NCJ Number
203613
Journal
Criminal Behaviour and Mental Health Volume: 13 Issue: 2 Dated: 2003 Pages: 140-150
Date Published
2003
Length
11 pages
Annotation
Since many professionals believe that self-report questionnaires used to predict recidivism have a low validity, this study tested the assumption that the validity of self-report data are vulnerable to self-presentation biases in offender samples.
Abstract
A total of 124 male offenders volunteered to participate in the study during the process of completing psychological assessments associated with release decisions. They were incarcerated in various institutions in the Ontario Region of the Canadian Federal Correctional System. The ages ranged from 18 to 54 years (mean of 28.9 years). The Self-Appraisal Questionnaire (SAQ) consisted of 72 "true" or "false" items and 7 clinical subscales and a validity subscale. Offender responses were compared with the criminal record to check for inaccuracies in responding. Participants were also administered the Balanced Inventory of Desirable Responding (BIDR), a 40-item self-report measure of the tendency to give socially desirable responses on self-reports, and the General Statistical Information on Recidivism (GSIR), which was initially developed as an actuarial instrument to predict recidivism among federally sentenced Canadian offenders. The study findings indicate that self-report data, despite the presence of socially desirable responding, can be an effective predictor of recidivism. This is due, in part, to the confirmed hypothesis that the SAQ's variance associated with risk exceeds the variance associated with social desirability; however, the relationship between social desirability and recidivism may explain why, despite the presence of social desirability, self-report can be an effective predictor of recidivism. 4 tables and 33 references