Thinking about Race: How Group Biases Interact with Ideological Principles to Yield Attitudes toward Government Assistance

When are people more likely to evaluate race-targeted government assistance based on ideological principles rather than racial prejudice? In order to answer this question, it is necessary to understand the mechanisms by which prejudice influences political attitudes. In this dissertation, I develop a theoretical model for explaining how deep-seated, automatic group biases interact with higher-order, ideological principles in order to influence attitudes toward race-targeted government assistance. I suggest group-based principles are more important than individualistic values or ingroup favoritism in explaining race-targeted policy attitudes. I argue that when people evaluate race-targeted policies, controlled neural processes translate automatic neural processes into broad group-based principles, which then become the primary tool people use to evaluate race-targeted policies. As such, the degree to which people’s race-targeted policy opinions are driven by principles rather than automatic group biases is a function of how much controlled processing has occurred. I find support for this model across an array of empirical investigations. In a survey experiment, I find group-based principles to outperform an array of other constructs in predicting race-targeted policy attitudes. Then, in a laboratory experiment, I replicate the primary findings from the survey experiment and show further that the only influence ingroup favoritism has on race-targeted policy attitudes is through automatic, implicit processes. Further, group-based principles are comprised of a combination of ingroup bias and individualistic values. Next, I investigate the translation process from automatic to controlled processes by examining the implications of discrepancies between people’s implicit and explicit attitudes – known as implicit ambivalence. I find that individuals with the most “resolved” racial attitudes are the most likely to evaluate race-related political objects ideologically. Finally, I directly examine the automatic and controlled neural processes hypothesized to underlay this model by using functional Magnetic Resonance Imaging (fMRI). I find evidence that activation in brain regions implicated in controlled processing is associated with explicit evaluations of race, and this relationship is strongest among individuals with more extreme group-based principles. I conclude by discussing implications of this model for existing literature as well as possible public policy interventions for reducing the role of prejudice in politics. Advisor: Elizabeth Theiss-Mors