Background: A core concept of social epidemiology is that societal factors are major determinants of individual well-being and life course, and therefore of population health and health disparities. The social environment, group and network-based relationships and processes, is grounded in the places where people live, manage households and daily life, learn, work and socialize. Yet because people are clustered and similar households sort into neighborhoods, social environments are complex sets of correlated factors and pose analytical challenges. Objectives: 1) Explore sets of objective and subjective measures of social environment for African-American women participating in cancer disparities research. 2) Examine relationships of built environment to social environment measures from multiple sources at neighborhood and individual levels. Methods: Combine indicators from multiple sources using generalized latent variable modeling (GLVM) to estimate correlated measurement error and test for relationships across multi-level model of key features in social environments that affect individual stress-related biomarkers via subjective measures of psychosocial stress and loneliness. Results: Analyses suggest clustering of social environment indicators and associations with psychosocial stress. Results from proposed GLVM using these indicators with community level indicators of the social environment will be presented. Discussion: Being able to measure and evaluate associations between physical and social environments of neighborhoods and individuals has implications not only for elucidating causal pathways of population health disparities, but also for making the case for the critical role of public health considerations in local governance and in evaluating effectiveness of interventions to restore the social fabric of neighborhoods at risk.