Between-School Ability Tracking and Ethnic Segregation in Secondary Schooling


Social Forces Vol. 98(1), pages 119-146, doi: 10.1093/sf/soy099


Between-school ability tracking—the assignment of students to different school types based on their prior achievement—is usually associated with increased ethnic segregation across schools. This article argues that stronger between-school ability tracking is not only associated with stronger ethnic sorting into school types, thus increasing segregation. At the same time, it hampers majority flight, majority members’ avoidance of exposure to minority members, thus decreasing segregation. To identify the two-fold effect of tracking the article exploits a unique feature of the German secondary school system: regional variation in the strength of between-school ability tracking. Analyses rely on administrative data entailing geocoded information on all secondary schools in Germany in 2008/2009. Results corroborate expectations of a two-fold effect: there is indication that between-school ability tracking increases segregation via more sorting into tracks while at the same time decreasing it via less school sorting within each track and via less spatial sorting. This suggests that school reforms changing tracked school systems into more comprehensive school systems may have a weaker desegregating impact than expected.

Name-based Measures of Neighborhood Composition: How Telling are Neighbors’ Names?

2017, with Jörg Dollmann

Survey Research Methods Vol. 11(4), pages 435-450, doi: 10.18148/srm/2017.v11i4.7214

krusedollmann2017Name-based ethnicity classification is a common tool in the sampling of minority populations. In recent years, however, it has become a popular technique to construct measures of neighborhood composition if more objective data are unavailable. In this article, we test the accuracy of such name-based measures of neighborhood composition, relying on the example of German neighborhoods. Drawing upon previous research, we assert that ethnic groups differ as to how well they are identifiable via name-based classification. Moreover, the ethnic mix in neighborhoods varies systematically, the ethnicities of immigrants residing in majority-dominated neighborhoods differing from those residing in minority-dominated neighborhoods. Taken together, these two notions imply that a name-based classification bias should be neighborhood-specific. Results indicate a tendency to overestimate majority shares in minority-dominated neighborhoods and slightly underestimate them in majority-dominated neighborhoods. All analyses rely on data from the “Children of Immigrants Longitudinal Survey in Four European Countries (CILS4EU)” as well as on neighborhood compositional data from local statistics of two German cities. The article closes with a discussion of potential strategies to cope with the name-based classification bias.

The SES-Specific Neighbourhood Effect on Interethnic Friendship Formation. The Case of Adolescent Immigrants in Germany


European Sociological Review, Vol. 33(2), pages 182-194, doi: 10.1093/esr/jcw056

Neighbourhoods with high ethnic concentrations are argued to steer immigrants into lives separate from the native population. Taking adolescent immigrants in Germany as an example, this article argues and shows why the link between the ethnic composition of immigrants’ neighbourhoods and their friendships depends on immigrants’ socio-economic status (i.e. SES-specific neighbourhood effects on interethnic friendships); why high-SES immigrants living around more natives have more native friends, whereas for low-SES immigrants, this association is weaker.


Introducing a formal account of interethnic friendship formation, the article proposes four potential explanations, three being empirically corroborated. First, SES differences are partly an artefact because of model misspecification. Second, low-SES immigrants attend more concentrated meeting contexts (i.e. schools) than high-SES immigrants, yielding different opportunities for native friends even when neighbourhood compositions are identical. Third, SES-specific friendship preferences may also be responsible. There is no indication that SES groups differ in how much they rely on their neighbourhoods when making friends. Additional analyses conducted on adolescent immigrants in the Netherlands confirm these patterns. All presented analyses make use of ego-network data from the first wave of the CILS4EU data, accompanied by small-scale neighbourhood information from an external data source.

From neighbors to school friends? How adolescents’ place of residence relates to same-ethnic school friendships

2016, with Sanne Smith, Frank van Tubergen, and Ineke Maas

Social Networks, Vol. 44, pages 130-142, doi: 10.1016/j.socnet.2015.07.004

kruseetal2016This study examines to what extent adolescents’ place of residence is related to the opportunities and the preferences to befriend same-ethnic classmates. Analyzing 3345 students within 158 German and Dutch school classes, we find that sharing a neighborhood provides additional meeting opportunities to become friends in class as adolescents are likely to befriend classmates who live nearby them or who live nearby a friend of them (propinquity mechanism). However, this hardly explains why adolescent friendship networks in school classes tend to be ethnically homogeneous. Also, we find no convincing evidence that an adolescent’s preference for same-ethnic friends in class varies with the share of outgroup members in his/her neighborhood (exposure mechanism).