Toward a multidimensional understanding of residential neighborhood: a latent profile analysis of Los Angeles neighborhoods and longitudinal adult excess weight

Abstract

People are embedded in a complex socio-spatial context that may affect their weight status through multiple mechanisms, including food and physical activity opportunities and chronic stress exposure. However, research to date has been unable to resolve what features of neighborhoods are causally related to weight status. We used latent profile analysis to identify three "types" of neighborhoods (based on five dimensions of neighborhood social status) in Los Angeles, CA. Our neighborhood types were both substantively interpretable and predictive of excess weight in both cross-sectional and longitudinal models. Our results are promising for a research community attempting to operationalize neighborhoods as multidimensional, complex systems.

Keywords: Latent profile analysis; Longitudinal model; Neighborhood effects; Obesity.