A study of 20,879 learners found that peer-feedback benefits come from providing feedback: the length and substance of what a reviewer writes is what predicts their own performance gains.
Yu, Q., & Schunn, C. D. (2023). "Understanding the What and When of Peer Feedback Benefits for Performance and Transfer." Computers in Human Behavior, 147, 107857.
By 2023, meta-analyses had already established that peer feedback helps people learn. Effect sizes varied widely, though, from one study, course, or assignment to the next: some programs produced large gains, others almost none. Yu and Schunn set out to explain that variation by isolating which part of the peer-review process actually drives the benefit.
They separated the peer-feedback experience into its component parts — giving feedback, receiving feedback, and revising work based on it — and tested which of these, and which qualities of the feedback itself, predicted whether a person's performance improved and whether that improvement transferred to new tasks.
This was a large-scale, cross-context analysis: real assignment data from 20,879 students across 505 assignments in 243 courses at 76 different institutions, captured through structured online peer review. That scale is what made it possible to isolate individual variables at this level of precision — separating the effect of "how much someone wrote" from the effect of "how helpful it was rated" requires a dataset this size.
learners across 76 institutions — the same dataset behind Peerceptiv's "20+ years, 76 institutions" research claim.
The researchers used meta-regression, multi-level modeling, and ANCOVA to test each feedback experience against performance and transfer outcomes, while statistically accounting for differences between courses and assignments.
Two findings define this study:
1. Learning comes from giving feedback. The act of evaluating a peer's work — diagnosing what's wrong, applying criteria, explaining reasoning — is what predicted a reviewer's own performance gains.
2. Substance drives the gain. How much a reviewer wrote predicted their own learning gain. The cognitive effort of articulating detailed feedback is what builds the skill.
Most corporate feedback programs — 360 reviews, manager coaching, mentorship pairings — are built around the person receiving feedback as the one who's supposed to grow. However, the reviewer is doing the learning. Programs designed only around the person being reviewed are training the wrong side of the interaction.
Most organizations funnel feedback through a single channel: the manager evaluates the employee. That design leaves real learning on the table. Evaluating someone else's work is what builds the skill, so having only managers give feedback concentrates that benefit in one role and withholds it from everyone else.
It also doesn't scale. One manager reviewing every direct report's work, in detail, on a regular cadence, runs into a hard ceiling: time. Peer review removes that ceiling. Every employee who reviews a colleague's work gets the same cognitive workout a manager gets from reviewing theirs, and the review capacity grows with headcount instead of bottlenecking on management bandwidth. This is the rationale behind how Peerceptiv structures peer review, with every participant reviewing multiple pieces of work using prompts designed to elicit real evaluative reasoning.
Chris Schunn led this study and is the senior researcher behind most of the peer-review research underpinning Peerceptiv's platform. He is a Professor of Psychology, Learning Sciences and Policy, and Intelligent Systems at the University of Pittsburgh and a Senior Scientist at the University's Learning Research and Development Center (LRDC), where he has directed research projects backed by more than $80M in federal grants. As Peerceptiv's Chief Learning Scientist, his research directly shapes how the platform structures reviewer prompts, rubrics, and feedback workflows.
