The US writer Christie Aschwanden demonstrated how easily nutritional epidemiology can go awry in a piece on the website FiveThirtyEight in 2016. She ran a survey of the site’s readers, asking them questions about their diets and a range of personal attributes, such as whether they were smokers or if they believed the movie Crash deserved to win a best-picture Oscar. One association she turned up was that people who are atheists tend to trim the fat from their meat.
Nice example of correlation-is-not-causation on foods.
