Methodological issues in investigations of the relative effects of schooling and age on school performance: the between-grade regression discontinuity design applied to Swedish TIMSS 1995 data

Abstract

The primary purpose of the study was to examine methodological issues in investigations of the relative effects of schooling and age on school performance. One means of separating these 2 effects involves the use of a between-grade regression discontinuity approach. However, amongst other things, this approach is based on assumptions concerning the mechanism for allocating students to grades. Using the Swedish sample of the Third International Mathematics and Science Study (TIMSS) 1995, which comprised 3 grades (Grades 6, 7, and 8), the study investigates the robustness of the regression discontinuity approach in terms of deviations from underlying assumptions. The main finding is that the technique is robust vis-à-vis deviations from a strictly age-based grade allocation present in the data. The pattern of substantive results also corresponds with those obtained in previous research, with the grade effect being about twice as strong as the age effect.