Episode 1 · Cognitive biases · Research notes

The most famous graph in psychology is fake

You have seen the Dunning-Kruger curve: a stick figure climbs "Mount Stupid," crashes into the "Valley of Despair," then rises up the "Slope of Enlightenment." That picture never appeared in the actual study, and a 2020 paper argues the effect itself is partly a statistical artifact.

What did Kruger and Dunning actually find in 1999?

Cornell students took tests in humor, logic, and grammar, then estimated how well they had done. The lowest scorers did not just perform badly; they believed they had done well. In the classic result, the bottom quartile scored around the 12th percentile but placed themselves near the 62nd.

The elegant part is the explanation: the skills you need to do well at something are often the same skills you need to recognize that you did not. If you do not know the grammar rule, you cannot catch yourself breaking it. Incompetence can be invisible to the person who has it.

Where did the Mount Stupid curve come from?

Not from the paper. The real 1999 figure shows four quartiles and two nearly flat lines: everyone estimated themselves around the 60th to 70th percentile regardless of actual skill. The rollercoaster with a towering peak of confidence and a despair valley is an internet illustration that spread because it tells a funnier story. It is packaging, not data.

Is the effect even real?

This is where it gets interesting. In 2020, Gignac and Zajenkowski argued the classic pattern is "(mostly) a statistical artefact": if you analyze random noise with the same method, a similar picture can emerge, because extreme scores regress toward the mean and self-estimates cluster around "slightly above average" for everyone. Other researchers push back with analyses where a real effect survives. The argument is live.

Three things hold up either way: people overrate themselves on average (in a famous 1981 result, 70 to 90 percent of drivers rated themselves above average), the worst performers are the most overconfident, and the famous curve is fiction.

The lemon-juice bank robber who started it

In 1995, McArthur Wheeler robbed two Pittsburgh banks in broad daylight with no mask, was shown the security tape after his arrest, and said, stunned: "But I wore the juice." He believed lemon juice made his face invisible to cameras, because lemon juice can work as invisible ink, and he had even "tested" it with a blurry selfie. David Dunning read the story and asked the question that became the research: can incompetence hide itself from the incompetent?