To ascertain whether impulsive responding in behavior disordered adolescents is amenable to change, 15-year-old normal and "acting-out" behavior disordered adolescents participated in an experiment designed to alter Kagan's Matching Familiar Figures Test (MFF) scores through a modeling psychoeducational procedure. No significant differences occurred on the MFF latency or MFF error measures as a function of the modeling psychoeducational treatment in either the normal or behavior disordered adolescents. While the present study failed to confirm the efficacy of a modeling psychoeducational treatment procedure in altering either MFF error or MFF latency measures among any of the adolescent groups, this finding suggested that specific instructional strategies are necessary, in addition to a modeling treatment in altering cognitive style in adolescents. (Author)