Cognitive Homogenization, Educational Achievement and Social Inequality A theoretical model and empirical analysis of the effects of strict ability tracking according to cognitive abilities on achievement and social inequalities in secondary education across the German federal states using data of the "national Educational Panel Study" (NEPS)

Abstract

Channeling students into different educational tracks based on their cognitive abilities (ability tracking) is assumed to foster educational achievement for all students by creating homogeneous learning contexts in school. Critics however question such beneficial effects and claim that ability tracking aggravates social inequalities. This view could be deemed the "standard position"of previous research. Starting from a general theoretical model, the present study examines the effects of ability tracking with different degrees of strictness on student achievement in secondary education. Using data of the National Educational Panel Study (NEPS), we compare the educational systems of the German federal states, which differ considerably with regard to the strictness of tracking regulations. Findings indicate that strict ability tracking improves educational achievement in secondary education, without aggravating influences of social origin. Students in cognitively homogeneous classrooms with lower average ability composition particularly benefit from strict ability tracking. The students in cognitively homogeneous classrooms perform worst when liberal tracking is implemented and best under strict tracking regulations. In classrooms with higher average ability levels, by contrast, cognitive homogeneity appears to be less relevant; students in cognitively heterogeneous classrooms here even tend to outperform those with a more homogeneous classroom composition. These results clearly contradict findings of previous studies that support the standard position. This discrepancy might reflect methodological shortcomings of previous studies. Particularly the bulk of contributions that are based on cross-national data, oftentimes lack appropriate means to test the theoretical assumptions and/or are unable to include all relevant factors in the analyses.