Although parenting programs are the most widely used approach to reduce children's disruptive behavior, there is a notable lack of understanding of the exact changes in parenting that underlie their effects. Challenges include the frequent use of composite measures of parenting behavior, and insufficient power to detect mediation effects and individual differences in these in individual trials. We pooled individual participant data from 14 European randomized controlled trials of social learning-based parenting programs to examine which specific parenting behaviors best explain program effects. Participants were 3,252 families with children aged 1 to 13 years. As putative mediators, we included parental use of praise, tangible rewards, physical discipline, harsh verbal discipline, and not following through on discipline. Additionally, we explored whether subgroups of families showing different mediational pathways exist. Changes in parenting partially mediated program effects, with all included parenting behaviors, except parental use of praise, serving as unique mediators. Less harsh verbal discipline and increased following through on discipline were the strongest mediators. We identified three subgroups with distinct responses to parenting programs. Most families benefited, partly through increased following through on discipline; those with the least or most difficulties were more likely to benefit less, or not at all. Our findings offer insight into the specific parenting behavior changes key to reducing disruptive child behavior, while highlighting the need for innovative research methodologies to gain a deeper understanding of individual differences in parenting program benefits and mechanisms.