The less someone knows about a subject, the more likely they are to overestimate their grasp of it — real expertise tends to bring more self-doubt, not less

Confidence is easiest before the hard part has introduced itself. A person who has just learned the vocabulary of a subject can feel strangely fluent. They know enough to speak, not yet enough to hear what is missing. That is why early certainty can be so seductive in work, leadership, investing, hiring, strategy and technology.

The psychology behind this is usually called the Dunning-Kruger effect. In its popular form, it is often drawn as a joke about people knowing nothing and feeling brilliant. The research version is more precise, and more useful: people with low skill in a specific domain often lack the metacognitive tools needed to recognise how poor their performance is.

The original paper by Justin Kruger and David Dunning, published in 1999 in the Journal of Personality and Social Psychology, tested people on humour, logic and grammar. Across studies, the lowest performers tended to substantially overestimate how well they had done.

The double problem

The central insight was not simply that people like to think well of themselves. It was that the same incompetence that produced poor answers also made poor self-evaluation more likely. If someone lacks the skill needed to do a task well, they may also lack the skill needed to tell the difference between a good answer and a bad one.

That is the double burden. A novice may not only make mistakes. They may also miss the signals that those mistakes are mistakes. They may leave a meeting, a code review, a market analysis or a negotiation feeling clear, not because the work was strong, but because the weak parts were invisible to them.

Later work pushed the idea further. In a 2008 Organizational Behavior and Human Decision Processes paper, Joyce Ehrlinger, Kerri Johnson, Matthew Banner, Dunning and Kruger explored why poor performers often remain unaware, testing explanations such as lack of incentives, self-presentation and limited self-insight.

Why expertise can feel less certain

Expertise changes the emotional texture of knowledge. Beginners often see the clean outline of a field. Experts see exceptions, edge cases, measurement problems, hidden assumptions, trade-offs, failure modes and the limits of their own evidence. What looks like self-doubt from the outside may simply be better resolution.

This is why the second half of the title matters. Real expertise does not always make people timid, but it often makes them more conditional. An expert is more likely to say “it depends” and mean it. They can see where the question is underspecified. They know which variables would change the answer.

Kruger and Dunning’s original paper also found that top performers sometimes underestimated their relative standing. One explanation is that skilled people may assume tasks that are easy for them are also easy for others. Their error is different from the novice’s error: less “I know everything” and more “surely everyone can see this.”

The workplace version

In companies, the pattern shows up wherever surface fluency is easy to mistake for understanding. Someone learns the language of AI, fundraising, product-led growth, cybersecurity, user research or organisational design, then begins speaking with a confidence that outpaces their actual grasp.

This is especially dangerous when the domain has high status. Buzzwords can create a costume of competence. A person who can repeat the right terms may be treated as knowledgeable long before they can make reliable judgments. In fast-moving sectors, the appearance of certainty often travels faster than skill.

David Dunning later widened the issue beyond one famous effect. In a 2004 Psychological Science in the Public Interest review, he described flawed self-assessment across health, education and the workplace, arguing that people often have trouble judging their own competence in consequential settings.

The meme version is too simple

The internet version of the Dunning-Kruger effect often shows a dramatic curve: a spike of beginner confidence, a valley of despair, then a slow rise toward expert confidence. It is memorable, but it is not the original finding. The actual research is about calibration: how self-assessment compares with performance.

That distinction matters because the effect has critics. In a 2020 Intelligence paper, Gilles Gignac and Marcin Zajenkowski argued that much of the observed pattern can be explained statistically, especially by regression toward the mean and the broader tendency for people to rate themselves above average.

That critique does not make overconfidence disappear. It does make the clean folk story less secure. Poor performers often overestimate themselves, but the size, cause and interpretation of the pattern depend on how performance and self-assessment are measured.

Feedback is the missing corrective

The cure for false confidence is rarely humiliation. It is contact with reality. Good feedback makes performance visible. It gives people standards, examples, error signals and comparisons that their own judgment may not provide.

In a 2006 Journal of Personality and Social Psychology paper, Katherine Burson, Richard Larrick and Joshua Klayman showed that perceptions of task difficulty can drive miscalibration in relative comparisons. People do not judge themselves in a vacuum; they judge through assumptions about the task and about other people’s performance.

That is why environments matter. Teams that reward confident performance more than accurate thinking create conditions where shallow certainty thrives. Teams that reward revision, evidence and calibrated confidence make it easier for people to notice what they do not yet understand.

Self-doubt is not the goal

None of this means expertise should look like paralysis. A good engineer, investor, researcher, founder or manager still has to decide. The difference is that expertise separates confidence from certainty. It can act while still naming assumptions.

Healthy doubt is not a personality flaw. It is a diagnostic instrument. It asks: What would change my mind? What evidence am I missing? Which part of this problem am I simplifying? Who has seen this fail before? Where am I mistaking vocabulary for understanding?

That kind of doubt is not the enemy of competence. It is often one of its first signs.

The quiet mark of knowing more

The more a person learns, the more the subject stops looking like a room and starts looking like a building with locked doors. They can still navigate it better than a beginner. They just no longer confuse the lobby with the whole structure.

This is why the least informed voice can sound the cleanest in a meeting. It has fewer constraints to hold. The expert’s answer may be slower, narrower and more caveated, but that caution is not weakness. It is contact with complexity.

Real expertise often sounds less impressive than early confidence because it has learned what confidence costs. It does not need to pretend the unknown is small. It has spent enough time with the subject to know that the unknown is usually where the important part begins.