International Large-Scale Assessments (ILSAs) are widely used to compare educational performance and equity across countries, yet such comparisons conflate the effects of schooling with non-school factors like family background. This study demonstrates how modest design changes could enable causal inference on schooling effects. Applying the Differential Exposure Approach (DEA) to TIMSS 2011–2023, we exploit within-country variation in test timing to estimate the causal contribution of schooling to achievement and inequality. Results reveal a global “schooling effect”: each additional month of exposure increases mathematics achievement and narrows socioeconomic gaps. However, limited variation in test windows reduces precision in country-specific estimates, suggesting that broader, randomized testing schedules would strengthen future ILSAs’ capacity to identify and compare schooling effects across countries.
Jensen et al. (Mon,) studied this question.