Impact and detection of response distortions on parenting measures used to assess risk for child physical abuse.

This study investigates the impact of instructional conditions (fake good, be honest, fake bad, and respond randomly) on the scores of 3 parenting measures: the Adult/Adolescent Parenting Inventory (AAPI; Bavolek, 1984), the Child Abuse Potential (CAP) Inventory (Milner, 1986), and the Parenting Stress Index (PSI; Abidin, 1995) in general-population parents and at-risk parents. In addition, the study explores the ability of the PSI Defensiveness Scale and the CAP Inventory validity indexes to detect response distortions. As expected, most parenting-measure scores changed significantly as a result of parents' attempts to distort their responses. Across the response-distortion conditions, the PSI Defensiveness Scale only detected protocols in the fake-good condition with detection rates below 50%, whereas the CAP Inventory validity indexes correctly detected as invalid 94.7% and 91.1% of the protocols generated by general-population parents and at-risk parents, respectively. With regard to correct detection and labeling rates, except for the labeling of faking-bad behavior in the at-risk group (57.9% correct), the labeling rates of the CAP validity indexes for each of the response-distortion conditions in the general population and at-risk groups were acceptable, ranging from 82.4% to 100% correct.