Students judge feedback comments as fairer when they are usable

Updated Sep 07, 2026

Constructive information can affect how fair feedback feels. David Playfoot, Ruth Horry and Aimee E. Pink tested this using comments on fictional assignments.

What the experiments found

The authors report recruiting 127 and 151 second-year psychology students for two experiments at a UK university. Comments varied in usability and supportive wording. One experiment varied whether a grade was shown; the other varied its generosity.

Usable comments received higher fairness ratings in both experiments. Supportive wording had a smaller significant effect in the first and no significant effect in the second. The grade manipulations did not significantly affect perceived feedback fairness.

These findings concern reactions to fictional work. Participants did not judge the fairness of their own grades, and learning improvement was not measured. The discussion explicitly cautions that personal involvement could change the importance of tone or marks.

There is also a reporting inconsistency: Experiment 1’s stated sample size does not readily reconcile with the reported ANOVA degrees of freedom. We therefore describe the authors’ findings without reproducing those detailed statistics or implying independent reanalysis.

An editorial application

Review whether comments explain what could be improved and give a usable next step, while remaining respectful. Ask students how they interpreted the feedback and what support they need to act on it.

When investigating fairness concerns, distinguish comments, criteria and grades. This study should not be used to assume that every complaint about marking is really a request for clearer wording.

Reference

David Playfoot, Ruth Horry and Aimee E. Pink. “Fairly useful feedback: characteristics of feedback comments perceived as fair by students.” Assessment & Evaluation in Higher Education. DOI: 10.1080/02602938.2025.2586836.

Reviewed 7 September 2026 against the full paper. The revision narrows generalisations, removes unsupported predictive claims and notes the source’s sample/statistic inconsistency.