In a recent paper on gender inequality in higher education Buchman and DiPrete (2006) assume that the decrease in the gender gap in college completion in the US can partly be explained by changes in the allocation of familial resources in favour of women. However, they do not test this hypothesis empirically. In this paper I examine the effects of sibling sex composition on the graduation of women in more detail by analysing data from the German Life History Study. I assume that resources are the key issue to explaining the effects of sibling configuration on educational attainment. Tertiary education is a good case for testing sex composition effects due to the unequal distribution of resources between and within families, because the direct costs and opportunity costs of higher education are relatively high compared to those of earlier educational decisions. Accordingly, I expect that working class daughters are most likely to be disadvantaged if they are raised with brothers. The empirical results show that in fact, not the presence of a brother as such hinders educational attainment of sisters, but older brothers have a negative influence on their sisters chances of graduation. In accordance to the hypothesis, this effect is stronger for university graduation than for graduation at Fachhochschule. For social class differences in sibling effects it turns out that working class daughters are particularly less likely to graduate compared to service class daughters if there are older brothers in the family.