Waves I-V Neighborhood Concentrated Disadvantage Data for Longitudinal Analyses User Guide

This document provides an overview of the Add Health ancillary study entitled, &ldquo;Wave I-V Neighborhood Concentrated Disadvantage Data for Longitudinal Analyses&rdquo; and its accompanying dataset (w5nbrhddisad.xpt). The data file contains five census indicators of neighborhood concentrated disadvantage, as well as a weighted composite score, for each of the five Add Health Waves (30 variables total). These data are linked to Add Health participant addresses and reflect concentrated disadvantage at the census block group level. The five indicators of neighborhood concentrated disadvantage are listed below. <ol> <li>Proportion of persons below the federal poverty level (last year)</li> <li>Proportion of persons 16+ years old in the labor force and unemployed</li> <li>Proportion of persons 25+ years old with less than a high school diploma</li> <li>Proportion of households on public assistance (last year)</li> <li>Proportion of households with children that are single-headed</li> </ol> These data differ from existing contextual data available from Add Health in several key ways, such that 1) the census block group boundaries are standardized to 2010 boundaries, 2) the values of each indicator variable are linearly interpolated to estimate the values of intercensal years and align with the year of Add Health assessment (e.g., 1994 values of the proportion of persons below the federal poverty line were estimated from 1990 and 2000 decennial census values), and 3) weighted composite variables of neighborhood concentrated disadvantage are included. To calculate the composite neighborhood concentrated disadvantage variables, a series of principal component analyses were conducted to extract weights for each indicator. Each indicator was then multiplied by the extracted weight, and the weighted items were then summed.