Recent family interaction studies are reviewed with an emphasis on looking for dimensions along which disturbed and normal families differ. Several areas of consistency in the literature were found, including: family coalition patterns, patterns of conflict, flexibility versus rigidity, family effectiveness and efficiency, and deviant styles of communication. It was concluded that several measures reliably discriminate disturbed from normal families and that one type of measure in particular is a reliable predictor of thought disorder in offspring. Implications for clinical practice and future research are discussed.