Accelerating the development of formal thinking in middle and high school students

8 schools plus 1 laboratory school; ~24 classes split between intervention and control (per the EEF 2016 report's description of the same dataset)

England; average-attaining 11-12 and 12-13 year-olds in comprehensive secondary schools.

The founding CASE study. Up to 30 hour-long "Thinking Science" intervention lessons delivered by science teachers over two years, displacing normal science lessons. Classes were matched, not randomised. The outcome measure is the tell: Piagetian Reasoning Tasks (NFER 1979) — the same Piagetian construct the intervention was built from, so the headline 0.89 SD is a gain on the trained task, not transfer. Shayer (1999) defends this explicitly. Analysis excluded two schools that dropped out or failed to implement, which the independent EEF evaluators later flagged as a potential false-positive bias. Effects appear only in specific age-by-sex cells; the 12+ girls and the middle-school students showed no gain over controls.

Boys who started aged 12+ showed a pre-post effect of 0.89 SD on Piagetian tests versus control classes — a shift from the 51st to the 74th British percentile. Neither the middle-school students nor the 12+ girls beat controls; gains appeared for girls only in one 11+ class and in two laboratory-school classes. Critically, the paper reports NO effect on science achievement tests during the intervention. So at end of treatment CASE moved only the Piagetian measure it was designed around; the science, maths and English claims all came later, from delayed GCSE analyses (Shayer & Adey 1993).

Medium-to-high: matched, non-randomised classes in self-selecting schools, with two non-implementing schools dropped from analysis. Intake differences between arms cannot be excluded.

The immediate cognitive-development finding has been reproduced repeatedly (Iqbal & Shayer 2000; Choi et al. 2002; McCormack et al. 2014) but always on Piagetian-based tasks, which the EEF evaluators note gives CASE participants an in-built advantage. The independent randomised effectiveness trial of the scaled programme (EEF/Hanley et al. 2016) found no effect on science and negative point estimates on standardised English and maths.