A Spatial Analysis Comparing Urban and Rural Neighborhood Characteristics and Maltreatment

Background and purpose:

Although most child maltreatment research has occurred in urban areas, the proportion of rural children who are maltreated is substantial and often greater than that for urban areas. In Michigan, the average overall maltreatment rate for rural census tracts is 31 per 1,000 children, compared to 16 per 1,000 for urban tracts. Maltreatment is influenced by characteristics of the context in which families live. In urban settings, neighborhood conditions and child maltreatment are clearly connected. Aspects of neighborhoods such as poverty rates (neighborhood structural factors) influence the way in which neighbors interact with each other (neighborhood processes), which in turn, affects child maltreatment. ¹ The impact of geographic distance on these relationships is unknown. Our knowledge about maltreatment and factors that shape it are largely urban based. It is unknown whether community risk and protective factors of maltreatment identified primarily in urban areas are applicable to rural areas and what unique characteristics of rural neighborhoods relate to maltreatment. The current study sought to examine the extent to which neighborhood factors commonly found to be related to maltreatment in urban areas apply to rural areas.

Methods:

The data are from the Michigan Statewide Automatic Child Welfare Information System. We use child maltreatment investigations per 1000 children within a census tract to measure maltreatment. Neighborhood variables were downloaded from the US Census Bureau American Community Survey. We first ran an exploratory factor analysis using key neighborhood variables found to be related to child maltreatment in urban areas, including poverty, unemployment, percent single headed households, percent of residents moving into the neighborhood, percent vacant housing, ratio of children to adults, and ratio of elderly to other adults. We assess and compare the factors contributing to maltreatment in rural and urban areas using negative binomial mixed effect models.

Findings:

We found three key factors in urban census tracts, socioeconomic disadvantage, residential instability, and childcare burden. These factors are consistent with extant literature on urban neighborhoods. Within rural census tracts, these factors do not align. Childcare burden consistent across the two contexts, but residential instability and poverty hung together and did not represent unique factors. Single headed-households did not map onto any other neighborhood-level variables. The study further finds that poverty and vacant housing rates are related to child maltreatment in rural census tracts (similar to urban tracts); but contrary to findings in urban areas, unemployment was not related to maltreatment in rural areas.

Implications:

This study found important differences between urban and rural census tracts in terms of the relationships between neighborhood factors and child maltreatment. These findings suggest that efforts to prevent maltreatment may need to be tailored for rural areas, especially those interventions targeted at the neighborhood level. The study suggests that findings from urban neighborhood research literature and the measures used to generate these findings may not apply within the rural context. Additional research is needed to understand rural child maltreatment.