Reliability, Construct, and Concurrent Validities of the Self-Appraisal Questionnaire

Abstract

The Self-Appraisal Questionnaire (SAQ) is a recently developed self-report questionnaire designed to predict violent and nonviolent offender recidivism. This 67-item, six-subscale tool measures quantitative criminogenic risk-need areas. The SAQ was administered to 303 federally sentenced Canadian male offenders. The total scale test-retest reliability coefficient was .95 and ranged from .69 to .93 for the subscales. Coefficient alphas ranged from .42 to .87. All the subscales with the exception of one had a one-factor solution. The SAQ subscales correlated with other instruments assessing similar constructs and with other measures of recidivism. Offenders with high SAQ totals and subscale scores committed significantly more offenses than did those with low SAQ scores. Similarly, offenders with a history of violence had higher SAQ total scores than the offenders with no history of violence. These preliminary results support the further study of the SAQ as an instrument for predicting violent and nonviolent recidivism.

Predictive validity despite social desirability: evidence for the robustness of self‐report among offenders

“…Offender responses are compared with the criminal record sheet to check for inaccuracies in responding. The reliability, construct and concurrent validity of the SAQ have previously been demonstrated (Loza et al, 2000), as has the predictive validity of the SAQ over a two-year (Kroner and Loza, 2001;Loza and Loza-Fanous, 2001) and five-year period (Loza and Loza-Fanous, in press). Also, the SAQ was found to be at least as effective as four other well-established professional rated and widely used measures for the prediction of recidivism (Loza and Loza-Fanous, 2001).…”

“…Further, self-reported antisocial personality characteristics, such as those reported by Blackburn and Fawcette (1995) have been shown to predict violent recidivism (Quinsey et al, 1998) and Kuriychuk (1990) found concordance between Blackburn's self-reported scales and the PCL-R. Self-reported psychopathology has also been shown to be predictive of institutional adjustment problems in samples of both violent and sexual offenders (Mills and Kroner, 2003a). In addition, the Self Appraisal Questionnaire (SAQ: Loza, 1996), a self-report questionnaire which was designed specifically to predict offender general and violent recidivism, was repeatedly found to be reliable and valid and at least as effective in predicting offender post-release outcome when compared with four other widely used and professional rated measures (Loza et al, 2000;Kroner and Loza, 2001;Loza and Loza-Fanous, 2001). Paulhus (1984) holds to a two-dimensional nature of social desirability: impression management and self-deception.…”

“…The SAQ has been shown to be related to rated risk prediction instruments (Loza et al, 2000) therefore it contains variance associated with risk prediction. As a selfreport instrument it is assumed that it will also contain variance associated with social desirability.…”

Characteristics of offenders: the HEXACO model of personality as a framework for studying offenders’ personality

“…Indeed, some studies indicate that offenders have stronger tendencies than nonoffenders to give socially desirable responses to self-report items (Haapasalo, 1990). Despite this, other studies have demonstrated the validity of self-report measures for studying offender groups (Loza, Dhaliwal, Kroner, & Loza-Fanous, 2000;Loza & Loza-Fanous, 2002). For example, in the Loza et al's (2000) study, the authors compared prisoners' responses to a self-appraisal questionnaire with their criminal histories, and found that overall scores on the questionnaire distinguished violent from nonviolent prisoners and predicted their number of violent and nonviolent offenses.…”

The Impact of a Video Game on Criminal Thinking

“…Criminal Thinking (total score; maximum possible score of 185) and Cold Heartedness (subscale; maximum possible score of 25) were measured by the Texas Christian University Criminal Thinking Scale (Knight, Garner, Simpson, Morey, & Flynn, 2006). Criminal Risk (total score; maximum possible score of 120) and Conduct Disorder (subscale; maximum possible score of 11) were measured by the Revised Self-Appraisal Questionnaire (Loza, Dhaliwal, Kroner, & Loza-Fanous, 2000). In accordance with hypothesis 2, results indicate that participants in the criminal videogame condition were significantly faster at post-test than at pre-test at responding to the me-crime association than in the three control conditions (for values see Figure 2).…”