All articles
5 min readlearning / training

A third of feedback makes performance worse: give it better

Every training day I run ends with 20 minutes on the work people produced that day: their prompts, their evaluation scripts, their slides. For years I ran those 20 minutes on instinct, opening with what went well and softening the rest. Avraham Kluger and Angelo DeNisi pooled 607 effect sizes covering 23,663 observations and found an average gain of d = 0.41, with more than a third of the feedback interventions lowering the performance they were meant to lift 1.

What the research shows.

Kluger and DeNisi explain the losing third by where the recipient's attention goes. Feedback that holds attention on the task tends to help. Feedback that pulls attention up toward the person, praise included, spends the recipient's effort on defending a self-image instead of repairing the work 1.

Benedikt Wisniewski, Klaus Zierer and John Hattie reran the question on 435 studies, 994 effect sizes and more than 61,000 learners. The pooled effect came out at d = 0.48, and the moderator that explained the spread was how much information the feedback carried. Feedback that supplied task information together with process guidance and cues for regulating the next attempt reached d = 0.99 2. The same analysis found bigger effects on cognitive and motor outcomes than on motivation and behavior, so a comment that fires someone up is doing less than it appears to.

Hattie and Helen Timperley sort feedback into four levels: the task, the process that produced it, the learner's self-regulation, and the self. The first three hand the recipient something to do next. The fourth, which covers most of what trainers call encouragement, leaves a verdict and no move 3.

Jodi Goodman, Robert Wood and Margaretha Hendrickx varied one dimension in a computer-based task: how specific the feedback was. More specific feedback raised performance during practice, and the gain survived neither a delay nor a change to the task. It also cut how much people explored during practice, and exploration is where later independent performance comes from 4. Goodman and Wood published the companion study the same year: more specific feedback helped people learn how to respond when things were going well, and worked against their learning about what to do when performance dropped 5.

James and Chen-Lin Kulik pooled 53 studies on timing. Immediate feedback won in applied studies built on classroom quizzes and real material, while laboratory studies of list acquisition favored delayed feedback 6. A training room is the first setting.

The receiver decides what lands.

Lauren Eskreis-Winkler and Ayelet Fishbach ran five experiments with 1,674 participants on binary-choice questions, the format where being told you are wrong also hands you the right answer. People who received failure feedback learned less than people who received success feedback. The result held across professional, linguistic and social content, and it held with money on the line. Participants who failed also remembered fewer details of the task, and that memory gap carried part of the effect 7.

James Smither, Manuel London and Richard Reilly looked at the aftermath of 360 reviews across 24 longitudinal studies. Improvement in ratings from peers, direct reports and managers came out small on average, and it concentrated in the recipients who accepted that a change was needed 8.

The failure finding is contested.

Stephen Spiller published a commentary in Psychological Science in 2025 arguing that the scoring in the original task is confounded. Someone who learns nothing from the feedback and answers the follow-up consistently gets marked correct after success and incorrect after failure, which can manufacture the gap without any difference in learning 9. Eskreis-Winkler and Fishbach replied in the same journal and stand by their reading. Treat the tune-out mechanism as plausible and unsettled. The step it suggests, checking out loud that the person can state the gap, costs 30 seconds either way.

The protocol.

  1. Write about the artifact. Delete any sentence whose subject is the person: "you tend to", "you are strong at", "you rushed this". Point at slide 7, the retry loop, line 40 of the prompt. Kluger and DeNisi's attention hierarchy is the reason 1.
  2. Carry three parts every time: where the work stands now, the standard it has to reach, and one move that closes the distance. That combination is what the d = 0.99 condition describes 2.
  3. Drop the praise wrapper. Approval aimed at the person reads as a verdict and leaves nothing to act on 3. Praise a move instead: "slide 3 puts the number in the title" gives them a rule they can repeat.
  4. Deliver it in the room, minutes after the work, rather than in a written recap the following week 6.
  5. Fade the specificity across passes. First pass, name the exact fix. Second pass on similar work, point at the location, name the rule, and let them find the move 45.
  6. Ask for the gap in their words before you leave the table. One sentence, out loud. Someone who cannot say it did not take it in 7.
  7. Close on one change and a date. Write down the change, the piece of work it applies to, and the day you will look at it again. The people who improve after a review are the ones who agreed a change was needed 8.

In your next feedback session, count the sentences whose subject is the person. Bring that count to zero, then work on the other six steps.

Sources.

  1. Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: a historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119(2), 254-284. doi.org/10.1037/0033-2909.119.2.254
  2. Wisniewski, B., Zierer, K., & Hattie, J. (2020). The power of feedback revisited: a meta-analysis of educational feedback research. Frontiers in Psychology, 10, 3087. doi.org/10.3389/fpsyg.2019.03087
  3. Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81-112. doi.org/10.3102/003465430298487
  4. Goodman, J. S., Wood, R. E., & Hendrickx, M. (2004). Feedback specificity, exploration, and learning. Journal of Applied Psychology, 89(2), 248-262. pubmed.ncbi.nlm.nih.gov/15065973
  5. Goodman, J. S., & Wood, R. E. (2004). Feedback specificity, learning opportunities, and learning. Journal of Applied Psychology, 89(5), 809-821. doi.org/10.1037/0021-9010.89.5.809
  6. Kulik, J. A., & Kulik, C.-L. C. (1988). Timing of feedback and verbal learning. Review of Educational Research, 58(1), 79-97. doi.org/10.3102/00346543058001079
  7. Eskreis-Winkler, L., & Fishbach, A. (2019). Not learning from failure: the greatest failure of all. Psychological Science, 30(12), 1733-1744. doi.org/10.1177/0956797619881133
  8. Smither, J. W., London, M., & Reilly, R. R. (2005). Does performance improve following multisource feedback? A theoretical model, meta-analysis, and review of empirical findings. Personnel Psychology, 58(1), 33-66. doi.org/10.1111/j.1744-6570.2005.514_1.x
  9. Spiller, S. A. (2025). Commentary on Eskreis-Winkler and Fishbach (2019): a tendency to answer consistently can generate apparent failures to learn from failure. Psychological Science. doi.org/10.1177/09567976251333666

Keep reading