Feedback is one of the most-studied influences in all of education — and one of the most misunderstood in practice. A grade at the top of an essay is feedback. So is a two-minute conversation that changes what a student does next. Only one of them reliably moves learning.
What the Evidence Base Shows
The Visible Learning MetaX synthesis of feedback research pools 10 meta-analyses, more than 1,100 studies, and over 3,400 effect sizes. The overall weighted mean effect size is 0.50 — "potential to considerably accelerate" achievement. But the range across contexts is the real lesson:
Notice the pattern: feedback helps most where it is specific, task-focused, and comes while the student can still act on it.
Four Rules of Feedback That Works
- Feed forward. Comment on work the student is about to revise, not work that's finished and filed.
- Name the next step. "Tighten your topic sentence in each paragraph" beats "good organization: B+".
- Less is more. One improvement target outperforms a page of margin comments — students act on the first one and skim the rest.
- Protect the relationship. Kluger and DeNisi's classic meta-analysis found roughly a third of feedback interventions actually hurt performance — usually feedback aimed at the learner's ego rather than the task.
Start Small
Pick one upcoming assignment and change only your feedback move: fewer comments, all task-focused, all actionable, delivered before the final draft. That single change is one of the highest-leverage hours available in teaching.
