Implicit biases, differential attitudes towards distinct groups, are prevalent and cause inequities. This research uses urban scaling theory and complex system modeling to predict that larger, more diverse, and less segregated cities exhibit lower implicit biases. Analysis of Implicit Association Test data from 2.7 million individuals across U.S. cities over a decade, combined with U.S. Census data, supports these predictions. Short-term changes in city social environments precede bias changes, but this relationship is bidirectional over longer timescales, suggesting that city social organization influences the strength of implicit biases.

Andrew J. Stier, Sina Sajjadi, Fariba Karimi, Luís M. A. Bettencourt, Marc G. Berman