Does Educational Tracking Affect Performance and Inequality? Differences-in-Differences Evidence Across Countries

Cross-country differences-in-differences across six international student assessments

School systems differing in the age at which children are tracked into separate school types.

Differences-in-differences: primary-school performance (before tracking anywhere) is used as the counterfactual for secondary performance in early- versus late-tracking countries. Graded C rather than B: identification rests on country-level comparisons and on the assumption that primary results are a valid pre-treatment baseline — weaker than the within-country reform and state-comparison designs in this cluster.

The first systematic causal-flavoured evidence that early tracking widens the achievement distribution without raising the mean. Directionally corroborated by better-identified later work, but its own design does not support a precise magnitude.

Cross-country design; country-level confounders (not student selection) are the main threat. Between-country differences in population composition are exactly what the primary-school baseline is meant to difference out.

The original cross-country result; superseded in credibility by within-country designs (Matthewes 2021, Meghir & Palme 2005, Pekkarinen et al. 2009, Canaan 2020) that reach the same conclusion.