Evolutionary Model of Racial Attitude Formation Socially Shared and Idiosyncratic Racial Attitudes

A growing body of research in political science has uncovered evidence of a " split personality " among Americans when it comes to racial attitudes, whereby people express different attitudes in public than they personally hold. A common assumption is that people adjust their personal attitudes to conform to dominant social norms. At present, however, there is no theoretical model that could account for the emergence of racial norms that are at odds with people's personal attitudes. This paper proposes a simple neural model of racial attitude formation that makes an important distinction between socially shared and idiosyncratic racial attitudes. Socially shared attitudes reflect evaluations that are culturally transmitted and may not necessarily represent an individual's personal views. In contrast, idiosyncratic attitudes represent a sense of interpersonal 'chemistry' that may be at odds with dominant social norms. A computational model based on Kimura's (1983) Neutral Theory of Evolution predicts that socially shared racist attitudes may be able to coexist with, and eventually be replaced by, more favorable idiosyncratic racial attitudes. In order to investigate racial attitudes unencumbered by considerations of social desirability and 'political correctness,' this study augments traditional survey measures with a number of reaction time based measures of non-conscious racial attitudes. Socially shared, non-conscious attitudes are measured using implicit racial priming based on a lexical decision task. Idiosyncratic non-conscious attitudes are measured using a timed trait rating procedure to measure feelings of implicit closeness towards African Americans and White Americans. Experimental results (N=555) support the predictions derived from the computational model. They suggest that socially shared racial attitudes (as measured by implicit priming) are biased in a pro-White and anti-Black direction, even among non-White participants. This bias is interpreted as a subconscious remnant of old racist norms. On the idiosyncratic level, however, an entirely different picture emerges. Attitudes that are measured by the timed trait rating procedure are much more favorable toward African Americans. A multivariate analysis suggests that implicit closeness to Blacks drives support for race related policies such as affirmative action. Thus, idiosyncratic racial attitudes may be able to overcome the lingering effect of socially shared racist attitudes. The implications of the theoretical model and the empirical findings of this study are discussed and future research projects are proposed.