This paper describes an ongoing project on the nature and effects of family structure and family interaction. The research presented involves the study of normal families, emphasizing the effects of system-level variables on the development of children within the family. A theoretical model is provided which focuses on the individuation process, i.e., the development of the individual within the family setting, and the way in which comfort with difference, accuracy of interpersonal perception, mutual validation and active discussion of opinions among family members contribute to and are affected by the personal development and functioning of individual family members. These system-level variables are examined as they are revealed in patterns of conversational interaction. Data collection from families with normal adolescent daughters, families of daughters with anorexia nervosa, and families with a history of child abuse is described and the development of a coding scheme to facilitate the microanalysis of the process of interaction is discussed. Analyses of differences among identifiable subsamples of the normal population are reported, as are comparisons involving the anorexic group and a subsample of the abuse group. The ability of the model to distinguish between strong and weak families within a normal population, as well as between normal and clinical populations, is also examined. (Author/NRB)