Effects of an early start in learning English as a foreign language on reading and listening comprehension in Year 9

more than 30,000 students

Representative large-scale assessment sample of Year 9 students (age ~15) in Germany

Multilevel regression on a representative national assessment. Early starters are students who began English before Year 3 or in Year 3; late starters began in Year 4 or 5. Observational: starting age is determined by federal state policy and school, not randomised, and the states that moved earlier differ from those that did not. The outcome is the national assessment's English reading and listening comprehension test - a standardized instrument, which is a real advantage over most of this literature.

Long-term advantages in English reading and listening comprehension for students who began English instruction earlier in primary school, still visible in Year 9. The paper is the strongest published counterweight to the Jaekel et al. (2017) reversal finding, on a much larger and nationally representative sample and at a later horizon.

Medium-to-high. Starting age varies by federal state and school, which correlates with regional affluence, migration composition and school-track allocation. No within-family or discontinuity identification.

Points the opposite way from Jaekel et al. (2017) on overlapping German populations. Neither is a randomised or discontinuity design, and the disagreement between two large German studies is the single best argument that the school-starting-age effect on foreign-language attainment is small enough to flip sign with specification and horizon.