Two processes, based upon derivations from reinforcement theory, were applied to the analysis of social interactions in families. Within this framework, it was hypothesized that reciprocity would hold in that for each dyad, the individuals would provide both positive and aversive consequences on a reciprocal basis. Observation data provided support for the hypothesis. The data also showed that both the deviant parents and the “deviant children” tended to be characterized by non-reciprocal interactions. In thse interactions , one member of the dyad coerced the other. The member who coerced tended to receive positive reinforcement while the behavior of the compliant-reinforcer was supported by the withdrawal of an aversive stimulus. It was proposed that the analysis could be extended to provide a description of a set of dyads, sujch as a family. It was assumed that effective behavior modification procedures would alter not only the output of deviant behaviors within a family system, but also increase the frequency of reciprocal interactions. Presumably, if this occurred it would increase the probability of long-term improvements. Some illustrative materials were presented which demonstrated some of the technology which might accomplish these effects. Data were also presented illustrating the chagnes produced.