On the dynamics of two-person interactions.

Abstract

A class of discrete-state probabilistic models is discussed that allow for the separate assessment of the effects of both self-regulatory processes, including predispositions, and interactive processes on the complementary, reciprocal, and reinforcing tendencies of observable responses. In most of these models, the two responses made at Time t + 1 by members of an interacting dyad are statistically independent, given the pair of responses made at Time t. For each model, estimable parameters are denned in terms of state transition probabilities and are labeled, for the most part, as measures of bias or of sensitivity to previous own and other behavior. The models are applied to data collected by other researchers on the self-fulfilling nature of social stereotypes, on aggression in a natural setting, and on parent-infant interactions, and the resulting parameter estimates are discussed. Possible applications of the models for as sessing the effects of small group interaction are also mentioned.