spurious correlation
A statistical association that does not reflect a genuine causal relationship—here, crediting 'good autocrats' for growth booms that occur for other reasons.
As defined in William Easterly on the Tyranny of Experts
Where it comes up · 5
- Simeon Djankov and Matt Warner on the Doing Business Report and Development Aid2 claims
- Machine Intelligence and Risk Management1 claim
- Tim Cook’s Dashboard1 claim
- Nature, Nurture, and Identical Twins (with David Bessis)1 claim
- William Easterly on the Tyranny of Experts1 claim
Most important claims · 6 of 6Search all claims for “spurious correlation” →
- Humans are powerful pattern recognizers that can detect signal, but they also find patterns where none exist, so a key skill is separating signal from noise.Machine Intelligence and Risk Management
- Poor countries have far more volatile, boom-and-bust economies than rich ones, so they produce both the best and worst growth rates; since poverty is associated with autocracy, autocrats happen to be on the scene during high-variance outcomes, and attributing booms to 'good autocrats' and busts to 'bad autocrats' is spurious—a co-authored paper found leadership changes do not line up in timing with booms and busts.William Easterly on the Tyranny of Experts
- The dramatic 'Jim twins' coincidences (both named Jim, both married Lindas then Bettys, both had adopted brothers named Larry, both had dogs named Toy) cannot be genetic—genes cannot telepathically influence adoptive parents' choices—and instead reveal that the twins were placed into similar conformist suburban Ohio families, so the apparent natural experiment actually captured a shared social background, not genetic determinism.Nature, Nurture, and Identical Twins (with David Bessis)
- Running every machine learning or statistical method on a dataset and picking the one with the best score is data mining that produces spurious results, and it reflects laziness; an analyst should start with a conviction-based hypothesis and test it.Tim Cook’s Dashboard
- The correlation between good governance/low red tape and low poverty may run in reverse: wealthier countries develop better governance, more transparency, and a more vibrant press as a consequence of wealth, so the relationship could be spurious rather than regulation causing prosperity.Simeon Djankov and Matt Warner on the Doing Business Report and Development Aid
- Every 5% increase in a country's Doing Business ranking is associated with reducing the poverty rate in the average country by about 1 percentage point, a relationship that holds even when examining year-over-year changes rather than levels (to control for slow-moving factors like health, education, and conflict).Simeon Djankov and Matt Warner on the Doing Business Report and Development Aid