MULTIVARIATE ASSESSMENT OF CONFLICT IN DISTRESSED AND NONDISTRESSED MOTHER‐ADOLESCENT DYADS

TLDR

The study employed a comprehensive battery—including retrospective judgments, frequency estimates, self‑monitored home recordings, and tape‑recorded problem discussions—to assess conflict across 78 mother–adolescent dyads in distressed and nondistressed families. Multivariate analysis identified nine key variables that perfectly classified families in the training set and correctly classified 84% in cross‑validation, highlighting strong discriminators such as maternal and adolescent reports and independent interaction ratings.

Abstract

A battery of measures was used to assess conflict between mothers and young adolescents (females and males, 11 to 15 years of age). Two groups of families, one composed of a distressed clinical sample (N = 38), the other a nondistressed normative sample ( N = 40), participated. The assessment battery included retrospective judgments, frequency estimates, self‐monitored home recording, and tape‐recorded discussion of a home problem. Content of assessment measures tapped aspects of parental control, decisionmaking, self‐reported interaction behavior, arguments, interaction behavior rated by independent “blind” observers, frequency and anger‐intensity of specific problematic issues, and perceptions of positive and negative behaviors of the other family member. Based on univariate analyses, 21 of the 26 defined variables discriminated significantly in the predicted direction. Maternal and adolescent reports of behavior and independent ratings of tape‐recorded interaction emerged as strong and consistent discriminators. Stepwise multivariate discriminant analysis provided successful classification of 100% of the families based on the inclusion of nine variables. In a cross‐validation sample, 84% of the families were correctly classified. Implications for systematic outcome research as well as clinical application are discussed.