Concordance of Reports of Childhood Abuse by Adults

Abstract

A random sample of 610 adults, enrolled in a population-based, prospective cohort study, were examined by questionnaire on two occasions (separated by an average of 20 months) about unwanted sexual experiences, physical abuse, and psychological abuse that might have occurred when they were younger. Although the screening questions for these types of maltreatment were not identical at the initial and followup surveys, relatively high levels of concordance were identified. Agreement on sexual abuse between both questionnaires was 85.6% (kappa =.64), 91.8% for physical abuse (kappa =.59), and 65.4% for psychological abuse (kappa =.25). Reports of noncontact sexual abuse accounted for approximately one fourth of the nonconcordant reports on sexual abuse. Logistic regression examining demographic variables (e.g., age, gender, marital status, education level) was used to identify predictors of concordance in subjects who indicated a specific abuse type on either questionnaire. Females had significantly greater concordance on sexual abuse, whereas marital status predicted concordance for psychological abuse. No variables were identified that predicted concordance on physical abuse.