Feedback That Teaches: What the Meta-Analyses Actually Say

01The Surprising Finding

John Hattie's synthesis of thousands of educational studies, published in Visible Learning (2009), identified feedback as one of the highest-leverage influences on achievement. But buried in that finding is a crucial caveat: roughly a third of feedback interventions in Avraham Kluger and Angelo DeNisi's landmark 1996 meta-analysis reduced performance compared with receiving no feedback at all. The average helps; the average hides a lot.

What separates the feedback that teaches from the feedback that stalls — or damages — turns out to be specific and replicable.

02What Works, and What Doesn't

The critical variable is where the feedback is directed. Hattie and Timperley's framework distinguishes feedback aimed at the task, feedback aimed at the process, and feedback aimed at the self. Evidence consistently shows that task-level and process-level feedback produce learning gains. Feedback directed at the person — praise or criticism of who someone is rather than what they did — tends to be neutral at best and actively harmful at worst, particularly when it draws attention to the ego rather than the work.

Timing matters too, but not in the way most people assume. Immediate feedback helps when a learner is still acquiring a skill and needs to correct errors before they consolidate. Delayed feedback often outperforms immediate feedback once basic competence is established, because the gap forces retrieval and deeper processing — an effect that overlaps neatly with the testing effect. Flooding a novice with corrections slows them; giving an expert only end-of-task summaries leaves errors unaddressed.

Specificity is the third variable. Vague feedback — "good effort," "needs improvement" — provides no actionable information. Effective feedback closes the gap between current performance and a clear standard, tells the learner what to do differently, and leaves them with a concrete next step.

03Applying This

For managers and teachers alike, the practical hierarchy runs roughly as follows. First, make the standard explicit before the task begins — feedback without a reference point is just noise. Second, focus comments on process and strategy, not talent or effort as fixed traits. Third, calibrate timing to the learner's level: novices need frequent, specific corrections; experienced performers benefit from being made to self-assess first. Finally, keep the signal-to-noise ratio high. Voluminous feedback trains people to skim; two sharp, actionable points are routinely more effective than a page of annotations.

The evidence is consistent across workplace training, sport coaching and classroom instruction: feedback is powerful precisely because it can be done badly. Getting the direction, timing and specificity right turns it from a ritual into a genuine teaching tool.