Short-term predictive validity of the static-99 and static-99-R for indigenous and nonindigenous Australian sexual offenders.

Résumé

Actuarial risk assessment (Static-99 and Static-99-R) scores were obtained for 399 Australian adult sexual offenders who were subsequently released from prison and followed up with searches of police arrest records (mean follow-up period = 29 months; range = 15-53 months). Indigenous offenders (n = 67; 16.8%) scored significantly higher on both the Static-99 (M = 4.04 vs. 2.89, p < .001) and Static-99-R (M = 3.72 vs. 2.22, p < .001), were more than twice as likely to be arrested for sexual offenses (9.0% vs. 4.1%, ns), and were significantly more likely to be arrested for nonsexual violent (28.4% vs. 1.9%, p < .001), any violent (including sexual; 37% vs. 5.9%, p < .001), and any offenses (58.2% vs. 21.6%, p < .001). For the combined groups, predictive accuracy of both instruments was comparable to results reported elsewhere. Predictive accuracy of the Static-99 was similar for indigenous and nonindigenous offenders. The Static-99-R was only marginally predictive of any violent recidivism (AUC = .65, 95% CI = [.52, .79]), and did not predict sexual (AUC = .61, 95% CI = [.45, .77]) or nonsexual violent recidivism (AUC = .65, 95% CI = [.48, .78]), for indigenous offenders. Higher risk scores, indigenous race, and unsupervised release all contributed unique variance to any violent recidivism. Results suggest that the Static-99 may be appropriate for assessing Australian indigenous sexual offenders, but more research is needed to test the validity of the Static-99-R for this population. We conclude that practitioners should consider the potential effects of racial differences and postrelease factors, as well as static risk factors, in their assessments.