Racial Differences in Networks: Do Neighborhood Conditions Matter?


  • *Direct correspondence to Mario L. Small, 1126 E. 59 St., Department of Sociology, University of Chicago, Chicago, IL 60637 〈mariosmall@uchicago.edu.〉 I thank Jeanette Chung, Sara McLanahan, Katherine Newman, Sandra Smith, and Christopher Winship for conversations and comments that have improved the quality of this article. I thank Wangyal Shawa, of Princeton University's GIS Library, for preparing the 1980 to 1990 Census tract translation file. The data and modeling information used to write this article are available on request to those wishing to replicate the findings.


Objectives. This study examines which of five neighborhood conditions help account for racial differences in social networks.

Methods. The data set is the Urban Poverty and Family Life Survey, a survey of blacks, whites, Mexicans, and Puerto Ricans clustered in Chicago Census tracts, matched to 1990 Census data. I estimate HGLM models predicting five indicators of social isolation and five indicators of number of social ties as a function of race, controls, and the following neighborhood conditions: neighborhood poverty, proportion black, residential stability, ethnic heterogeneity, and population density.

Results. Although initial estimates confirm the existence of racial differences in network size, most of these differences are not robust to controls for neighborhood conditions. Among the neighborhood variables, only neighborhood poverty is consistently associated with size of social networks.

Conclusions. Findings suggest that while residential segregation has created conditions in which some races are more likely to live in high-poverty neighborhoods, it is the poverty, not the racial composition, of the neighborhoods that is significantly associated with weaker social ties.