Data from comparison studies were reviewed, which showed social learning procedures to be more effective than no treatment, client centered or traditional treatment. The limited findings relating to changes from baseline to termination in behavior of other members of these families were also presented. These findings were related to the hypothesis that treatment alters the system as well as the problem child. In addition, it was assumed that reducing the rates of coercive child behavior would strengthen the parenting behaviors which produce this effect. This mechanism, plus the alterations in the family system, were thought to correlate with the persistance of treatment effects. Data were presented showing that the family interrvention effects persisted. No data are available correlating changes in the sysem with persistence of treatment effects.