Strange Correlations

Spurious Correlation: Two things can rise and fall together for years with zero connection between them. It happens for a handful of reasons: pure coincidence when you compare millions of datasets, a hidden third factor driving both (season, wealth, population), shared upward trends over time, reverse causation, and regression to the mean. The golden rule of statistics, correlation is not causation, exists because our brains are wired to see stories in patterns.

Google Correlate: This 2011 Google tool let anyone upload a data series and instantly find search terms that tracked it. It churned out bizarre matches between unrelated products and searches because anything with the same seasonal rhythm, like summer produce and spring car buying, lines up on a graph. Google quietly shut it down in 2019.

Nicolas Cage: From 1999 to 2009, the number of films he appeared in each year tracked the number of Americans who drowned by falling into swimming pools. Tyler Vigen found it by letting software compare thousands of random datasets. Test enough pairs and some will match perfectly by pure chance.

Spurious Correlations (Book): Tyler Vigen's 2015 book shows US per capita cheese consumption matching deaths by bedsheet entanglement with a 0.947 correlation. Both simply drifted upward over the same decade, and any two rising trends will look married on a chart.

Franz Messerli: In 2012 this cardiologist published a tongue in cheek paper in the New England Journal of Medicine showing countries that eat more chocolate win more Nobel Prizes, with Switzerland on top. The real driver is wealth: rich nations buy more luxury sweets and fund more research.

Robert Matthews: In 2000 he published "Storks Deliver Babies (p = 0.008)" showing European countries with more storks have higher birth rates. The hidden factor is land area: bigger countries have more room for storks and more people making babies.

Super Bowl Indicator: Coined by sportswriter Leonard Koppett in 1978, it claimed the stock market rises when an original NFL team wins and falls when an AFL team wins. It was eerily accurate for decades, then collapsed, the textbook sign of a coincidence mistaken for a law.

Hemline Index: Economist George Taylor suggested in 1926 that skirts get shorter in booming economies and longer in downturns. Fashion historians later found hemlines actually lag the economy by about three years, making it useless as a predictor.

Lipstick Effect: Estée Lauder chairman Leonard Lauder noticed lipstick sales climbing after the 2001 recession. The theory: when people can't afford big luxuries, they treat themselves to small ones. A rare strange correlation with a plausible human reason behind it.

Alan Greenspan: The former Federal Reserve chairman reportedly watched men's underwear sales as an economic signal. Underwear is invisible and easy to postpone, so when money is tight, men stop replacing it, and sales flatline before other data catches up.

Benjamin Sandler: During the 1948 polio epidemic this doctor blamed ice cream and sugar, and North Carolina ice cream sales plunged. Both polio and ice cream simply peaked in summer, when heat drove people into crowded pools where the virus spread.

Women's Health Initiative: Observational studies suggested hormone replacement therapy protected women's hearts. This massive trial, halted in 2002, found the opposite. The women taking hormones had been wealthier and healthier to begin with, a trap called the healthy user effect.

Flying Spaghetti Monster: In a 2005 satirical letter to the Kansas school board, Bobby Henderson charted the decline of pirates against rising global temperatures and declared that fewer pirates cause global warming. It became one of the internet's most famous lessons in fake causation.

Sports Illustrated Cover Jinx: Athletes who land on the cover often slump right after. The cause is regression to the mean: you only make the cover at a career peak, and after an extreme high, ordinary performance feels like a curse.

Francis Galton: In 1886 he noticed very tall parents tend to have children shorter than themselves, and very short parents taller children. He called it "regression towards mediocrity," discovering the statistical effect behind countless jinxes and miracle cures.

UC Berkeley Admissions (1973): Men were admitted at 44% versus 35% for women, suggesting bias. Broken down by department, women were admitted at equal or higher rates. Women had simply applied to the most competitive departments, a famous case of Simpson's Paradox.

Abraham Wald: In WWII, returning bombers showed bullet holes clustered on the wings and fuselage, so the military wanted armor there. Wald realized the planes hit in the engines never came back, and said to armor where the holes weren't. The data only showed survivors.

University of Pennsylvania Night Light Study: A 1999 Nature paper linked sleeping with a light on as a baby to nearsightedness later. Follow up research found the real link: nearsighted parents were more likely to leave lights on and to pass the gene to their kids.

Strawberry Pop-Tarts: Before Hurricane Frances in 2004, Walmart's data showed strawberry Pop-Tarts selling at about seven times their normal rate ahead of storms. No fridge, no cooking, long shelf life, and kids like them. Beer was the top seller.

Target: Statistician Andrew Pole found that shoppers buying unscented lotion, zinc, magnesium, and cotton balls were likely pregnant. The model reportedly sent baby coupons to a Minneapolis teenager before her father knew, as reported by the New York Times in 2012.

Pentagon Pizza Index: Domino's franchisee Frank Meeks claimed late night pizza orders to the Pentagon and White House spiked the night before Iraq invaded Kuwait in 1990. Staff working overtime during a crisis need to eat, so pizza becomes a secret signal.

Osco Drug: A 1992 data mining analysis of its stores reportedly found men buying diapers and beer together between 5 and 7 p.m. The beer and diapers story became data science legend, though the chain never actually moved the shelves.

Waffle House Index: FEMA administrator Craig Fugate used the status of local Waffle Houses to gauge disaster damage. Green means full menu, yellow means limited menu, red means closed. The chain is so good at staying open that a closed one means things are truly bad.

Redskins Rule: Discovered in 2000 by Steve Hirdt of the Elias Sports Bureau, it held that if Washington's NFL team won its last home game before a presidential election, the incumbent party kept the White House. It worked every election from 1940 to 2000, then broke in 2004.

Google Flu Trends: Launched in 2008 to predict flu outbreaks from search activity, it overestimated the 2013 flu peak by nearly double the CDC's numbers. Media panic drove healthy people to search flu symptoms, proving that searches track attention, not illness. Google shut it down in 2015.