In this study we investigated longitudinal associations among parenting, children's temperamental negative affectivity, and internalizing and externalizing behavior. Second, we tested whether findings confirmed the diathesis‐stress model or differential susceptibility theory when conducting stringent interaction tests. The sample included 129 children and their families. Parenting quality (age 5) was measured by parent–child interaction observations. Parents evaluated child negative affectivity (age 7) and teachers reported on problem behavior (age 12). Multiple regression analyses revealed an interaction effect of negative affectivity and parenting on externalizing behavior. Visual inspection suggested ‘for better and for worse’ effects of parenting for children with negative affectivity. However, more stringent tests failed to show convincing evidence for differential susceptibility theory. For internalizing behavior, negative affectivity may render children vulnerable regardless of parenting. Our results point at the importance of further testing interaction effects to distinguish between differential susceptibility theory and the diathesis‐stress model.