The purposes of this study were to learn about: (1) ways of identifying children not accepted by their peers, (2) the behavioral correlates of peer acceptance, and (3) strategies of teaching social skills to isolated children. Sociometric assessment, behavioral observations, and teacher and peer interviews were obtained in eight 3rd-, 4th-, and 5th-grade classrooms. Within each classroom, three isolated children were selected. One child was instructed in social skills, using a standardized coaching procedure; another child received a more individualized version of coaching based upon the child's style of relating to peers; and the third child received no coaching. Results from the assessment phase of the experiment indicated that scores from a 1-5 rating scale, by means of which children rated each of their classmates, correlated highly with both a positive peer nomination measure and a negative peer nomination measure. This suggests that the rating scale method may be a way of identifying rejected, as well as accepted, children without focusing on negative evaluation of peers. Data on behavioral correlates of peer acceptance were contrary to expectations. Behavioral observations in the classroom revealed no significant differences between children receiving high versus low sociometric ratings. Results from the intervention phase of the experiment suggest that individualized coaching is more effective than the standardized coaching procedure. (Author/BF)