Abstract
Grouping students together according to their ability – in vocational versus academic tracks, in different grades or schools – is frequently denounced as being ineffective and/or a source of additional inequality. Yet very few international studies have evaluated the effects of these practices on educational performance. This study attempts to fill this gap using standardized scores in maths, science and reading literacy at the age of 15 published by the OECD in 2000. Results are that ability grouping has no impact on effectiveness (country mean scores). And the intensity of inequality (within country dispersion of scores) is also hard to predict. It is only for maths that higher inter-school segregation, but not tracking or grade repetition, leads to higher inequality.
Notes
1 Ancova models are regression models combining categorical variables (generally analyzed with Anova models) and continuous ones (i.e. covariates).
2 i.e. is attending (modal) grade 10.
3 Programme for International Student Assessment.
4 The reader interested by the methodologies underlying each of these indices should refer to OECD (Citation2002).
5 In both cases, standardization means imposing that Mean = 0 and Standard deviation = 1.
6 Modal grade is used as reference (grade 10 in most countries).