(Withdrawn) Understanding the Roles of Nonprofit Infrastructure, Gun Violence, and Social Disorganization in Child Maltreatment; An Ecological Study of Nonprofit Density and Child Maltreatment Rates

Background:

Advocates and researchers have emphasized the role of neighborhood processes, such as social disorganization, in understanding child maltreatment rates. Nonprofit organizations are important neighborhood actors that solve collective action problems and maintain a range of strategies to enhance neighborhood processes and support nearby families. In addition to direct services, nonprofits may facilitate the development or maintenance of social networks, as well as formalize advocacy efforts resulting in increased attention from government. The goal of this ecological study is to understand the relationship between nonprofit density and child maltreatment rates. Prior research has often emphasized the role of social services with mixed results; however, the sociology of density dependence suggests accounting for agglomeration is essential. Accordingly, this study also dis-aggregates nonprofits by six organizational types, capturing heterogeneous relationships.

Methods:

The study region is Cuyahoga County Ohio (USA). Aggregating by census tract, this study couples four primary data sources. The dependent variable, child maltreatment, includes substantiated and indicated cases of maltreatment from agency records. Indicators of social disorganization include counts of gun involved events, poverty, unemployment, and vacancy rates, as well as measures of residential mobility and diversity. Nonprofit density is measured using geolocated form-990 information to obtain nonprofit locations. Heterogeneity is addressed by classifying nonprofits based on The National Taxonomy of Exempt Entities, as primarily concerned with human services, education, religion, health, arts, or other. After excluding areas with no children, the final sample includes 442 census tracts. The study aims are addressed using hierarchical Poisson models, accounting for spatially structured and tract-specific variation, implemented through a Bayesian methodology.

Results:

Model comparison shows the spatial and non-spatial components are important to the model, and the majority of variation in child maltreatment rates is spatially structured. The results indicate the presence of nonprofits are protective factor: negatively associated with child maltreatment rates when pooled (posterior mean: -0.26, CI: -0.45, -0.10 +/- 1 SD). Additionally, the results suggest substantial heterogeneity by organizational type, as the presence of organizations is either negatively associated with child maltreatment rates (including education, religion, human services organizations), or show posterior density consistent with zero effect (health, arts, other).

Conclusions and Implications:

The results show a negative relationship between child maltreatment rates supporting nonprofits as a protective neighborhood factor. The results of the decomposition by organizational type suggest it is unlikely that the effects are driven entirely by human services nonprofits, which have been the focus of much prior research. Rather, substantial effects are estimated for education and religious organizations, as well. These organizational types are primarily, although not exclusively, concerned with promoting civic life and fostering social ties. For example, the presence of educational nonprofits presents community support and education opportunities that increase parental involvement, while religious institutions are strong community builders. The study emphasizes the importance of a strong nonprofit infrastructure, as well as the equitable distribution of organizations. Policy makers may use existing tools to incentivize alternative locations for new nonprofits or support entrepreneurship in neighborhoods with few formal organizations.