Objective: To examine the effects of individual and group factors on the changes in child-rearing practices in a parenting program. Method: At risk and non at-risk parents (N = 496), who varied by sociodemographic factors, participated in 94 groups that differed by size and composition. Latent growth curve modeling was used to examine the impact of individual and group variables on pre–post changes in inductive, coercive, and permissive-neglecting practices. Results: Inductive parenting increased and coercive and permissive-neglecting parenting decreased over the intervention and the proportion of change was influenced by family structure, parents’ gender and group composition. Low-risk groups and mixed (at-risk and non at-risk) groups did better than medium/high-risk groups in changing inductive and coercive practices whereas all groups did well in changing permissive-neglecting practices. Conclusion: To test the effectiveness of a group-based intervention requires taking into account both individual and group sources of variation in the implementation process.