Testing the Predictive Validity of the LSI-R Using a Sample of Young Male Offenders on Probation in Guangzhou,China

Abstract

This study explored the predictive validity of the Level of Service Inventory–Revised (LSI-R) with 112 young male probationers on recidivism for new offense and rearrest in Guangzhou, China, from 2010 to 2013. Using bivariate correlations, logistic regression, and receiver operating characteristics (ROC) analysis, the LSI-R total score was found to be a stable and significant predictor of recidivism. The LSI-R total score and subscales of Criminal History and Leisure/Recreation were significantly correlated to recidivism. There was significant association between time at risk and recidivism. Increase in the LSI-R total score was associated with greater likelihood that recidivism would occur. There was no recidivism in low-risk group, and higher recidivism rates were found in moderate-and high-risk groups. The LSI-R was effective in specifying recidivists and non-recidivists with area under the curve (AUC) coefficient of .733. The current study supports the utility of the LSI-R, which is applicable to different cultural and social contexts, as a risk/need assessment instrument for young male probationer recidivism in Guangzhou. Sufficient training and accumulated experience of administering the LSI-R would be necessary to reduce regional variation. Future studies with larger samples across different offender groups are needed. Limitations and implications for offender assessment in China are discussed.