By George Crowne
Motivation
In an increasingly fragmented America, the question of how to socially integrate different groups of people is growing in importance. Research has shown that cross-group social integration has effects on a variety of outcomes ranging from educational attainment to overall health to income inequality. One theory for reducing cross-group bias stems from Gordon Allport’s seminal work on the effects of physical intergroup contact. Allport’s theory suggests that under certain conditions, intergroup contact can decrease intergroup biases by reducing the salience of group differences and increasing the frequency and quality of interactions. Additionally, influential work over half a century ago by Jane Jacobs emphasizes the impacts of the urban built environment on mobility patterns and social behavior. But can urban form affect the people with whom we interact? Do interactions caused by effective urban design meet the conditions for Allport’s contact hypothesis?
Hypothesis and Methodology
In my thesis, I empirically examine the hypothesis that the urban built environment can promote interactions between individuals in different socioeconomic groups, and that these interactions can reduce class-based friending bias. To this end, I study three features of the urban built environment that Jacobs’ framework implies might increase random interactions: small city blocks, entertainment place density, and mixed primary-use buildings. I begin by showing that these three features strongly predict experienced cross-class interactions, as measured by individual-level GPS cellphone mobility data. This result reinforces Jacobs’ theory that effective urban form can create random, unplanned interactions. I then turn to the second step in the causal chain: do these random cross-class interactions also increase cross-class social cohesion? I show that experienced cross-class interactions are strongly correlated with greater cross-class relationships, as measured by large-scale Facebook friendship data. Next, I find a similar relationship when I instrument cross-class interactions with the three features of the urban built environment, implying a causal effect of physical interactions on cross-class bias, and supporting Allport’s contact hypothesis.
Discussion
These results suggest that city design can play a significant role in promoting cross-group interactions and reducing group based bias. By designing cities in ways that promote unplanned interactions we can create opportunities for people from different socioeconomic backgrounds to interact and create cross-class friendships. With approximately 80% of the US population, and a growing fraction of the global population living in cities, these results can have significant implications for upward economic mobility, income inequality, and other positive effects of increased cross-group social integration.