We use Danish register data to investigate whether the effects of schoolmates' gender and average parental education on individual educational achievement, employment and earnings vary with individual family characteristics such as the gender of siblings and own parental education. We find that boys with sisters have worse employment prospects than boys with no sisters when exposed to a higher share of girls at school. The opposite is true for girls who have sisters. We also show that the benefits from exposure to "privileged" peers accrue mainly to "disadvantaged" students. These benefits decline when the dispersion of parental education increases. Overall, the size of the estimated effects is small.