Berkeley's admissions looked biased against women — until the data was split up
In 1973, UC Berkeley's graduate admissions showed men getting admitted at a noticeably higher overall rate than women, and looked like blatant discrimination. But when researchers broke the numbers down department by department, women were admitted at the same rate or higher in most departments. How can a bias disappear, or even reverse, once you look closer?
Reveal the answer
Women had disproportionately applied to more competitive departments with lower acceptance rates for everyone, while men applied more to less competitive ones — the aggregate rate reflected that application pattern, not per-department bias. This reversal is called Simpson's paradox, named for statistician Edward H. Simpson's 1951 paper, and the Berkeley case was documented in a landmark 1975 Science paper.