Robust estimates of the price of crime, measured as the costs of crime to victims, inform a wide range of policy analysis. The most commonly cited studies are constrained by limited data and rely on indirect methods to estimate prices. In these studies, health statistics are used to estimate direct losses from crime, jury award data are used to estimate indirect damages from crime, and self-reported crime data are used to weight injury prevalence within broad crime categories. While the relationship between injury and damages can be observed at the individual level in civil court records, individual level data have not previously been available that link crimes and injury. Since both individual and aggregate data are combined in these studies, prior research has not corrected sampling bias, and the estimates of victimization costs have been reported only as point estimates without confidence intervals. Estimates have been developed for only a few broad categories of crime and these estimates have not been robust to study design.

This study analyzes individual-level data from two sources: jury award and injury data from the RAND Institute of Civil Justice and crime and injury data from the National Incident-Based Reporting System. Propensity score weights are developed to account for heterogeneity in jury awards. Data from the jury awards are interpolated onto the NIBRS data based on the combination of all attributes observable in both data sets. From the combined data, estimates are developed of the price of crime to victims for thirty-one crime categories. Until data become available linking information about criminal incidents to jury award data, the strategy used here is likely to yield the most robust estimates of the costs to crime victims that can be generated from the jury compensation method.