Invalidity of "Disconfirmation of the Predictive Validity of the Self-Appraisal Questionnaire in a Sample of High-Risk Drug Offenders" (2006): A Reply  

Oleh:

Dhaliwal, Gurmeet K. ; Loza, Wagdy ; Reddon, John R.

Jenis: Article from Journal - ilmiah internasional
Dalam koleksi: Journal of Interpersonal Violence vol. 22 no. 8 (Aug. 2007), page 1077-1090.
Topik: Offender Risk Assessment; Psychometrics; Self-Appraisal Questionnaire
Fulltext: 1077.pdf (93.05KB)

Isi artikel

In their article, “Disconfirmation of the Predictive Validity of the Self- Appraisal Questionnaire in a Sample of High-Risk Drug Offenders,” criminologists Mitchell and MacKenzie (2006) purported to evaluate psychometric properties of the Self-Appraisal Questionnaire (SAQ; Loza, 1996, 2005), the first self-report risk/need measure estimating violent and nonviolent recidivism. There are several serious problems with this study, which dispute their conclusions such as “disconfirmation” of the SAQ. Briefly, in their first study, Mitchell, MacKenzie, and Perez (2005) used the SAQ as a pre- and postprogram measure to evaluate antisocial attitudes and cognition of adult male offenders randomly assigned to either a boot camp program (TBC; N = 102) or program in a correctional facility (MTC; N = 93) that did not employ the military-based component of boot camp. They found no differences in SAQ Total or Criminal Tendencies subscale scores for offenders who participated in either program. This finding did not support their hypothesis that boot camps would be more effective in reducing antisocial attitudes than identical programming in a correctional setting. Consequently, in their second study, they combined the two groups of offenders and evaluated the predictability of the SAQ with 159 of these offenders who completed all of the SAQ items and were in the community for at least 90 days before recidivism data were collected. They reported that SAQ Total scores exhibited high reliability but subscales revealed “substandard” reliability. Using tetrachoric correlations and a “parallel” analysis derived from Pearson product–moment correlations with continuous data, they suggested that all but one scale were multidimensional. In addition, they relayed that the SAQ failed to predict rearrest whereas age and number of prior arrests were better predictors. In conclusion, these authors “cast considerable doubt on the reliability, factorial validity and predictive validity of the SAQ in similar samples” (p. 462) that they considered “high-risk drug offenders” regardless of the various analytical techniques used. Inadequacies in the Mitchell and MacKenzie’s (2006) sample selection process, their questionable methodology, and the problematic analytical techniques they employed obfuscates their findings and invalidates their conclusions.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali