Where Race Matters: A Geographic Approach to the Racial Threat Hypothesis

Abstract

One of the oldest debates in the study of racial attitudes and voting contrasts the racial threat hypothesis - which posits that racial identities and attitudes intensify in proximity to the outgroup - and the contact hypothesis - which argues that exposure to outgroup members will actually temper such attitudes, rather than heighten them. Studies to date have not emphasized geography and multivariate regression to examine this question. This paper uses geographically weighted regression (GWR) to estimate the relationship between racial composition and voting on Proposition 209, the 1996 anti-affirmative action measure that passed in California. The fine-grained results show that neither mechanism serves as a good universal predictor of the relationship between race and racial voting, but that racial threat may be at play in Los Angeles. These results also pinpoint particular places where further qualitative work may shed light on the context-bound mechanisms of racial behavior.