The effectiveness of three counseling models and a no-counseling control situation for students with behavioral and learning difficulties was examined. The first model consisted of counseling parents, teachers and students; the second of consulting and counseling teachers and students; and the third of counseling students only. Subjects were 58 fourth- through sixth-grade students. Results indicated that students with behavioral and learning difficulties who received counseling improved significantly in reading performance as compared to students who had no counseling. The counseled groups also showed an improvement in classroom behavior and self-concept as measured by the Walker Problem Behavior Identification checklist and the Inferred Self-concept Scale. However, these gains were not statistically significant. The second counseling model was found to be statistically more effective than the first and third models in reading improvement, but there was no significant difference among the three treatment models in classroom behavior and self-concept scores. Appendices include detailed descriptions of the program, from student selection criteria through final evaluation forms, making the program easily replicable by other interested schools. (Author/BP)