Do measures of child protection recurrence obscure the differences between reporting and substantiation?

Abstract

After an initial child protection involvement, the same factors predict all types of recurrence including subsequent reports and subsequent substantiations. It has been concluded that there are no meaningful differences between children reported and substantiated because child protection decisions are determined primarily by availability of evidence rather than genuine differences in underlying need. However, an alternative explanation is that differences exist between children most likely to be reported and most likely to be substantiated, but that these differences are obscured in measures of recurrence. The aim of this study was to test this alternative explanation. Administrative data used in a previous recurrence study conducted in Queensland, Australia were reanalysed. A sample of 9608 children subject to their first ever screened-in report in 2010–11 were followed for 12 months. Odds ratios were calculated to measure bivariate associations between 21 independent variables and decisions to report and substantiate harm. Despite finding no differences in the factors that predicted different types of recurrence, different factors predicted high rates of reporting and high rates of substantiation. Furthermore, children subject to high rates of reporting were relatively unlikely to be subject to high rates of substantiation. There are meaningful differences in the factors associated with reporting and substantiation. The similarity in factors associated with different types of recurrence is an artefact of the way recurrence measures are constructed. Findings have implications for performance measurement and risk assessment instruments constructed and tested using measures of recurrence, which are poor indicators of need for statutory intervention.