A Brief Form of the Child Abuse Potential Inventory: Development and Validation

Abstract

A brief version of the Child Abuse Potential Inventory (CAP) was developed using a development sample of N = 1,470, and cross-validated using an additional sample of N = 713. Items were selected to maximize (a) CAP variance accounted for; (b) prediction of future child protective services reports; (c) item invariance across gender, age, and ethnicity; (d) factor stability; and (e) readability and acceptability. On cross-validation, scores from the resulting 24-item risk scale demonstrated an internal consistency estimate of .89, a stable 7-factor structure, and substantial correlations with the CAP Abuse Risk score (r = .96). The CAP risk cutoff was predicted with 93% sensitivity and 93% specificity (area under the receiver operating characteristics curve = .98), and the Brief Child Abuse Potential Inventory (BCAP) and CAP demonstrated similar patterns of external correlates. The BCAP may be useful as a time-efficient screener for abuse risk.