The Dunning Kruger Effect

The most dangerous person in any room is the one who doesn’t know what they don’t know.

They’re confident. They’re decisive. They speak with authority. They’re absolutely certain they’re right. And they’re making terrible decisions that they don’t even realise are terrible, which is precisely what makes them so dangerous. It’s not the incompetent person who knows they’re struggling that you need to worry about. It’s the incompetent person who thinks they’re nailing it.

This is the Dunning-Kruger Effect, and once you understand it, you’ll see it everywhere. In meeting rooms. In politics. On social media. In the mirror, if you’re paying attention. And in my profession, when it goes unchecked, it can literally kill people.

What Unconscious Incompetence Actually Means

The Dunning-Kruger Effect is the tendency for people with limited knowledge or skill in a particular area to overestimate their own ability. Not just to be wrong about how good they are, but to be so wrong that they can’t even identify the mistakes they’re making. Their incompetence creates a cognitive blind spot that renders it invisible to them.

In learning terms, this is called being unconsciously incompetent. You don’t know what you don’t know, and because you don’t know what you don’t know, you have no way of knowing you don’t know it (Kruger and Dunning, 1999). It’s the kind of sentence that sounds circular until you’ve seen it play out in real life, at which point it stops being funny very quickly.

The popular shorthand is ‘ignorance is bliss.’ And it might be bliss for the person experiencing it. It’s considerably less blissful for everyone around them who can see the gap between their confidence and their actual capability and is quietly hoping nothing goes seriously wrong.

Why This Matters in Medicine

I train other doctors. Trainees would rotate through for six months to a year, sometimes longer. They range from newly qualified doctors who are bright, enthusiastic, and know things I’ve long since forgotten, to experienced clinicians approaching the end of their specialist training. Every single one of them gets the same treatment in those first few weeks: I go through every consultation note, every test result, every referral letter looking for areas of improvement.

Mostly, what I find is impressive. Real skill, genuine compassion, clinical knowledge that’s rigorous and up to date. But occasionally I find gaps. Sometimes, predictable gaps are identified. Sometimes not. And what I’m always specifically looking for, the thing I take most seriously when I find it, is evidence of the Dunning-Kruger Effect.

Because in medicine, overconfidence isn’t just embarrassing. It’s dangerous. One of the most common ways any of us learn is through trial and error, through making mistakes and adjusting. In most fields, that’s an acceptable, even healthy, part of development. In medicine, you’d prefer the errors to be somewhat limited in number and consequence. And if you’re running the Dunning-Kruger Effect, you’re not even registering the mistakes as mistakes. You’re filing them away as correct decisions and carrying on.

Research shows that medical errors attributable to overconfidence and knowledge gaps are a significant cause of adverse patient outcomes (Berner and Graber, 2008). You genuinely cannot fix what you don’t know is broken. And a doctor who doesn’t know something is broken, but is absolutely certain everything is fine, is the most concerning clinical scenario there is.

I had a trainee once, fresh out of medical school, bright as anything, who missed a serious diagnosis because he was too confident in his initial impression to go back and question it. He thought he knew enough. He didn’t. The patient was fortunate. It was the kind of near-miss that should never happen, and it happened entirely because the trainee’s confidence was running well ahead of his knowledge.

I have a standing rule: when in doubt, ask. Not asking is never the clever option. Better to feel momentarily stupid for five minutes than to make a mistake that has lasting consequences. That rule has saved patients. I’m certain of it.

The Confidence Curve

You’ll often see the Dunning-Kruger Effect illustrated as a curve: confidence spikes early, dips into a ‘valley of despair’, then climbs slowly back up as real competence develops. It’s a popular illustration of the idea rather than a direct finding from Kruger and Dunning’s own research — their original study showed something flatter and less dramatic: poor performers substantially overestimated their ability, and top performers slightly underestimated theirs. But the popular curve still points at something real, and it’s worth understanding, because it says something useful about where you might be in your own development.

When you first start learning something new, your confidence spikes fast. You’ve picked up the basics. You can do a few things. You’ve got the vocabulary. You feel capable. That’s the early overconfidence the Dunning-Kruger Effect describes — you know just enough to feel capable and not nearly enough to realise how much you’re missing.

Then, as you learn more, something shifts. You start to see the complexity you didn’t know was there. You encounter problems you can’t solve. You realise how much genuine experts actually know, and the distance between where you are and where they are becomes visible for the first time. Your confidence drops, sometimes quite dramatically. That’s often where real learning begins, because it’s where you stop operating on assumption and start operating on genuine inquiry.

Eventually, with enough deliberate practice and real experience, your competence rises, and your confidence rises with it. But this time the confidence is calibrated. It knows its own limits. Dreyfus and Dreyfus’s classic stages of skill acquisition, from novice to expert, trace a version of this same arc (Dreyfus and Dreyfus, 1986). Beginners are overconfident. Intermediates are appropriately uncertain. Experts are calibrated. The goal is calibration, not confidence.

Why Intelligence Doesn’t Protect You

Here’s where it gets uncomfortable. You might assume that intelligent people are immune to the Dunning-Kruger Effect. They’re not. In some respects, they’re more vulnerable to it because intelligence in one area creates a kind of ambient confidence that can bleed into areas where it has no business being.

Being genuinely good at one thing doesn’t make you good at everything. A brilliant software engineer can be completely lost when it comes to personal finance. An accomplished cardiac surgeon can have spectacularly wrong ideas about neurosurgery. A talented lawyer can make terrible decisions about his own health. Intelligence doesn’t transfer across domains. It just feels like it should, which is precisely the trap.

