Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

William J. Fremouw

Committee Co-Chair

Michael Crabtree

Committee Member

Cheryl B. McNeil

Committee Member

Aaron Metzger

Committee Member

Tracy L. Morris


The purpose of the present study was to investigate the validity and reliability of the Child Abuse Potential L scale with a heterogeneous sample of caregivers referred for parenting capacity evaluations. One aim of the study was to assess the measurement properties of the L scale. A second aim was to evaluate the discriminative validity of the L scale by way of its receiver operating characteristics. A third aim of the study was to examine potential correlates of desirable responding on the L scale. The findings from this study provide new information about the psychometric properties of the CAP L scale and its application in clinical and forensic settings. Consistent with past investigations, caregivers produced a high rate (74.4 %) of invalid CAP profiles by way of elevated L scale scores. The L scale showed little variation across caregivers from families with different maltreatment histories. Item analyses and estimates of internal consistency showed homogeneity of the L scale, though several problem items were identified. Deletion of these items, however, produced only marginal improvements in internal consistency. The 14-item revised scale that resulted from the item deletions showed tradeoffs in sensitivity and specificity compared to the original 18-item scale. Classificatory accuracy of the 18-item scale (with emphasis on sensitivity to detect fake-good responding) was best using a cutoff score that was one to two points higher than recommendations given in the CAP manual (Milner, 1986). Last, the L scale showed inverse associations with stress and aggression. These findings suggest that caregivers perceive the context of evaluation to be coercive, pointing up the importance of procedures and pacing that increase rapport. Also, caregivers who report low levels of anger and stress produce higher L scale scores. Furthermore, findings highlight tradeoffs in L scale sensitivity and specificity that evaluators can select as a function of referral question or other relevant considerations. In sum, findings add to the scientific merit of the CAP in relation to Daubert criteria for testimonial admissibility.