International comparative achievement studies reveal not only considerable achievement variations among students in a country but also between the participating schools. In Austria, for example, results from the international reading literacy study PIRLS show a difference of 128 points between the best and the lowest performing school. This huge difference of nearly two competence levels is remarkable – particularly regarding the fact that in Austria primary school is a comprehensive school, meaning that the allocation of students to schools is independent of students’ competencies but results from their living place. This analysis reported in this paper aims to evaluate those factors that have an influence on the differences in achievement at school level using PIRLS and TIMSS data. Furthermore it will be investigated whether this difference is also present in other (European) participating countries and whether the difference can be explained by the same factors. Using the Austrian PIRLS data, first analyses indicate that the average cultural capital, the amount of migrants and the average education level of the parents can explain more than half of the variance in achievement on school level. The comparison of these results with other (European) countries will provide valuable information about the ability of the schools (and schools systems) to compensate for existing social differences among students. It seems that at present Austria manages this compensation insufficiently.