Background: The non-cognitive skills self-control and grit have often been cited as predictors of school performance, but little research has investigated whether this relationship is causal. We investigated the causal nature of this association in a classical twin design, with mono- and dizygotic twin pairs. Specifically, we evaluated the direct impact of self-control and grit on school performance, while controlling for genetic or environmental influences common to all three traits (i.e., confounding). Methods: Teachers of 4,891 Dutch 12-year-old twin pairs (of which 3,837 were complete pairs) completed a survey about school performance (school grades), self-control (ASEBA self-control scale), and the perseverance aspect of grit. We regressed school performance on self-control and grit within the twin model to establish the phenotypic, putatively causal, regression relationship. We modelled genetic or environmental confounding (influences common to the three phenotypes) to determine their influence. In all analyses, we corrected for sex, rater effects of the teacher, and parental socioeconomic status. Results: Self-control and grit explained 28.4% of the school-performance variance in the phenotypic regression analysis (assuming no confounding). However, allowing for genetic confounding (due to genetic pleiotropy) revealed that the association was largely attributable to genetic influences that the three traits share. In the presence of genetic pleiotropy, the phenotypic regression of school performance on self-control and grit accounted for only 4.4%. Conclusions: The association between self-control and grit as predictors of school performance is attributable to both genetic pleiotropy, and to a lesser extent, direct effects of self-control and grit on school performance, which are putatively causal.