Glossary
Dunning–Kruger Effect
What is the Dunning–Kruger effect?
The Dunning–Kruger effect is a pattern in which lower performers overestimate their performance or relative standing on a task. One proposed explanation is that the skills needed to answer correctly also help people recognize when they are wrong.
What the original research found
Kruger and Dunning compared people’s estimates with their scores on humor, grammar, and logic tasks. People in the lowest performance group substantially overestimated their standing. Improving task skills also helped participants recognize their limitations.
This is more specific than the popular idea that confident people are incompetent. To assess the pattern, you need both actual performance and the person’s estimate. Experts can be overconfident, too.
A Dunning–Kruger effect example
Imagine someone predicts a high score on a grammar test but performs poorly. The gap illustrates overestimation. Researchers must compare estimates and performance across people to assess the Dunning–Kruger pattern; one confident mistake is not enough to establish its cause.
Is the effect real?
The explanation and size of the pattern remain contested. Gignac and Zajenkowski argued that common quartile plots can produce misleading patterns through statistical regression. Their alternative analyses did not support the predicted effect in an intelligence sample.
Jansen, Rafferty, and Griffiths later compared models in two large grammar and logic studies. They found support for lower performers being less able to judge whether their answers were correct. The task and measurement method matter to the conclusion.
What the effect does not mean
The finding concerns performance on a particular task, not a type of person or a disease. A knowledgeable person can misjudge a different skill. Popular “knowledge curve” diagrams are not a measured sequence that every learner follows; the original comparisons grouped people by performance.
Compare predictions with results and seek feedback on the actual skill. That is more useful than labeling someone you disagree with. Confidence alone tells you neither how good a person is nor how accurately they judge themselves.
The illusion of validity concerns confidence in judgments despite weak or unreliable supporting information.