AbstractFifty‐eight first‐ and second‐grade children exhibiting behavior problems were identified by teachers for inclusion in a behavior intervention program. Children were assigned to experimental and control groups matching for scores on the Teacher's Report Form (TRF), ability, and achievement. The experimental subjects received classroom‐based behavior intervention mediated by teachers. Pre‐ and postintervention TRFs were compared and showed significant correlations, indicating stability of the instrument. Children in both experimental and control groups improved significantly in their TRF scores. No significant differences were found between the two groups in the amount of improvement in their TRF scores, even though the experimental group evidenced significantly greater improvement in behavioral measures. Therefore, although the TRF appears to generate stable indices of behavior problems, it may not be useful for detecting behavioral changes resulting from classroom‐based in