Examining County-Level Implicit Stereotypes and Prejudice as Predictors of Racial and Gender Disparities in Incarceration Across the United States

Background: This dissertation investigated the role of local macropsychological factors - specifically, county-level implicit racial bias and implicit gender stereotyping - in shaping inequalities in incarceration across the United States. Despite overall declines in incarceration, trends in the incarceration of women and people of color (particularly Black women) show disparities compared to White men. Regional aggregations of implicit biases have been proposed as robust markers and drivers of systemic inequalities and are examined in current research. Research Questions: Three overarching research questions (RQ) guided this dissertation: RQ1 tested the relationship between county-level implicit racial bias and proportions of Black (relative to White) incarceration; RQ2 tested the relationship between county-level implicit gender stereotyping and the proportion of female (relative to male) incarceration; and RQ3 tested the interactive effect of implicit racial bias and implicit gender stereotyping on Black female (relative to White female) incarceration. Hypotheses: Given prior research demonstrating positive relationships between implicit racial bias and other criminal justice outcomes (e.g., traffic stops and use of fatal police force against Black individuals), it was anticipated that counties with higher levels of implicit racial bias would exhibit higher proportions of Black incarceration. Regarding the effect of implicit gender stereotyping on female incarceration, competing hypotheses were proposed in line with the ambivalent nature of sexism. That is, either a positive "punishing" effect of gender stereotyping on women's incarceration was anticipated, or a reductive "chivalry" effect on incarceration. Furthermore, considering a "selective chivalry" perspective, it was anticipated that chivalry may only reduce the incarceration of White women, leading to higher relative incarceration rates for Black women. Therefore, counties with higher levels of both implicit racial bias and implicit gender stereotyping were expected to have greater incarceration rates for Black females relative to White females. Method: Twelve years (2005-2016) of publicly available archival data on implicit cognitive associations and incarceration demographics were collected from Project Implicit and The Vera Institute of Justice. For a nuanced examination of incarceration, all variables were tied back to the county level, and incarceration indicators included both admissions and populations in both jails and prisons, where possible. Analyses: Cross-sectional and longitudinal multilevel modeling were used to test each of the hypotheses in this dissertation. Counties were clustered within states to acknowledge their state-level sentencing policies, county demographic controls were entered into each model (e.g., target group population proportions and crime rates, among others), and explicit attitudes were controlled for to test the unique impact of unconscious biases "in the air" of the community. Results: Results for RQ1 suggested that counties with higher implicit racial bias had higher proportions of Black (relative to White) incarceration, but only at the prison level and in cross-sectional analyses. For RQ2, results supported both the chivalry and punishing hypotheses for female (relative to male) incarceration, where chivalry was evident in jail-level incarceration, while punishment appeared at the prison level; these findings were observed in longitudinal models but not in cross-sectional ones. Finally, RQ3 cross-sectional (but not longitudinal) models indicated that greater implicit gender stereotyping led to relatively lower (though still not equitable) levels of Black female (relative to White female) incarceration, but only in counties that were also low in implicit racial bias. Discussion: Findings support the notion that the prejudicial cognitive biases shared across communities can affect gendered and racial representation in incarceration in both jails and prisons. Policy implications are discussed for mitigating the context-level effects of implicit prejudice on systemically disparate incarceration, and how federal data collection can be reformed to address significant barriers to conducting intersectional research on criminal justice outcomes.