Poverty has been identified as one of the strongest risk factors for child maltreatment. However, it is unknown whether this relationship has become stronger or weaker over time. To address this knowledge gap, we examined whether the county-level association between child poverty (CP) rates and child maltreatment report (CMR) rates changed over time in the United States from 2009-2018. We examined overall CMR rates (i.e., rates among all children), as well as CMR rates specific to subgroups of child sex, age, race/ethnicity, and maltreatment type (e.g., rates among age 0-5 children).
Methods:
CP rates (percentages of children in poverty) per county per year were obtained from the American Community Survey data. CMR rates (numbers of children with a screened-in report per 1,000 children) per county per year were computed based on the National Child Abuse and Neglect Data System Child Files. We included year dummies (Y2010, Y2011, ..., Y2018; reference year=2009) and their interactions with CP rates (CP×Y2010, CP×Y2011, ..., CP×Y2018) in models to examine how the CP coefficient in each year from 2010-2018 differed from the CP coefficient in 2009. We used subgroup-specific CP rates (e.g., age 0-5 CP rates) for subgroup-specific CMR rates (e.g., age 0-5 CP rates). We used linear multilevel models with state-level random effects to handle the nesting of counties in states. The nesting of county-year observations in counties was handled by the year dummies (i.e., year fixed effects). We adjusted for a range of controls, including % Black, % Latino, % foreign-born, % child, % elderly, % male adult, % children with disability, % moved, and urbanicity.
Results:
The CP coefficient was 0.96 in 2009, meaning that when the CP rate increased by 1-percentage-point, the overall CMR rate statistically significantly increased by 0.96 per 1,000 children. The CP coefficient grew over time almost linearly and became 1.48 in 2018, meaning that per 1-percentage-point increase in the CP rate, the overall CPR rate statistically significantly increased by 1.48 per 1,000 children. We found similar growing trends in the CP coefficients for most subgroup-specific CMR rates, including CMR rates among male, female, age 0-5, age 6-11, age 12-17, White, and Black children, as well as for neglect and physical abuse report rates. The CP coefficient on Latino CMR rates and sexual abuse report rates changed little over time. We conducted sensitivity analyses using negative binomial multilevel models, and the findings were consistent between the linear and negative binomial models.
Conclusions/Implications:
We found that the county-level relationship between CP rates and CMR rates became stronger over time in the United States, not only overall, but also within most subgroups (child sex, age, race/ethnicity, and maltreatment type). Our findings suggest that poverty may have become more detrimental to the parenting and safety of children. Greater attention to lowering CP rates and buffering their impacts on children may be increasingly beneficial given the stronger CP-CMR relationship in recent years. Further research is required to confirm our findings, especially at the individual level.