Friday, December 9, 2011
Leontien van der Knaap , Marise Ph. Born , L. E. W. Leenarts and P. Oosterveld (INTERVICT, Tilburg University , affiliation not provided to SSRN , affiliation not provided to SSRN and affiliation not provided to SSRN) have posted Reevaluating Inter-Rater Reliability in Offender Risk Assessment (Crime & Delinquency, 2010) on SSRN. Here is the abstract:
Offender risk and needs assessment, one of the pillars of the Risk-Need-Responsivity model of offender rehabilitation, usually depends on raters assessing offender risk and needs. The few available studies of inter-rater reliability in offender risk assessment are, however, limited in the generalizability of their results. The present study examined inter-rater reliability in Dutch offender risk assessment of 38 raters who independently assessed 75 offenders. Results show substantial reliability (Tinsley & Weiss’ T value ≥ .61) for risk of reconviction and moderate (T value ≥ .41) to substantial reliability for offender needs, such as accommodation, finances, or education. These results are discussed in light of a recent British study on the inter-rater reliability of a comparable risk assessment instrument. Results from the present study show similar to better reliability, leading to the conclusion that greater external validity does not negatively influence inter-rater reliability results.