Ethnic segregation and educational outcomes in Swedish comprehensive schools: A multilevel analysis

Abstract

We ask whether ethnic segregation in Swedish comprehensive schools, fueled by increased residential segregation and increased immigration during the 1990s, are associated with depressed educational outcomes as measured either by teacher-assigned school grades in ninth grade (age 16) or eligibility to enter an academic study programme at the upper secondary level. We compare immigrant children and Swedish-born pupils with foreign-born parents with those of Swedish ancestry (the reference group). The data are based on two entire cohorts who graduated from comprehensive school in 1998 and 1999 (188,000 pupils and 1,043 schools), linking educational information from schools with Census data on parents’ education, socioeconomic position, and household composition. Using multilevel analysis we find grade differences to the advantage of the reference group. For second generation immigrants, but not immigrant children, these are accounted for by socioeconomic characteristics of the family of origin and of the composite socioeconomic status in schools. We argue that differences in school quality are unlikely to produce these results. Rather, a clustering of immigrant children is likely to create a more difficult learning environment, partly because of language problems, and relative lack of positive role models. This has negative consequences for the achievements of all children in these schools – though most for immigrant children – despite the fact that more resources are allocated to schools with many immigrant pupils. Desegregating policies as well as intense language training for newly arrived immigrant children could be efficient means towards overall increasing academic standards as well as decreasing educational inequality.