The psychoanalytic theory of identification and the cognitive-developmental and social-learning theories of imitation are briefly described. Pertinent empirical research in the following areas is summarized and critically evaluated: imitation in infants, observational learning, clinical use of modeling, and the relation of imitation to aggression, prosocial behavior, and self control. Among the author's general conclusions: (1) there is very little pertinent research on identification, but it appears that identification is not an all-encompassing unitary process; (2) the child's general cognitive level has an impact on what he or she can learn by observing models; (3) children are active cognitive processors of the model's words and actions; (4) Piaget's work with infants is the closest approach to date of a stage analysis of imitation; (5) young children may acquire, from brief periods of observation, certain motoric and verbal behaviors that appear to be associated with aggression in real-life situations; (6) the results of research on the long-term effects of television violence are equivocal; (7) experimental studies of observational learning of sharing behavior indicate that models do increase prosocial behavior in children; however, (8) prosocial behavior does not seem to be enhanced by watching television programs with prosocial content; (9) exposure to models who yield to the temptation to perform a prohibited act, and are not punished, has a disinhibiting effect on children; the effects of exposure to models who resist temptation, however, is less clear due to demand characteristics of laboratory research. (Author/SS)