Hartmut Esser, Ilona Relikowski
Is Ability Tracking (Really) Responsible for Educational Inequalities in Achievement? A Comparison between the Country States Bavaria and Hesse in Germany

IZA-Discussion Paper
31 S.
Forschungsinstitut zur Zukunft der Arbeit

It is still taken for granted that (early) ability tracking increases the impact of social origin on achievement in (lower) secondary education, but without gains in the overall level. This contribution addresses the question of whether this common conviction is really correct. The various deviations and inconsistencies obtained from analyses that use other approaches and data bases form the starting point. On the basis of a general theoretical model, the Model of Ability Tracking, we specify the preconditions for identifying the effects of ability tracking. These include considering the school level as well as cognitive abilities prior to ability tracking at the end of elementary school. Both conditions aren’t included in common analyses using PISA data. As a consequence, effects of social origin have been systematically overestimated and those of cognitive abilities haven’t been detected in the respective studies at all. Because PISA data are lacking information on cognitive abilities in the institutional sorting at the end of primary school and no other appropriate data set to compare educational systems is available, these assumptions will be tested with another data base: the BIKS-study. This study allows using the different levels of strictness of the institutional rules concerning ability tracking in the two country states Bavaria and Hesse in Germany. The results support the presumptions of the Model of Ability Tracking: If school effects on the one hand and cognitive abilities on the other hand were taken into account, all effects of a reinforcement of social origin disappear and increases in school effects of abilities on achievement are observed in Bavaria, the country state with an especially strict rule for ability tracking. Applying the misspecifications of the other approaches to these data, one again obtains their misleading findings, and they disappear by approaching the analyses to the specifications of the Model of Ability Tracking.