> ## Content Index > Fetch the complete content index at: https://www.altiusjournal.com/llms.txt > Use this file to discover other available public pages before exploring further. # Why Feedback Sometimes Makes People Perform Worse - URL: https://www.altiusjournal.com/why-feedback-sometimes-makes-people-perform-worse/ - Published: 2026-11-03T22:57:35.000Z - Updated: 2026-11-03T22:57:35.000Z - Description: Why can accurate, well-intentioned feedback sometimes undermine performance? An evidence-led examination of criticism, attention and workplace learning. - Author: Giani Boldeanu - Tags: series, Leadership, The Human Side of Leadership, Human Performance, Psychology *Correcting a mistake should help someone improve. But a large body of research suggests that what people hear, where their attention goes and what happens afterwards can change the result.* You leave a review meeting with a clear instruction: be more decisive. For the next week, you consult three colleagues before making a decision you would previously have made alone. The feedback was meant to strengthen your judgement. Instead, you are spending more time checking it. This is an illustrative situation, not a reported case. It captures a puzzle that extends well beyond awkward performance reviews. How can advice that accurately identifies a weakness make the next performance worse? An influential meta-analysis by psychologists Avraham Kluger and Angelo DeNisi examined 607 effect sizes drawn from 23,663 observations. Feedback interventions improved performance on average, yet more than a third of the effects were negative. The number describes observations across analyses, not 23,663 distinct people, and the findings do not mean that a third of all everyday conversations about performance will be harmful. They do challenge a simpler idea: that giving people corrective information necessarily makes them better at what they do. \[1\] **The short answer:** Feedback changes more than what a person knows. It may change what they pay attention to, what they infer about their competence and whether they have a workable next step. The same criticism can be useful, irrelevant or disruptive depending on the task, the recipient and the environment. ## The message is not the mechanism Imagine that a department head delivers a poor presentation. The organisation wants a clearer argument next time. There are several ways to describe the shortcoming: the conclusion arrived too late; the evidence did not support the recommendation; or the presenter lacked executive presence. These statements are not interchangeable. The first two identify features of a task that can be examined. The third is a much broader judgement, with uncertain instructions attached. A person can agree that it is true and still have little idea what to do differently on Monday. Kluger and DeNisi's feedback intervention theory offers a useful explanation, although it is not a complete account of every workplace exchange. Feedback can draw attention towards learning the task, towards motivation to perform it, or towards more self-focused concerns. The authors found that feedback became less effective as attention moved away from task processes and closer to the self. \[1\] A presenter's attention might turn towards organising evidence and testing an argument. Or it might be captured by the question of whether colleagues now consider them inadequate. The latter reaction is plausible, not an inevitable effect of criticism. People differ, and their responses depend on the circumstances. The important point is that an accurate observation and a helpful learning process are separate achievements. ## Being judged is different from knowing what to change Criticism can carry practical information and social meaning at the same time. A senior employee may hear a question about a particular decision as a question about their standing. Someone new to a role may interpret an otherwise ordinary correction as evidence that they do not belong. It would be a mistake to describe every uncomfortable response as fragility. A person can feel embarrassed and still learn. It would be equally mistaken to assume that all negative feelings come from poor feedback. Sometimes an uncomfortable assessment is accurate, consequential and necessary. The issue is whether the assessment remains usable. A vague instruction such as 'show more leadership' can make a recipient more conscious of scrutiny without making the task clearer. In contrast, 'bring two options, the evidence for each, and the trade-off you recommend' describes an observable next action. There is no universal phrase that guarantees a good outcome. But specificity allows people to test whether they have understood the point. It gives a subsequent conversation something more precise to evaluate than an impression. ## The organisational setting matters A manager may offer thoughtful advice to an employee who has neither the authority nor the time to act on it. A team may demand independent decisions while requiring approval for minor changes. In those cases, the obstacle is not necessarily a failure to listen or a lack of motivation. A 2021 meta-analysis of the *feedback environment* examined 112 independent samples representing 31,089 workers. Perceptions of the everyday feedback environment were positively associated with supervisor-rated performance and feedback orientation, and negatively associated with burnout. These were associations, not proof that changing the environment would produce those outcomes. The authors also caution against treating a positive feedback environment as a cure-all. \[2\] This distinction is important. A development conversation is embedded in a system of responsibilities, incentives and working relationships. If an employee is told to exercise judgement but punished whenever an uncertain decision fails, feedback about decisiveness cannot be evaluated separately from the rules surrounding the job. The relationship may matter too. A criticism from someone who understands the work and applies consistent standards may be interpreted differently from an identical message delivered by someone perceived as arbitrary. That is a reason to investigate the context, not to assume that every perception is accurate. ## What the recipient brings to the conversation Feedback does not land on a blank page. People have different habits of seeking, interpreting and using information about their performance. A 2023 meta-analysis of feedback orientation, covering 46 independent samples and 12,478 workers, found positive associations with job performance (corrected correlation 0.35) and feedback-seeking behaviour (0.43). It also found an association with job satisfaction. These findings do not show that instructing someone to 'take feedback better' will cause an improvement in their work. They indicate that receptivity is a meaningful part of the broader picture. \[3\] It is tempting to turn this finding into advice to accept criticism without resistance. That