Jonson-Reid, M., Drake, B., & Zhou, P. (2012). Neglect subtypes, race, and poverty: Individual, family, and service characteristics. Child Maltreatment 18(1) 30-41.
This was a detailed analysis of the relationships between race, poverty, child neglect and initial child welfare response, with two particularly interesting features: 1) the simultaneous consideration of two levels of poverty—family poverty and community poverty, and 2) the comparison of seven subtypes of neglect. The sample comprised child abuse/neglect reports, for families first reported in 1993-94, that were investigated in a Midwestern metropolitan area; only children categorized as Black (N = 4,296) or White (N = 2,522) were included, as there were too few children in other racial groups.
A major finding was that the context of poverty for Black and White children was very different; Black children in the welfare system live in far poorer communities than White children. The authors argue that “the importance of this finding is substantial,” and that both the family and community levels of poverty should be considered in cases of maltreatment. The use of family poverty alone is likely to systematically understate the problems faced by Black children, which may also reflect community factors, such as resource availability and levels of social disorganization. Specific neglect categories reported for Black and White children also showed a number of significant differences—for example Black children were more commonly reported and substantiated for severe and basic needs neglect.
Much of the continuing analysis of racial disproportionality has focused on agency response to Black children. These current findings underline the importance of the effects of income and other risk factors, and the need to “focus policy on ameliorating such disparities in living conditions.” The authors also speculate that neglect categories “could potentially be turned into a checklist” to guide workers in selecting the most appropriate intervention.
All data were extracted from electronic information systems in state and regional agencies. Address data were geocoded to allow for the linkage of administrative records from multiple agencies to census data. Use or non-use of Aid to Families with Dependent Children (AFDC) served as a dichotomous proxy measure of family poverty; census tract data (median income, percentage below poverty and residential mobility) were used as measures of community poverty. This use of exclusively administrative data was in one sense a limitation; most child maltreatment events are not officially reported, so results cannot be generalized to the entire population of maltreated children.
In the analysis, Black and White children were treated as separate, parallel samples (Tables 2 & 3) as opposed to the more commonly-used procedure of treating race as a control variable in a single sample. Groups were compared for child, family and community characteristics by the use of standard bivariate procedures (crosstabs & chi-square, t-tests and one-way ANOVA). Logistic regressions, including control for clustering at the community level, were used to analyze possible predictors of agency action (substantiation, need for service). Overall, this analytical approach appeared to be appropriate and well-conducted.