The Dunning-Kruger Effect at Work: Why the Least Competent Speak the Loudest

Picture the scene. A strategy meeting at a large retail company. Ten people around the table. A newly appointed operations director presents his plan to restructure the entire logistics chain within three months. He speaks with unwavering certainty. Lays out a clear vision. Has an answer ready for every objection.

At the other end of the table sits the head of supply chain. Fifteen years of experience. She has seen three similar projects, watched two of them fail. She takes notes. Asks almost no questions. At the end she says: “There are a few things we need to think through carefully.”

The director looks at her pleasantly. “I hear you, but I think we can do this.” The meeting closes. The plan goes ahead.

Six months later: the project has been halted, two managers have left and an external consultancy has been brought in to contain the damage.

This is the Dunning-Kruger effect at work. The most confident voice in the room was the least qualified. The most qualified voice stayed quiet, precisely because she understood how complex it really was.

What is the Dunning-Kruger effect?

In 1999, David Dunning and Justin Kruger published a paper that surprised the academic world, though it will probably surprise you less as a manager. They studied students on logical reasoning, grammar and humour, and found a consistent pattern: the students who performed worst overestimated their own performance the most.^[1]

The reason is painfully straightforward. To recognise incompetence, you need exactly the skills you lack. You cannot spot a bad chess move if you do not know what a good one looks like. You cannot identify a weak strategy if you have no reference point for what a strong one involves.

This is a System 1 process. The brain classifies the situation at lightning speed and gives you a feeling of understanding, even when that understanding is shallow. System 2, the slower, more critical mode that questions assumptions and tests hypotheses, never gets activated. Why would you doubt yourself when you have already figured it out?

The effect has three dimensions that recur constantly in organisational life. People with low competence overestimate their own skill systematically. They fail to recognise genuine competence in others. And they are poor at calibrating their performance even after training or feedback.

You cannot see your own incompetence with the instruments that make you incompetent.

There is an ironic flip side. People with high competence tend to underestimate themselves. They unconsciously assume that what feels easy to them must feel easy to others. This makes them more cautious in meetings, more tentative about taking up space. The genuinely good people hold back. The less skilled people fill the silence.

Three scenarios where it does the most damage

The manager who restructures a department in eight weeks

It is a pattern I encounter repeatedly in our Behavioural Design strategy work. A new manager joins the organisation. Not yet hindered by the complexity that has frustrated his predecessors for years, he sees the solution clearly. He acts fast and decisively. Presents a reorganisation plan. Sends out implementation steps before he has really got to know the people.

His certainty is convincing. His energy is infectious. Senior leadership is pleased by his decisiveness. The people on the ground watch with growing concern, but say nothing. Because who are you to contradict the new boss?

What the new manager does not see: the informal structures that keep the organisation running, the unwritten rules about who decides what, the two key players whose informal networks make execution possible. Those things you only learn by taking your time. By listening. By accepting that you do not yet know everything there is to know.

Dunning and Kruger described it this way: the skilled novice believes he is an expert. The actual expert recognises how far there is still to go. In leadership, that gap is the difference between an organisation that grows and one that lurches from crisis to crisis.

The brainstorm problem: juniors dominate, seniors go quiet

You know the scene. An innovation session. Thirty ideas in an hour. The junior team members are everywhere: throwing out ideas, building on each other's suggestions, filling the room. The energy is high. It feels productive.

Then there is the senior designer who has already launched three comparable products. She listens. She occasionally asks a question that nobody quite follows. Near the end of the session she says, almost in passing: “I saw something similar at Unilever. It went differently than expected.” But the momentum has gone. The group is sold on its own plan.

This is what the Dunning-Kruger effect does to team dynamics. Beginners are untroubled by doubt because they do not yet know what can go wrong. They speak with certainty, and in meetings certainty is confused with quality. Experts see the complexity and become more careful, more nuanced, quieter.

