We attempted to replicate an intervention program by Bornstein and Quevillon (1976), which had shown that the disruptive classroom behavior of Head Start children could be dramatically reduced through self‐instructional training. Although the subject population and procedures were quite similar across studies, our self‐instructional training did not produce socially significant, durable increases in either appropriate classroom behavior or changes in teacher ratings of the children's behavior. These results suggest that additional variables may have been responsible for Bornstein and Quevillon's succ