Factor Structure and Predictive Validity of Questionnaire Reports on Family Problem Solving

Abstract

This study examined mother, father, and youth (mean age 14) questionnaire reports on family problem solving in 55 Asian American and Pacific Island American families. Confirmatory factor analysis was used to test whether the ratings could be best represented in terms of behaviors, perceptions, relationships, or a family problemsolving style. The predictive validity of the family problem-solving ratings was evaluated with tests for associations with youth delinquency and selfesteem. Perceptions linked to the respondent's family role best accounted for the pattern of ratings. Predictive validity depended on family role. Mothers' perceptions of family problem solving predicted youth delinquency, and the youths' perceptions predicted self-esteem. Key Words: family assessment, minority families, parent-child relations, problem solving, structural equation models. Research and theory in several disciplines have maintained that families have distinctive strategies or styles for solving problems (Haley, 1976; Klein & Hill, 1979; Reiss, 1981; Straus, 1968). Work in this area has grown, in part, because family problem solving has been linked to important outcomes for family members, including psychological adjustment, school achievement, delinquency, social competence, and psychopathology (Coughlin & Vuchinich, 1996; Forgatch, 1989; Grotevant & Cooper, 1986; Hauser et al., 1984; Lamborn, Dornbusch, & Steinberg, 1996). Such evidence has been the basis for prevention and intervention components, that seek to improve outcomes by influencing family problem solving (Forgatch & Patterson, 1989; Haley, 1976; Hibbs & Jensen, 1996; Kazdin, Siegel, & Bass, 1992; Robin & Foster, 1989; Spoth, Redmond, Haggerty, & Ward, 1995). As the research and applications have expanded, however, variations in and controversies about the measurement of family problem solving have emerged (Keiran, Maguire, & Hurlbut, 1996; Lamborn et al., 1996; Rueter & Conger, 1995a; Tallman, 1988; Vuchinich, Angelelli, & Gatherum, 1996). Such measurement controversies limit the extent to which research on this family concept can accumulate and provide guidance for better applications. The goal of this article is to clarify key measurement issues by briefly reviewing how family problem solving has been conceptualized and measured, and by examining the factor structure and predictive validity of questionnaire reports by family members on dyadic family problem solving. MEASURING FAMILY PROBLEM SOLVING Like other family-level constructs, such as cohesion, adaptability (Olson, 1986), family environment (Moos & Moos, 1976), or family decisionmaking (Lamborn et al., 1996), family problem solving usually refers to a characteristic of a family that is distinct from the characteristics of any of the individuals or any dyadic relationship within the family (Sabatelli & Bartle, 1995). Over the years, there has been growing consensus that this family-level characteristic refers to the ability of the family as a group to openly address difficulties within the family or problems imposed from outside the family, to consider alternative ways of adapting to the situation, and to settle on a reasonable approach to the issue at hand. The concept of families having a distinct problem-solving strategy or style is consistent with several different theoretical approaches to family life (Haley, 1976; Reiss, 1981; Robin & Foster,1989). For example, a family's problem-solving strategy may be based on behavioral learning (Forgatch & Patterson, 1989), a shared worldview (Reiss, 1981), or intrafamily structural alignments (Haley,1976). Research on most family-level constructs (e.g., cohesion, adaptability, family environment) has used responses to forced-choice questionnaires from one, and occasionally from more than one, family member to assess family characteristics (e.g., Sabatelli & Bartle, 1995). …