Substantiated Reports of Sexual Abuse Among Latino/Latina Children: A Multilevel Model of National Data with Implications for Social Work Practice

Background: Despite a rapidly growing Latino/Latina child population in the U.S., as well as in the child welfare system (CWS), limited research exists on Latino/Latina children reported to the CWS for child sexual abuse (CSA). Moreover, given that states and counties administer CWS services, state and regional differences in Latinos’/Latinas’ CWS outcomes may be masked when data is aggregated at the national level. To begin addressing these research gaps, this study examined the following research question: are Latino/Latina children reported for sexual abuse more likely to have a report substantiated compared to non-Latino/Latina Black or White children, even after for controlling for other factors?

Methods: We conducted secondary data analysis using data from the 2012 National Child Abuse and Neglect Data System Child File to examine a sample of children reported to the CWS for CSA (n = 179,199). We used generalized linear mixed modeling (GLMM) to estimate the likelihood of substantiation among children reported for CSA to determine whether race/ethnicity predicted substantiation when controlling for child age, child gender, the interaction of race/ethnicity and age, and the interaction of race/ethnicity and gender. The intra-class correlation (ICC) was calculated to determine the impact of state clustering on the variance in rates of substantiation, and nested models were compared to assess model fit using the Aikake Information Criteria (AIC) and deviance. We calculated the conditional modes of the random effects to predict the relationship between race/ethnicity and probability of substantiation for each state and across counties with sufficient Latino/Latina populations.

Results: Bivariate analyses indicated that Latino/Latina children reported for CSA were significantly less likely to have their report substantiated (20.49%) compared to both non-Latino/Latina Black (21.69%, p < .0001) and non-Latino/Latina White (22.75%, p < .0001) children nationally without controlling for other factors. Further, results of GLMM analyses indicated that non-Latino/Latina Black (OR = 1.12) and White (OR = 1.17) children reported for CSA had significantly higher odds of having substantiated CSA cases compared to Latino/Latina children controlling for other factors. Older children (in mean-centered years; OR = 1.09) and females (OR = 2.52) had higher odds of substantiation. We also found significant interaction effects between race/ethnicity by age and race/ethnicity by gender. The relationship between race/ethnicity and substantiation decreased as child age increased and also was lower for females compared to males.

Implications: Our findings indicate that Latino/Latina and non-Latino/Latina children involved in CSA cases experience significantly different patterns of case substantiation when compared with children of other races/ethnicities, and that these patterns vary depending on the age and gender of the child. Each of these patterns needs further exploration. However, results underscore the importance of attention to race/ethnicity, language, and culture in designing intervention and prevention strategies for children who experience or are at risk of experiencing CSA.