Moving cases from one group to another can raise the average of both groups, so measured outcomes improve while no individual is better off.
Feinstein, Sosin and Wells named it in 1985 from cancer staging. Better scans found small metastases that older methods missed, so patients formerly counted as early stage moved into the advanced group. They were the sickest of the early group and the healthiest of the advanced one, so survival statistics improved for both while nobody lived a day longer. The name comes from Will Rogers's joke that Oklahomans moving to California raised the average intelligence of both states.
In practice
- A school moves its weakest top-set pupils into the middle set and reports higher averages in both sets.
- A country narrows its definition of unemployment so that discouraged workers are counted elsewhere, and two statistics improve at once.
- Standing
- A pure artefact of reclassification, which is why comparisons over time need the classification held fixed. Compare Simpson's paradox and Goodhart's law.
- Subject
- Science & Epistemology
- Essay
- The ways evidence misleads