Two studies evaluated a technique for combining perceptions family members have of their families into measures of the family as a unit. Study 1 tested the reliability of an individually administered measure of family support and affiliation created to serve as a basis for aggregated family indices. Two existing family-assessment methods were adapted and administered to 71 subjects, mostly graduate and undergraduate students, who rated their own family relationships from a child or parent perspective. Across five subscales and four samples (parents, children, males, females), internal consistency ranged from .60 to .95, with an average of .88. In study 2, the instrument was administered along with the Family Environment Scale (FES) (Moos and Moos, 1986) to families in crisis, families in counseling, and community families (n=24 families). Indices based on agreement in family members' ratings (convergence) correlated positively with FES Cohesion and Expressiveness and negatively with FES Family Incongruence, supporting the notion of convergence as a measurable family trait. Discriminant analyses demonstrated a consistent superiority of agreement-based family indices over individual scores for predicting family criterion group membership. Appendixes contain the questionnaire and two clinical cases. (Contains 10 references and 5 tables.) (Author/SLD)