Psychometric Characteristics of Translated Versions of the Child Abuse Potential Inventory

Objective: The present study was designed to assess the psychometric adequacy of translated versions of the Child Abuse Potential (CAP) Inventory. Method: Numerous terms were used to search databases for studies on translated versions of the CAP Inventory. Results: Across translations, meta-analyses produced Abuse scale internal consistency estimates of .88 and .91 for general population and comparison/maltreating parents, respectively. Abuse scale correct classification rates based on discriminant analyses ranged from 83.0% to 100% for maltreating parents and from 86.5% to 100% for comparison parents. When English (U.S.) item weights and cut scores were used, overall correct classification rates were about 10% lower. Construct validity data for translated versions of the Abuse scale were comparable to those for the English (U.S.) version. Even so, in many cases the CAP inventory Lie scale using English (U.S.) cut scores produced excessive false-positive classifications. Conclusions: Across translations, the Abuse scale had reliability and classification rates similar to those reported for the English (U.S.) version. Construct validity data suggested that similar risk factors are measured by both translated and English (U.S.) versions of the Abuse scale. However, U.S. norms for the CAP Inventory Lie scale may not be appropriate for use with translated versions of the Lie scale.