Date of Submission

2006

Degree Type

Dissertation

Degree Name

Doctor of Psychology (PsyD)

Department

Psychology

Department Chair

Robert A. DiTomasso, Ph.D., ABPP

First Advisor

Frederick Rotgers, Psy.D., ABPP, Chairperson

Second Advisor

Robert A. DiTomasso, Ph.D., ABPP

Third Advisor

Glenn P. Walters, Ph.D.

Abstract

This correlational study investigates the incremental validity of the Reactive (R) composite scale of the Psychological Inventory of Criminal Thinking Styles (PICTS) relative to Psychopathy Checklist: Screening Version (PCL:SV) total score for predicting medium security federal adult male inmate adjustment over a 12 month period. Adjustment, the criterion variable, was operationalized as the number of incident reports received over 12 months. Archival data for 146 offenders were obtained for incident reports and categorized as aggressive or non-aggressive. Scores for the PICTS P and R composite scales were obtained from archival electronic files. Data for rating the PCL:SV were gathered from each inmate’s presentencing investigation file. Two statistical models were employed for analyzing the eight hypotheses of this study. The first model included the traditional linear correlation and logistic regression analyses. The second model, which included univariate and multivariate negative binomial regression analyses, was implemented to specifically address the unique qualities of the distribution of the criterion variable. In the preliminary part of this correlational investigation, five demographic variables (age, education level, ethnic status, marital status, type of confining offense) and two psychometric variables and selected subscales (PCL:SV total, Part I and Part II scales and the PICTS P and R scales) were assessed to determine their relationships with the criterion variable. The primary hypothesis investigated incremental validity between the PCL:SV total score and the PICTS R composite scale. The findings using the two statistical models indicated that the logistic regression provided limited findings; however, the negative binomial regression, which in addition to replicating the findings of the logistic regression model, found additional significant findings. The primary hypothesis was partially supported in that significant incremental validity was found for the PCL:SV total score and the PICTS R composite scale for total incident reports, however not beyond age.

Share

COinS