if psychological variables predicted which individuals within threatened groups showed larger achievement gaps and if gaps in psychological variables paralleled racial achievement gaps. Our results focus on student-level variables that are important for understanding when and why this intervention could be expected to work in future scale-ups. Understanding mechanisms also has implications for designing future psychological interventions to reduce achievement gaps.

Perspectives

Although self-affirmation has been shown to be effective at closing racial achievement gaps in smaller-scale field experiments (Cohen et al., 2006; 2009), current studies have not addressed if this intervention is equally effective for all students. Laboratory research (Aronson et al., 1999) suggests that stereotype threat is most salient for individuals who identify with the domain (i.e. viewing the domain as important and having confidence in their abilities). Furthermore, self-affirmation theory (Cohen et al., 2009; Sherman & Cohen, 2006) suggests that multiple mediators may have important effects over time, so we measured five theoretically relevant mediators.

Methods/Data Sources

During the 2012-13 school year, we gave 7th grade students in the second cohort (N=1098) a survey to measure identification with school, self-confidence, locus of control, self-complexity, evaluative anxiety, group belonging, and social belonging at the beginning of the school year. Sixth-grade test scores were used to estimate the magnitude of the racial achievement gaps prior to the intervention and survey. We compared threatened groups (primarily African American and Latino students) to non-threatened groups (White and Asian students).

Our analyses were twofold. First, to test for moderators, we examined whether the psychological variables moderated the effect of group status on the achievement gap. Second, we tested the psychological variables as potential mediators of stereotype threat by examining whether there were group differences, or “psychological gaps,” on the measures and, also, whether the psychological variables correlated with achievement.

Results/Scientific Significance

We found that both identification with school and self-confidence moderated the effect of group status on the achievement gap. That is, threatened group members who showed higher levels of identification with school and self-confidence had the largest racial achievement gaps. Thus, we expect that these students may benefit the most from self-affirmation. These results support previous laboratory findings (Aronson et al., 1999).

Locus of control and self-complexity showed a racial psychological gap, and also correlated with the achievement gap. Specifically, threatened group members showed a more external locus of control and lower levels of self-complexity, and these variables were also correlated with lower baseline achievement. Thus, these variables showed a racial psychological gap consistent with the racial achievement gap, suggesting potential psychological mechanisms that self-affirmation or future novel interventions might target in order to narrow gaps in academic achievement.

Authors

• Chris Rozek, University of Wisconsin - Madison

• Jaymes Ray Pyne, University of Wisconsin

• Geoffrey D. Borman, University of Wisconsin - Madison

• Sarah Bruch, University of Iowa

• Paul Hanselman, University of Wisconsin - Madison