The effectiveness of extended day programs: Evidence from a randomized field experiment in the Netherlands

Seven Dutch elementary schools

Dutch elementary school pupils

Economics of Education Review 36, 1-11. READ AT ABSTRACT DEPTH ONLY - the article is paywalled and no numeric effect sizes were obtained. A randomised field experiment adding FIVE WEEKLY INSTRUCTION HOURS (two of language, two of mathematics, one of excursions) for three months. It is included despite the thin read because a randomised null on extended-day programmes is exactly the kind of result the archive exists to record, and because the authors note that 'hardly any evidence that these programs are effective' exists generally.

'This three-month program had no significant effect on math or language achievement.' No effect sizes were obtained. Read alongside the Danish trial, which added three weekly hours for sixteen weeks and found 0.15 SD in reading, the pair bracket the honest uncertainty about short extended-day interventions: two randomised trials of similar dose and duration, one positive and one null.

Low. Randomised assignment balances family and child characteristics in expectation.

The null replicates across the large-dose extended-time literature - Massachusetts, Chile and Florida's extra hour all deliver between zero and 0.07 SD, with Florida's decaying to zero within three years - and is contradicted by the Danish trial. The most parsimonious reading is that adding hours produces small and unreliable effects that depend heavily on what fills them, which is what Rivkin and Schiman's classroom-quality interaction predicts.