Document Type



Doctor of Philosophy


Criminology and Criminal Justice

Date of Defense


Graduate Advisor

Janet Lauritsen, Ph.D.


Richard Rosenfeld, Ph.D.

Lee Ann Slocum, Ph.D.

Cathy Spatz Widom, Ph.D.


Figures from the National Child Abuse and Neglect Data System (NCANDS) suggest that national rates of child maltreatment declined during the last decade of the 20th century and into the 21st century. These data, which are derived from official state child protective service agency record systems, have frequently been used to measure changes in child abuse and neglect in the U.S. and in individual U.S. states. However, because the NCANDS has yet to be assessed for methodological issues surrounding the validity of the data to measure temporal change, it is unknown if the decline revealed in the NCANDS data reflects the true nature of the changes over time from 1990 to 2013 in the occurrence of child maltreatment. Using correlational and cointegration techniques, I find that child physical abuse and child sexual abuse rates at the national level are highly correlated in both level and trend with both national rates of child homicide and children exposed to violence in their home. I also find that most (though not all) states’ rates in child maltreatment correspond to state rates in child homicide, and that these two child victimization trends are cointegrated, sharing a significant long-run equilibrium with one another. Once the NCANDS data validity concern has been addressed, this project also assesses how several competing hypotheses (e.g., greater economic prosperity, increased number of agents of social control, increased use of psychiatric medications, higher incarceration rates) might account for state-level variations in child maltreatment trends. Using panel regression models, these findings suggest that at least some of the declines in rates of child victimization over this period are real and not due to artificial declines based on policy and procedural changes over time. I find that various factors are significantly associated with rates of child victimization from 1990 to 2013, with particular factors better suited for explaining trends in certain types of child victimization. Policymakers should be aware of these relationships and recognize that macro-level declines in rates of child victimization may only be possible if these other macro-level conditions are also addressed.

Included in

Criminology Commons