Bottom Quartile Knew Nothing, Middle Quartile Knew Exactly Where They Stood

The Dunning-Kruger effect is real, and David Dunning's 1999 study genuinely documented something. People at the absolute bottom of a skill set—the bottom quartile on logic tests, grammar tests. Humor recognition—wildly overestimated their performance.

They thought they were median when they were terrible. That part is not a myth—it is an actual empirical finding and it explains something true about human psychology.

But everyone has forgotten what the data actually showed. The people in the second and third quartiles assessed themselves accurately. The overestimation problem was concentrated at the very bottom, not distributed across all incompetent people everywhere.

The effect escaped the lab and became a cudgel. Twitter users deploy it to explain away anyone confident who disagrees with them—ignorant people are too ignorant to know they're ignorant. It became a closed loop, a way to dismiss rather than understand. You see it everywhere in AI arguments, expertise debates, politics, and wellness. Someone makes a claim with certainty.

Confident people must be ignorant.

The actual tension is weirder and less useful as a weapon than the version that lives in culture now. When you encounter someone certain, you have to actually check whether they're in the bottom quartile of the specific domain or whether they're just confident—because those are different problems requiring different responses. One requires instruction. The other requires listening to an argument. Dunning and Kruger gave us precision. We traded it for a way to feel smart while dismissing everyone louder than us.