This study examined the effects of two parent training procedures in programming for generalization of parent and child behaviors to extra training settings. Five families were sequentially exposed to a child management training condition and planned activities training within a multiple baseline, across subjects design. Observational data were collected in two different settings, a training setting, and a range of generalization settings in the home and community. Child management training resulted in generalized changes in parent behavior in both training and nontraining settings. However, only one of five parents was effective in reducing levels of deviant behavior. Planned activities resulted in further improvements in child behavior in both the training and generalization settings for three families and maintained reduced levels of deviant behavior for a fourth family. Both trainning phases were ineffective in modifying problem behavior for the fifth family.