Ak. Cumberland et Gj. Boyle, PSYCHOMETRIC PREDICTION OF RECIVIDISM - UTILITY OF THE RISK NEEDS INVENTORY, Australian and New Zealand journal of criminology, 30(1), 1997, pp. 72-86
The efficacy of psychometric measures of recidivism is an important is
sue. Accordingly, the present study examines: (1) the predictive (crit
erion-related) validity of the Risk Needs Inventory (RNI); (2) whether
RNI scores alone are better predictors than are scores moderated by C
ommunity Corrections Officers (professional override principle); and (
3) whether inclusion of age and marital status as independent variable
s improves predictive validity. It was found that RNI total and risk c
ategory scores for both the unadjusted and override conditions signifi
cantly predicted both number and severity of reoffences for in-program
and total recidivism measures. Age predicted number and severity of r
eoffences (inversely), but generally did not add significantly to the
variance accounted for by the RNI items. While the RNI appears to have
some utility, the prediction of recidivism remains a complex psychome
tric issue. However, it is likely that use of the RNI provides a more
structured/objective assessment of recidivism risk. Although the varia
nce accounted for is small (8-15%), nonetheless, use of the RNI result
s in correct classification of the majority of offenders. In addition,
use of professional override would appear to be essential in order to
differentiate between the reoffending behaviours of medium and high r
isk groups.