Demonstrating how the famous curve can emerge from pure statistical noise
What you're seeing: This simulation demonstrates a critical flaw in the famous Dunning-Kruger effect research. We've generated completely random data - no psychological effects, no actual correlation between competence and self-assessment - yet the characteristic "D-K curve" still appears.
How it works: We create 100 random data points with independent "actual performance" and "perceived ability" scores. Then we sort participants by actual performance, divide them into quartiles, and plot the average scores for each group.
What this means: The left chart shows the classic Dunning-Kruger pattern - low performers appear "overconfident" (above the diagonal line) while high performers appear "underconfident" (below the line). But since our data is completely random, this reveals that the pattern can emerge from statistical methodology alone, not psychological phenomena.