Purpose: The paper aims to examine to what extent and why parental involvement as well as characteristics of ethnic school population influence social skills scores (social position, behavioural skills) of students. Design/methodology/approach: The study used the COOL5-18 database (2010) that included 553 Dutch primary schools and nearly 38,000 students in Grades two, five and eight (aged approximately 5, 8, 11, respectively). Multilevel regression analyses were used for analysis. Findings: The findings indicate that parental involvement has a positive impact on the social skills scores of the students; behavioural skill scores are higher in ethnically homogeneous schools and lower in schools with a high share of non-Western ethnic minority students. There is no impact of characteristics of school population composition on social position scores. Research limitations/implications: A possible disadvantage is the way in which social skills and parental involvement were measured. These measurements are possibly negatively influenced by the teacher's judgement of the language skills of the parents. Originality/value: Existing research focuses on the impact of parental involvement and the composition of ethnic school population on cognitive skills. This study shows that parental involvement has a positive impact on social skills. This study also shows that in schools with a homogeneous ethnic composition or a high share of native Dutch children, behavioural skills scores are higher, but social position scores are not impacted.