Misleading Data in Aggregation

Simpson's Paradox is a statistical phenomenon where trends apparent in different groups disappear or reverse when the groups are combined. A famous example is the Berkeley sex bias case, where the university appeared biased toward admitting men overall. However, when admissions data were analyzed department by department, each department was found to favor women applicants. This paradox reveals how data can be misleading when viewed at different levels of aggregation.

Previous Fact Next Fact