A replication of earlier research, this study investigates the relationship between moral reasoning and classroom conduct. According to Kohlberg (1978), children who reason at preconventional levels (Stages 1 and 2) would be likely to disrupt a classroom, while children who reason at the conventional levels (Stages 3 and 4) should cause less classroom disruption. Bear and Richards (1981) found that conduct problems declined monotonically (almost linearly) with advances in moral reasoning for a sample of middle class Iowa fourth graders. These findings were consistent with theoretical expectations. In this replication of Bear and Richard's study, 87 fourth, sixth, and eighth graders from lower class homes in rural western Virginia were assessed according to Kohlberg's interviewing procedure. Classroom behavior was assessed by teacher reports on the Conduct Scale of the Behavior Problem Checklist. In most respects, results accorded well with the views of Kohlberg and replicate those of Bear and Richards. Level of moral reasoning increased with age, and conduct ratings were strongly influenced by sex and social class. However, the predicted monotonic trend between moral reasoning and conduct emerged only for boys; low-reasoning girls exhibited few behavior problems. (Author/CB)