would miss the point. Useful receptivity requires judgement: Is the observation supported? Does the speaker understand the task? What information may be missing? Can the proposed change be attempted and evaluated? Disagreement can be productive. A recipient may be aware of constraints that the evaluator has overlooked. The aim is not unquestioning compliance. It is a better account of the work and a clearer basis for the next decision. ## Why leaders must examine the process, not only the person There is a management habit worth challenging. When an employee does not improve after feedback, the organisation may conclude that the employee resisted it. Sometimes that is correct. But other explanations deserve examination first. Was the instruction concrete enough to apply? Did the employee have a chance to practise? Was the evaluation based on a single outcome rather than the information available when the decision was made? Did the organisation provide conflicting expectations? Could the evaluator have mistaken a different working style for poor performance? These questions do not excuse negligence. They distinguish accountability from the easier practice of attaching blame to an individual without examining the conditions in which the work occurs. I find this distinction especially useful when thinking about human performance: capability, behaviour and circumstances are related, but they are not interchangeable. A person may know what to do and still operate within constraints that make the desired behaviour difficult. Conversely, an accommodating environment cannot indefinitely replace a missing skill. That observation is an analytical perspective, not evidence from an undocumented client case. The scientific research should carry the weight of factual claims. Psychological safety research adds a related, but distinct, consideration. In a 1999 field study of 51 manufacturing work teams, Amy Edmondson reported an association between a team's shared sense that interpersonal risks could be taken and its learning behaviour. This is not a study showing that feedback automatically improves performance when people feel safe. It is relevant because discussing errors, asking questions and trying alternatives are part of what happens *after* a performance conversation. \[4\] The question for a leader is therefore broader than 'Have I told this person what is wrong?' It is also 'What makes the next attempt possible?' ## The uncomfortable limit of simple advice If the evidence seems to resist a neat rule, that is because feedback research itself covers several different things. Some studies examine a discrete intervention; others examine a person's orientation towards feedback, the relationship with a supervisor or the surrounding workplace climate. Frederik Anseel and Elad Sherf's 2025 review of 25 years of organisational feedback research argues that the field has often relied on divergent assumptions and loosely connected research traditions. Findings should not be collapsed into one universal formula for giving advice. \[5\] This is an important corrective to familiar workplace prescriptions. Praise first, criticism second. Give feedback immediately. Make it direct. Make it gentle. Each formula may sound sensible in some circumstances, but the evidence does not license a single script for every task and person. The better question is what kind of learning the situation requires. A procedural error may call for a clear correction. A strategic decision may require an examination of assumptions. An inexperienced employee may need practice and examples. A senior specialist may need conflicting evidence or a careful challenge to their reasoning. These are different problems, even when they all appear under the heading of feedback. ## From evaluation to development An organisation can complete every performance review on its calendar without knowing whether its people have become better at their work. Counting completed conversations measures the process, not the learning. A more useful follow-up might ask three things: What specifically should change? What would enable that change? And how will we know whether the next attempt was better? These are practical editorial questions drawn from the evidence and reasoning above, not a clinically validated assessment tool. The answers may reveal that a person needs clearer information. They may reveal a need for practice, more authority, different working conditions or a more accurate evaluation. Occasionally, they may reveal that the original feedback was wrong. Feedback deserves neither automatic praise nor automatic suspicion. It can be indispensable, and it can be unhelpful. The difference is not settled by how firmly or kindly it is delivered, or by whether the recipient initially likes hearing it. A correction describes a gap between what happened and what was wanted. Development begins only when that gap becomes something a person can understand and work on. Confusing those two moments is one reason organisations can become very good at evaluating performance without becoming equally good at improving it. --- ## Research notes and 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\. [https://doi.org/10.1037/0033-2909.119.2.254](https://doi.org/10.1037/0033-2909.119.2.254?ref=altiusjournal.com) — 607 effect sizes, 23,663 observations, d = .41 on average; more than one-third of effects negative. Note distinction between effect sizes, observations and unique participants. \[2\] Katz et al. (2021). *Feedback environment: A meta-analysis.* *International Journal of Selection and Assessment*, 29(3–4), 305–325\. [https://doi.org/10.1111/ijsa.12350](https://doi.org/10.1111/ijsa.12350?ref=altiusjournal.com) — K=112, N=31,089; correlations, not intervention effects. \[3\] *Feedback orientation: A meta-analysis.* (2023). *Human Resource Management Review*, 33(4), 100986\. [https://doi.org/10.1016/j.hrmr.2023.100986](https://doi.org/10.1016/j.hrmr.2023.100986?ref=altiusjournal.com) — k=46, n=12,478; corrected correlations, noncausal. \[4\] Edmondson, A. (1999). *Psychological safety and learning behavior in work teams.* *Administrative Science Quarterly*, 44(2), 350–383\. [https://doi.org/10.2307/2666999](https://doi.org/10.2307/2666999?ref=altiusjournal.com) — Field study of 51 teams, not an experimental test of feedback delivery. \[5\] Anseel, F., & Sherf, E. N. (2025). *A 25-year review of research on feedback in organizations: From simple rules to complex realities.* *Annual Review of Organizational Psychology and Organizational Behavior*, 12, 19–43\. [https://doi.org/10.1146/annurev-orgpsych-110622-031927](https://doi.org/10.1146/annurev-orgpsych-110622-031927?ref=altiusjournal.com) — Review highlights heterogeneity in assumptions and research streams. ### Keep reading ALTIUS Subscribe for new essays and analysis, or sign in if you already have an account. [**Subscribe**](https://www.altiusjournal.com/#/portal/signup) · [Sign in](https://www.altiusjournal.com/#/portal/signin)