The goal of the study was to provide empirical findings on the effects of different types of parents' school involvement (PSI) on their children's mathematics performance. For PSI, we explored parent participation in parent-teacher conferences, voluntary parental involvement in school activities, parent informal contact with school or teachers, and parental phone contact with school or teachers. In addition, we paid focused attention to families of linguistic and racial minority groups who may have different dynamics of PSI for children's schooling and academic achievement. We conducted comprehensive descriptive statistics and multilevel longitudinal modelling using a nationally representative database, the Early Childhood Longitudinal Study (ECLS-K), to establish study conclusions with high validity and generalizability. The results showed that four different types of PSI had differential effects for different linguistic and racial groups. The study findings have practical suggestions and implications for educational researchers and practitioners.