Research confirms that highly intelligent people are just as prone to overconfidence in unfamiliar domains as anyone else, and sometimes more so (Ehrlinger et al., 2008). They’re used to being the competent person in the room. When they enter a new domain, the feeling of competence persists even when the actual competence hasn’t arrived yet. The gap between feeling capable and being capable is where the Dunning-Kruger Effect lives, and intelligence doesn’t close that gap. Only genuine learning and honest feedback does.

I see this constantly. The teacher who’s convinced he knows more about his symptoms than his GP. The accountant who read the top three articles on a Google search now has firm opinions about her medication. The senior executive who’s certain he can manage his mental health without professional support because he’s managed difficult problems his whole career. All intelligent people. All running the Dunning-Kruger Effect in a domain where their intelligence gives them no particular advantage.

The Metacognitive Problem

The reason the Dunning-Kruger Effect is so persistent and so difficult to self-diagnose is that it’s fundamentally a metacognitive problem. Metacognition is the ability to think about your own thinking, to assess your own competence accurately. And here’s the catch: you need knowledge to recognise the absence of knowledge. You need the skill to evaluate whether your skills are adequate. The very tools you’d use to identify the problem are the tools the problem has compromised.

This creates a loop that’s hard to break from the inside. Your incompetence prevents you from recognising your incompetence. The gaps in your knowledge make the gaps in your knowledge invisible. You can’t audit yourself accurately using the same faulty equipment you’re trying to audit.

Research on metacognition shows that improving someone’s actual skill level also improves their ability to assess their skill level accurately (Dunning, 2011). As you genuinely get better, you get better at knowing how good you are. The two develop together. Which means the exit from the Dunning-Kruger Effect is the same as the exit from incompetence generally: learn more, practise more, get honest feedback, and accept what the feedback tells you, even when it’s uncomfortable.

The solution isn’t to destroy your confidence. Confidence has real value when it’s earned. The solution is to calibrate it. To match your self-assessment with reality rather than with how you’d like things to be. To develop what researchers call intellectual humility, the ability to hold your beliefs with appropriate uncertainty and revise them readily when evidence suggests you should. Studies show that people with genuine intellectual humility learn faster, make better decisions, and maintain more accurate self-assessments over time (Porter and Schumann, 2018).

Ignorance Versus Stupidity

Dunning and Kruger were awarded the Ig Nobel Prize in Psychology in 2000, the satirical prize for research that makes you laugh and then think, for their paper titled ‘Unskilled and Unaware of It.’ The title does most of the work. The paper confirmed what good clinicians, good teachers, and good managers had suspected for a long time: the people who are worst at something are reliably the least aware of how bad they are.

There’s a distinction worth making here, one often attributed to Benjamin Franklin: ‘We are all born ignorant, but one must work hard to remain stupid.’ Whether Franklin actually said it is doubtful — it doesn’t appear anywhere in his writings — but the distinction it draws holds up regardless. Ignorance is a starting position. It’s the absence of knowledge you haven’t acquired yet. Everyone is ignorant about most things. That’s fine. That’s human. Ignorance is fixable with learning.

Stupidity, in this context, is choosing to remain ignorant. It’s refusing to learn when learning is available. Refusing to listen when someone who knows more is offering to help. Refusing to admit you might be wrong when the evidence strongly suggests you are. The Dunning-Kruger Effect describes ignorance. What you do with that information determines whether it becomes something worse.

Sustained progress depends on your ability to accurately assess where you actually are versus where you think you are. If those two things are significantly misaligned, you’ll make decisions based on a map that doesn’t match the territory, invest effort in the wrong places, skip the learning you actually need because you’ve already decided you don’t need it, and be blindsided by failures you could have seen coming if you’d been looking honestly.

What You Need to Do

Start by assuming you know less than you think you do. Particularly in areas where you feel most confident. Particularly when you’re new to something, the feeling of competence is not the same thing as competence, and treating it as though it is is one of the most reliable ways to stall your own progress.

Seek feedback actively and from people who’ll tell you the truth. Not the people who’ll make you feel good, the people who’ll make you better. Mentors, coaches, experienced practitioners in whatever area you’re developing, people who know what good actually looks like and can tell you honestly how far away from it you currently are. That feedback is uncomfortable. It’s also invaluable. The trainee who asks ‘Am I missing something here?’ will always outperform the one who doesn’t.

When doubt appears, treat it as useful information rather than something to overcome. Doubt is your brain registering that it’s uncertain, and uncertainty is the appropriate response to a situation you don’t fully understand. Pause. Ask. Research. Check. The confident feeling will come back once you’ve earned it, and when it does, it’ll be worth something.

Learn to distinguish between confidence and calibration. Confidence says, ‘I can do this.’ Calibration says, ‘I can do this, and here’s specifically what I can and can’t do within that.’ The second is far more useful for making good decisions and sustaining real progress over time.

Stay curious. Stay genuinely open to the possibility that you’re wrong. Ask the question you feel you should already know the answer to. The most capable people in any field are almost always the ones with the most questions, not the ones with the most certainty. Certainty is a destination. Curiosity is how you keep moving.

Take it slow, ask questions and calibrate. High fliers tend to land the hardest.

Further reading

• Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one’s own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77(6), 1121–1134.
• Berner, E. S., & Graber, M. L. (2008). Overconfidence as a cause of diagnostic error in medicine. American Journal of Medicine, 121(5), S2–S23.
• Ehrlinger, J., Johnson, K., Banner, M., Dunning, D., & Kruger, J. (2008). Why the unskilled are unaware: Further explorations of absent self-insight among the incompetent. Organizational Behavior and Human Decision Processes, 105(1), 98–121.
• Dunning, D. (2011). The Dunning-Kruger effect: On being ignorant of one’s own ignorance. Advances in Experimental Social Psychology, 44, 247–296.
• Dreyfus, H. L., & Dreyfus, S. E. (1986). Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. Free Press.