Dimensions of Structural Disadvantage: A Latent Class Analysis of a Neighborhood Measure in Child Welfare Data

Abstract

ABSTRACT Contextual constructs, such as neighborhood structure, may contribute to child welfare involvement. Secondary data analysis is used with the nationally representative, longitudinal National Survey of Child and Adolescent Well-Being-II (N = 5,872). This study employs latent class analysis (LCA) to identify the number of classes needed to explain the distribution of caregiver responses on the Community Environment Scale. LCA is ideal for this study to identify meaningful groups of caregivers involved with child welfare using neighborhood risk factors. Three latent classes are identified: high social order/medium social capital; high social order/low social capital; low social order/low social capital. Multinomial logistic regression tests whether there are significant differences across the classes, partially validating the LCA that poor, minority caregivers live in neighborhoods with lower social order and capital. Understanding neighborhoods as “high” versus “low” risk may not fully illuminate contextual risk factors in order to develop neighborhood-based interventions. This study reveals that there is a third group of caregivers who reside in places with higher social order but lower social capital. Social capital might be an important factor in preventing child maltreatment. Future work is needed to understand additional individual and neighborhood characteristics that predict membership in each class.