Measuring Trajectories of Neighborhood Poverty in California, 1970-2009

Claire Margerison-Zilko, University of Texas at Austin
Catherine Cubbin, University of Texas at Austin

Background. Most literature examining neighborhood socioeconomic status (SES) influences on health uses cross-sectional measures of neighborhood SES, which do not capture dynamic processes such as deterioration and gentrification. Methods. We compared four methods of measuring neighborhood poverty in California from 1970-2009 using the Neighborhood Change Database (NCBD) and the American Community Survey 2005-2009 (ACS). We categorized census tract poverty based on 1) percent poverty in 2009, 2) a priori defined poverty trajectories, 2) latent class growth modeling (LCGM), and 3) non-parametric clustering. We compared these categories to other neighborhood-level variables. Results. Defining poverty trajectories a priori allowed us to distinguish between early and late deterioration and gentrification processes. LCGM highlighted differences between long-term concentrated poverty and recent poverty. Cluster analyses illustrated that almost half of neighborhoods exhibited deterioration over time, with substantial variation in patterns of deterioration. Next steps. We will examine associations between neighborhood poverty trajectories and health outcomes.