Parental-child interactions are highly relevant for child development. However, there exists little evidence for a causal relationship between the two, and researchers do not usually separate different forms of interactions when studying the effectiveness. This paper is an attempt to establish causality while focusing on a specific form. It uses survey panel data from China to study how parental-reported attitudes toward school performances affect educational outcomes. An essential feature of this paper is that the reduced form analysis controls for child-perceived parental attitudes, previous educational outcomes, and schools’ characteristics, all of which are confounders that most previous researchers have left out in their studies. The result shows that a unit increase in the constructed parental-reported attitude variable, i.e. parents are more concerned about school performances, leads to a 0.0531-point increase and a 0.0111-point increase in children’s total subject score and cognitive test score. Additionally, the paper uses instruments to address concerns about measurement errors and omitted variable bias. The result is again statistically significant and confirms the conventional wisdom that “tiger parents” produce children with higher academic achievement. Interestingly, while children’s perceptions are significant and sometimes have a stronger correlation with educational outcomes than parental-reported attitudes do, perceptions do not matter once endogeneity is accounted for.