The Effect of Embedding Formative Assessment on Pupil Attainment
Anders J, Foliano F, Bursnall M, Dorsett R, Hudson N, Runge J, Speckesser S · 2022
grade Arctindependentunreplicatednumbers spot-checked
Sample
140 English secondary schools / 25,393 pupils
Population
Year 10-11 pupils in English secondary schools; GCSE Attainment 8 outcome.
Design
The peer-reviewed publication of the EEF Embedding Formative Assessment trial — the SAME trial as Speckesser et al. 2018, not an independent replication. Recorded separately because it reports the sensitivity and complier analyses the evaluation report did not.
Key findings
Peer-reviewed confirmation of the +0.09 at-scale estimate, with 0.11 under sensitivity and complier analyses. The prior-attainment gap narrows while the disadvantage gap does not — the intervention helps low prior attainers, which is not the same as helping poor children.
Genetic confound
Minimal — school-level randomization, national-register outcome.
Replication notes
Same trial as the EEF report; treat the two records as one piece of evidence, not two.
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Attainment 8 (GCSE), pre-registered primary analysis | ES | 0.09 | standardized | end of 2-year treatment | business-as-usual | end-of-treatment | domain-skill |
| Sensitivity analysis and complier-average causal effect | ES | ~0.11 (excluding schools already running a similar programme at baseline; CACE similar) | standardized | end of treatment | business-as-usual | end-of-treatment | domain-skill |
| Gap effects | pattern | may narrow the gap between high and low PRIOR-ATTAINMENT pupils, but not the disadvantage (FSM) gap | standardized | end of treatment | business-as-usual | end-of-treatment | domain-skill |
Cited by
- Feedback and formative assessmentmixedconf: highgc: low