When Feedback Does More Harm Than Good – Kluger & DeNisi
That more feedback is always better sounds obvious. It isn't. When researchers Avraham Kluger and Angelo DeNisi reviewed nearly a century of studies on feedback and performance in 1996, they found something uncomfortable: in over a third of the cases examined, feedback led to worse performance, not better.
Their meta-analysis – a study that statistically combines the results of many earlier studies – is among the most cited works on feedback ever published. And its most important lesson is not that feedback is dangerous, but that it matters enormously where it directs the recipient's attention.
What does the research say?
Kluger and DeNisi compiled hundreds of comparisons between groups that received feedback and groups that did not. On average, feedback improved performance – but the variation was large, and a substantial share of the effects were, in fact, negative. Explaining away the bad outcomes as sloppy studies wasn't enough; the pattern demanded a theory.
The theory they formulated is called Feedback Intervention Theory. Its core: attention is limited, and feedback steers where it goes. When attention is directed at the task – what was done and how it can be done better – performance usually improves. When it is directed at the person – what the result says about my competence, how I compare with others – energy goes into defending the self-image instead of solving the task, and the effect can turn negative.
This explains why even praise can backfire: "you're so talented" shifts the focus to the person just as criticism does. And it explains why detailed, task-focused feedback – "here you interrupted before the question was finished" – works even when the message is uncomfortable.
What does this mean for you as a project manager?
Examine your own feedback before you give it: is each sentence about something the person did, or about what kind of person they are? Cut anything that resembles a character judgment – positive or negative – and replace it with a description of the behavior and its consequence.
Be especially careful with comparisons between people. Rankings and "you're the best on the team at this" direct everyone's attention toward self-image and status instead of the work. Compare with the goal instead, or with the person's own previous level – those are comparisons that can be influenced by the next piece of work.
The same filter applies to feedback you receive yourself. When a steering group or client says something that feels like an attack on you: deliberately translate it to the task level. What in the delivery or the communication is it about, and what could be done differently? That's how you rescue the lesson from a clumsy package. It takes practice, because the first reaction to person-directed criticism is almost always to defend yourself – but every time you manage the translation, you have turned a risky feedback situation into a useful one.
How to train this
Kluger and DeNisi's findings are built into how Project-simulator gives feedback: every assessment is anchored in a concrete quote from your own conversation and followed by a task-focused suggestion for improvement – never a judgment of you as a person. You can also practice giving that kind of feedback yourself in simulated one-on-ones, with AI counterparts who react when you drift onto the personal level. Try it free for a week.
In Project-simulator you practice these conversations against an AI counterpart and get feedback grounded in research like this.
Try free for 7 days