Agreement about violence in U.S. couples

Numerous studies have demonstrated that cohabiting and married partners often do not agree about the occurrence of intimate partner violence. This study assessed agreement about intimate partner violence in a large, representative sample of U.S. couples. Using a multistage probability-sampling design, separate face-to-face interviews were conducted with both members of 1,635 respondent couples in their homes in the 48 contiguous states. Using statistics appropriate for the evaluation of agreement, the results suggest that concordance about intimate partner violence in the U.S. General population is low. Classification of one partner's report as violent on the basis of the other partner's report of violence yielded unacceptable high misclassificationerrors. Use of the proxy method for investigations of intimate partner violence is not recommended.