The result is a meeting room that systematically surfaces the wrong information. Not the most valuable input wins, but the most confident voice. And that has nothing to do with competence.

The performance reviews nobody trusts

Study after study shows the same thing: people rate themselves higher on average than their managers rate them. But the distribution is not even. The most significant overestimation sits consistently with the lowest performers. The highest performers are the ones who tend to underestimate themselves.^[2]

This has direct consequences for HR. If an employee rates their own work an eight while their manager would give a five, the conversation has a structural problem: the employee experiences the manager's rating as unfair. After all, they have supplied their own evidence that they are doing well. That evidence is their own perception, and that perception feels as solid as external data.

Performance reviews without structured, objective criteria become a source of friction rather than development. The employee who most needs feedback is the least receptive to it. And the employee who is genuinely performing well underreports their own achievements, meaning they receive less recognition than they deserve.

Why awareness is not enough: an Influence Framework analysis

When I analyse the Dunning-Kruger effect through the SUE Influence Framework, it immediately becomes clear why the standard approach of telling people the bias exists simply does not work.

The Influence Framework looks at four forces that determine whether someone changes their behaviour or not. When you apply those four forces to the Dunning-Kruger effect, you see something striking: the forces that maintain the current behaviour are nearly impossible to overcome through information alone.

The Pains of overconfident behaviour, the damage it causes, are real but invisible to the person who has the bias. You cannot feel the costs of a bad decision if you cannot see that it is a bad decision.

The Gains of better-calibrated behaviour, better decisions, more trust from colleagues, a more robust strategy, are abstract and located in the future. They do not compete with the immediate experience of certainty.

Then there are the Comforts. This is where the Dunning-Kruger effect draws its power from. Overconfidence feels good. Certainty is comfortable. Not doubting yourself costs no mental energy. The feeling of competence, even when that feeling is unfounded, is one of the strongest psychological rewards we know. The brain actively protects it. It filters out information that threatens it. It interprets feedback as a failure of understanding on the other person's part rather than as a signal.

The Anxieties complete the picture. Questioning your own competence is existentially threatening. What if you have been less good than you thought, for years? What does that say about your decisions, your career, your professional identity? That fear is large enough to override any rational argument about “becoming more aware of your biases.”

The conclusion of this analysis is always the same as with any other behavioural challenge: you do not change people by informing them. You change the system they operate in. The intervention is environmental design, not awareness raising.

Five interventions that actually work

1. Make expert consultation mandatory. Ensure that major decisions cannot be made until the relevant experts have been formally consulted and their input documented. Not as an option, but as a step in the decision-making process. This lowers the threshold for experts to voice concerns and raises the threshold for overconfident behaviour. You are redesigning the environment, not the person.

2. Introduce pre-mortems as standard practice. Before any major decision, ask the team: “Imagine it is six months from now and this has gone wrong. What happened?” This reframes the conversation from confirmation to exploration. It gives people social permission to name problems without being seen as a pessimist. And it activates System 2 thinking at the moment it still matters.

3. Use structured peer feedback with specific criteria. Generic feedback like “good job” or “could be better” does nothing for calibrating self-perception. Specific criteria, tied to observable behaviour, make the gap between self-perception and reality visible without feeling personal. This is the only kind of feedback that can reduce the Dunning-Kruger effect over time.

4. Make uncertainty status-enhancing. In some organisational cultures, expressing doubt is a weakness. In well-designed organisations, it signals depth. When leaders visibly ask questions, admit what they do not know and consult experts, they model the behaviour that breaks the Dunning-Kruger dynamic. Culture does not change through posters. It changes because the person at the top is visibly doing something new.

5. Design performance processes that calibrate self-assessment. Ask employees to rate their own performance before they see the manager’s rating. Then use the gap between the two as the starting point for the conversation. Not to judge, but to make the difference in perception visible. This is a proven technique that increases receptiveness to feedback, precisely because it lets the employee see the discrepancy pattern themselves.

