To content

New journal article in Journal of Educational Psychology published

-
in
  • News
  • Publikationen
© American Psychological Association
Becker, M., Kocaj, A., Jansen, M., Dumont, H., & Lüdtke, O. (2022). Class-average achievement and individual achievement development: Testing achievement composition and peer spillover effects using five German longitudinal studies. Journal of Educational Psychology, 114(1), 177–197. https://doi.org/10.1037/edu0000519

In recent studies, the existence and relevance of achievement composition effects on students’ individual achievement have been called into question because of the methodological challenges arising in multilevel analyses. The study examined how class-average achievement is related to students’ achievement development across one school year. It used data from Germany, which has a secondary school system with large achievement differences between schools and classrooms due to rigid, explicit between-school tracking practices. The researchers accounted for two methodological challenges, controlling for both selection bias and measurement error. Adopting an approach based on integrative data analysis (IDA), they systematically (re)analyzed five German longitudinal large-scale data sets. This IDA approach allowed them to quantify the extent to which results vary across (a) different longitudinal data sets and (b) different analytical strategies (i.e., ways of accounting for confounding variables and measurement reliability). Overall, the study found both general achievement composition effects and narrower peer spillover effects (i.e., effects of student composition above and beyond the effects of tracking) in the German setting, even after controlling for measurement error and selection bias. The results counter recent suggestions that composition effects on achievement development may be mere phantom effects due to methodological misspecifications. However, estimates of composition effects varied substantially based on the analytical approach. The study concludes with considerations regarding how to interpret composition effects in multilevel modeling and which effects are of interest for educational research.