Maguire-Jack, K., Lanier, P., Johnson-Motoyama, M., Welch, H., & Dineen, M. (2015). Geographic variation in racial disparities in child maltreatment: The influence of county poverty and population density. Child Abuse & Neglect, 47, 1-13.
There is ample evidence that Black children in the US are overrepresented in the child welfare system. Possible reasons for this—systematic bias or socioeconomic risk factors—have been studied. Current evidence indicates that discrimination has put Black families at a significant economic disadvantage; poverty, at both family and community levels, has then increased the risk of maltreatment.
This present study extends the analysis to include an additional community variable, population density, reported on a 9-point scale from 1= metro (≥ million) to 9 = rural (≤ 2,500). The relationships between population density, poverty and child maltreatment were compared for White, Black and Hispanic children, using US national data with county as the unit of analysis. Outcome variables were disparity ratios (DRs)— poverty and maltreatment rates for Black and Hispanic children were divided by the corresponding rates for White children; higher maltreatment DR indicated more serious overrepresentation. The principal findings were as follows:
• Maltreatment DRs were highly significantly correlated with poverty DRs for both minority groups.
• Both poverty and maltreatment DRs had a curvilinear relationship with population density; highest for metro counties and lowest for suburban counties, with a small uptick for rural counties.
• Maltreatment DR patterns for the two ethnic groups were different; Black DRs were higher in metro settings, Hispanic DRs were higher in rural settings. In suburban settings, Hispanic DRs dropped below 1.00, suggesting maltreatment rates even lower than rates for White children.
• This ethnic difference was reflected in multivariate regression analyses. Poverty and population density were significant predictors of maltreatment DR for both ethnic groups, but for Hispanics the correlation with population density was negative.
The authors suggest several possible ways in which geographic area might be driving these maltreatment disparities. They argue that future research must look for actual mechanisms through which neighborhood and individual poverty work, and that policy-makers must prioritize programs aimed at reducing poverty.
Data were obtained by linking various sources, using the Federal Information Processing Standard (FIPS): the NCANDS Child File (2011); the National Center for Health Statistics (2014) and the US Census American Community Survey (2008-2012). Population density data came from the US Dept. of Agriculture (2013) Rural-Urban Continuum Codes (RUCC).
The authors note several limitations; for one thing, the Hispanic category in the NCANDS data lumps together an ethnically and culturally diverse group, possibly a factor in the reversed population density effect. However, the methodological approach used here seems to have been unusually elaborate and painstaking. For example: spatial analysis was used to explore the effect of geographical location; sensitivity tests were performed to examine the effect of different cutpoints for the number of individuals of a specified race in each county; county-level findings were replicated, using data pooled at state and national level; exploratory descriptive mapping was used to examine the distribution of DRs across the US. All this tends to support the validity of the reported results.