Neighborhood Contexts Associated with Internalizing Behaviors of Latino and African American Youth: Evidence from Denver, Colorado

Tuesday, 17 July 2018: 18:00
Oral Presentation
Anna Maria SANTIAGO, Michigan State University, USA
National and local studies in the United States report that more than one in five children exhibit behavioral health problems falling in the borderline or clinical range. Previous studies have focused on exposure to interpersonal and community violence as antecedents to such problems; more recent literature has focused on other neighborhood factors that might lead to the development of internalizing behavioral problems (depression, anxiety and PTSD) -- among the most common behavioral health problems during childhood. These studies find that children residing in neighborhoods of concentrated disadvantage, limited institutional resources, high fractions of unsafe or poor-quality housing, and high levels of social disorder are more likely to experience childhood depression; anxiety; and PTSD . Yet, numerous questions remain as to the magnitude and mechanisms of purported neighborhood effects and the extent to which they might vary by ethnicity.

In this paper, data from a natural experiment in Denver (CO) are used to assess the relationships between neighborhood contexts and internalizing behaviors for a sample of 600 Latino and African American youth aged 7 to 18 who spent a significant part of childhood residing in subsidized housing. We use instrumental variables and accelerated failure time (AFT) models with frailties to estimate variations in the timing of internalizing behaviors diagnoses during childhood.

Findings suggest that 10% of all youth were diagnosed with internalizing behaviors after initial random assignment to neighborhood. Accelerated failure time models show that several dimensions of the neighborhood—safety, socioeconomic status, nativity, and residential instability—were strongly predictive of the acceleration or delay in being diagnosed with one or more internalizing behaviors. Results differ markedly by ethnicity with models being most predictive for African American youth.