Simpson Paradox
Simpson’s Paradox is a phenomenon where a trend observed within each subgroup of data disappears or reverses when the subgroups are combined (SEP).
How can a trend reverse on aggregation?
Aggregating across a confounding variable can flip the apparent direction of an association. The “right” answer depends on which causal question you are asking.
UC Berkeley admissions (1973)
Men were admitted at higher overall rates than women, suggesting bias. But department by department, women were admitted at equal or higher rates. Women applied disproportionately to more competitive departments.
Overlaps with Selection Bias in spirit but is distinct: Simpson’s is about how aggregation hides a confounder, not about who got into the sample.