How it connects to other biases

The Dunning-Kruger effect rarely operates alone. In the workplace it combines with other biases in ways that amplify the damage.

Confirmation bias reinforces it: someone who overestimates their own competence is less inclined to look for contrary evidence, and less likely to find it when it exists. The two biases lock together in a closed loop.

The halo effect makes it visible in hiring and promotion: people who project confidence tend to be rated as more competent, regardless of the actual quality of their work. Certainty gets confused with capability. This means Dunning-Kruger employees dominate rooms and receive higher evaluations while objectively performing below the level of the quieter experts beside them.

Social proof reinforces the group dynamic: when the most confident person in a meeting pushes their idea, the group follows. Not because the idea is best, but because confidence functions as a social signal for competence. The group then validates the overconfident judgement, completing the loop.

And optimism bias makes the forward-looking damage even greater: people with Dunning-Kruger patterns overestimate not just their current competence but also their future performance, the speed at which projects will be completed and the probability that their plan will work.

Frequently asked questions

What is a real example of the Dunning-Kruger effect at work?

A new employee who has used a software system for two weeks confidently explains to the IT team how it really works. A senior developer with ten years of experience sits quietly beside them. The newcomer does not know what they do not know, which makes them fearless. That is the Dunning-Kruger effect in action: the absence of knowledge makes the absence of doubt possible.

How do I recognise the Dunning-Kruger effect in a manager?

Signs include: making major decisions without consulting experts, showing little interest in feedback, rarely asking for clarification on complex issues, and presenting uncertain situations with absolute confidence. Paradoxically, the less they know, the less they realise they should be asking questions. Another clear signal is how they respond to contrary information: people with Dunning-Kruger patterns interpret criticism as a failure of understanding on the other person’s part, not as a warning worth taking seriously.

Can you fix the Dunning-Kruger effect with training?

Training helps at the margins but does not solve it. The Dunning-Kruger effect is a System 1 process: you cannot reason your way out of it. The structural solution lies in redesigning decision-making environments: mandatory expert consultation, pre-mortems as standard practice, and cultures where expressing doubt is a sign of depth rather than weakness. You change the environment, not the person.

What is the difference between confidence and the Dunning-Kruger effect?

Confidence is a positive trait that supports performance. The Dunning-Kruger effect is a cognitive error: confidence that is not grounded in actual competence but in the inability to see one’s own incompetence. The difference is visible in how someone responds to feedback. A person with genuine confidence actively seeks feedback and remains open to correction. Someone exhibiting a Dunning-Kruger pattern does not need feedback because, as far as they are concerned, they have already figured it out.

Why do experts stay quiet in meetings while beginners dominate?

Experts understand how complex a problem really is. They see the exceptions, edge cases and risks, which makes them more measured in their statements, more nuanced, sometimes slower to speak. Beginners do not see that complexity and therefore speak with more certainty. In meetings, certainty is routinely confused with competence. The most confident voice gets the most airtime, regardless of the quality of the insight behind it. This is not a culture problem. It is a structural problem that you can address by redesigning how meetings are run.

Conclusion

The Dunning-Kruger effect is not a problem of bad intentions. It is a problem of architecture. People do not know what they do not know, and that inability is not a character flaw but a cognitive mechanism you cannot train away.

What you can do: design the environment so that the best information surfaces, regardless of who presents it with the most confidence. That starts with accepting that the loudest voice in the room is not the most reliable one.

Want to learn how to structurally improve decision-making in your organisation and diagnose cognitive biases in real-world contexts? In the Behavioural Design Fundamentals Course you learn to apply the Influence Framework and the SWAC Tool to genuine organisational challenges. Rated 9.7 by 5,000+ alumni from 45 countries.

PS

At SUE we believe the most valuable contribution behavioural science can make to organisations is not labelling people but redesigning systems. The Dunning-Kruger effect is not a stamp you press onto colleagues. It is an invitation to ask a different question: how do we design an environment where the best information wins, rather than the most self-assured person in the room?