Students use the EAT framework to review assessments and support peers

Updated Sep 07, 2026

Updated, 7 September 2026: We checked the complete paper and added its sampling and evaluation limits. The study reports assessment changes and higher self-rated confidence after peer sessions; it does not establish a controlled effect on attainment.

What the study involved

Sheila Amici-Dargan, Kaisa Ilmari, Annabelle Buckley, Mathusha Chandrakumaran, Carol Evans and Stephen Rutherford describe activities at Cardiff and Bristol using the Equity, Agency and Transparency framework, or EAT. It organises assessment review around literacy, feedback and design.

At Cardiff, students and staff reviewed six first-year bioscience assessments. Seventeen undergraduate interviews explored assessment experiences. Staff and student ratings differed in both directions; the paper reports no single consistent pattern across all assessments. The review exercise involved a small, engaged subset.

At Bristol, 87 first-year students attended optional sessions led by final-year students, using exemplars for a graphical-abstract assessment. There were 69 survey responses before and 45 after. Responses were anonymous and unmatched; the authors used unpaired tests and reported higher confidence across six questions.

Those results concern self-rated understanding and confidence. Unequal, unmatched responses without a control group do not establish individual gains, improved marks or a causal intervention effect. The authors also describe subsequent assessment changes, while wider effects of the framework remained questions for future work.

Questions for assessment review

Our suggested starting point is to choose a specific task and compare how students and staff understand its purpose, criteria and feedback. A difference in ratings can open a discussion without showing which group is correct or why the difference occurred.

If students help design a support activity, define their influence and the support they need. Assess the resulting experience with measures suited to the intended outcome. Confidence, understanding of instructions and demonstrated performance should be reported separately.

Record who participated and who answered each follow-up. The student feedback analysis glossary can help establish shared terms before comparing comments. These are our local evaluation suggestions. Neither this case nor a consistent set of comment themes establishes that a particular analysis product improves assessment or trust.

Reference

Amici-Dargan, S., Ilmari, K., Buckley, A., Chandrakumaran, M., Evans, C. and Rutherford, S. (2026). Empowering students as partners in enhancing assessment practice, using the research-informed EAT framework. Student Engagement in Higher Education Journal, 7(3). Open paper.