This article was first published in the July 2015 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online
Are Americans getting tangled to death in their bedsheets because they're eating too much cheese? The answer, clearly, is no. But you could be forgiven for considering the possibility -- for a moment at least -- when you see Tyler Vigen's graph showing a close correlation between the number of US citizens meeting their end that way and per capita consumption of cheese over a nine-year period. "Two sets of data might appear to be strongly linked but correlation does not equal causation," says Vigen (right), a student at Harvard Law School who set up the website Spurious Correlations in May 2014, to hammer home this point light-heartedly. "There is a danger that people see correlations presented in the media and assume a causal link, so we need to look at this sort of thing critically."
Vigen, 23, wrote a program that combs large data sets published by the US government for statistical correlations, and converts them into graphs. It produces a new graph every minute -- 42,888 so far -- showing connections between bizarre incidents such as margarine consumption and divorce rates in Maine, for example, or the number of people drowning in swimming pools each year and the number of films in which Nicholas Cage appeared. Visitors can discover their own wacky correlations by picking variables to compare from a menu. Vigen posts the most interesting, as voted by visitors, on the website.
A month after he launched Spurious Correlations, Vigen saw a study that reported a strong correlation between Atlantic hurricanes with female names and number of fatalities. The authors claimed that hurricanes with women's names are more deadly because we don't take them as seriously. But other researchers demonstrated that the correlation was likely a statistical fluke. "I want people to be wary of these statistics we see in the news," says Vigen, whose book of meretricious data relationships, Spurious Correlations, is out now. "It just shows how you have to question these things."
This article was originally published by WIRED UK
