Why did a famous statistician dispute that smoking causes cancer?

Ronald Fisher, described as the single most important figure in 20th-century statistics, disputed that smoking causes lung cancer because, he argued, a correlation between the two did not prove that one caused the other. That warning is real and worth understanding, but he was also a heavy pipe smoker and a paid tobacco consultant, and he kept denying the danger long after the evidence came in.

The tool is called confounding, a form of systematic error that can distort estimates of cause and effect in studies where nothing was experimentally controlled. A confounding variable is something that's linked to both the thing you're studying and the outcome, without actually being part of the chain connecting the two, and if you don't account for it, you can end up blaming the wrong cause entirely. In the classic version of this problem involving smoking and lung disease, something like alcohol use or diet could act as a confounder: both are linked to smoking and to health outcomes, but neither one is actually caused by smoking itself, so ignoring them risks a misleading picture. Confounding is why correlation doesn't automatically mean causation.

Fisher himself had introduced the word 'confounding' back in 1935, in a book about designing experiments, though he meant a narrower technical problem than this one. So when a 1950 study showed that smoking causes lung cancer, he publicly spoke out against it, arguing that correlation does not imply causation.

But Fisher's own circumstances tell a different story than pure scientific caution. He was paid as a consultant by tobacco companies during this very controversy, and he was a heavy pipe smoker himself. His biographers argued that his real motive was a dislike of puritanical tendencies of all kinds, and perhaps the comfort he found in tobacco, while others have suggested his analysis was biased by professional conflicts and his own love of smoking. After moving to Australia in 1957 he kept denying tobacco harm until his death in 1962, even enlisting a German eugenicist to his cause, and he called the idea that cigarettes are dangerous 'propaganda.' Confounding is a real and important concept, but Fisher's story is also a reminder that a real tool can be used to defend a wrong conclusion.