Changing Mindsets: Evaluation Report and Executive Summary

Trial A (pupil workshops) 6 schools, 286 pupils, pupil-level randomisation; Trial B (teacher professional development) 30 schools, 1,505 pupils, school-level randomisation

England (Portsmouth, Southampton, Hampshire); Year 5 pupils aged 9-10.

The EEF pilot whose 'evidence of promise' launched the English scale-up. Trial A used an ACTIVE control - equal extra study-skills support without the mindset content - which is the correct comparator and rare in this literature. Trial B used business as usual and had baseline imbalance on free-school-meal status and prior attainment. EEF security was rated moderate-to-low for Trial A and moderate for Trial B.

Trial A gave +2 months in English and +2 months in maths, both non-significant. Trial B gave 0 months in maths and MINUS 2 months in English, non-significant. Both arms raised self-reported theories of intelligence, but the report states the security of those findings is low. The process evaluation records that intervention AND control schools were already using growth-mindset practices.

Low - randomised in both arms.

This is the pilot that failed. Scaled up as Foliano et al. 2019 with 101 schools and 5,018 pupils measured on national tests, the +2 months became exactly zero on maths, reading and grammar, including for disadvantaged pupils. The pair is the cleanest pilot-to-scale evaporation in the archive.