Examining Validity of Three Self-Report Measures of Criminogenic Risk
Presentation Type
Poster
Faculty Advisor
Chistopher King
Access Type
Event
Start Date
26-4-2024 11:15 AM
End Date
26-4-2024 12:15 PM
Description
The use of structured tools to assess individuals’ risk for repeat crime has become standard practice through the criminal legal system. Most of these tools are completed by evaluators, and are hence labor intensive and thus many not always feasible. However, a few self-report measures of criminogenic risk, with practical potential, have been developed and have shown some psychometric promise. The current study examined the convergent and predictive utility of three such tools. The sample consisted of 338 who were incarcerated and administered three self-report criminogenic risk assessment measures. Participants were tracked for rearrest and reconviction for 3.5 to 5 years, during which 21% were rearrested and 19% reconvicted. Convergent validity across total risk scores yielded by the three measures was examined using correlations; predictive utility was examined using correlations and logistic regression. All three risk scores correlated significantly, positively, and strongly with one another, and all three correlated significantly, positively, and weakly with reconviction. However, only one of the measures significantly, positively, and weakly correlated with rearrest. In a multivariable model, the three measures did not evidence significant incremental predictive utility over one another for reconviction. Findings suggested good convergent and modest predictive validity of the three self-report tools. The latter set of results may have been impacted by the relatively low base rate of criminal recidivism in the sample. Beyond additional predictive studies, examining how self-report tools may aid the treatment planning process would be another worthwhile future direction.
Examining Validity of Three Self-Report Measures of Criminogenic Risk
The use of structured tools to assess individuals’ risk for repeat crime has become standard practice through the criminal legal system. Most of these tools are completed by evaluators, and are hence labor intensive and thus many not always feasible. However, a few self-report measures of criminogenic risk, with practical potential, have been developed and have shown some psychometric promise. The current study examined the convergent and predictive utility of three such tools. The sample consisted of 338 who were incarcerated and administered three self-report criminogenic risk assessment measures. Participants were tracked for rearrest and reconviction for 3.5 to 5 years, during which 21% were rearrested and 19% reconvicted. Convergent validity across total risk scores yielded by the three measures was examined using correlations; predictive utility was examined using correlations and logistic regression. All three risk scores correlated significantly, positively, and strongly with one another, and all three correlated significantly, positively, and weakly with reconviction. However, only one of the measures significantly, positively, and weakly correlated with rearrest. In a multivariable model, the three measures did not evidence significant incremental predictive utility over one another for reconviction. Findings suggested good convergent and modest predictive validity of the three self-report tools. The latter set of results may have been impacted by the relatively low base rate of criminal recidivism in the sample. Beyond additional predictive studies, examining how self-report tools may aid the treatment planning process would be another worthwhile future direction.