In this article, we propose a new approach for assessing the extent to which schools are successful in mixing students from different backgrounds. It is based on a comparison of variation in the composition of the student population in small-scale residential neighborhoods with variation in the composition of the student population at local schools. From this we compute a measure that corresponds to the number of small scale neighborhoods that needs to be sampled in order to arrive at the observed mixing of students in schools. Using this measure, we can show that in 2012, in a large majority of Swedish municipalities, schools are successful in mixing students from different types of neighborhoods, but in 25% of the municipalities mixing is not so good. Three fundamental determinants of mixing are large-scale residential segregation, average school size, and number of students in the municipality. These factors are strong determinants of mixing and when they are included, other contextual factors provide very little additional explanation of why mixing varies among municipalities. With the fundamental determinants excluded the contextual factors have an effect. For example, tertiary education, many migrants, and high proportions of independent schools tend to lower the level of mixing.