Eric Schwartz

In 1953, Ancel Keys published a paper showing a relationship between dietary fat consumption and heart disease mortality across six countries. The relationship was striking. The graph was compelling. The problem: Keys had data from twenty-two countries. He selected six. The six that supported his hypothesis. If you plot all twenty-two countries, the relationship disappears. This was pointed out at the time, by Yerushalmy and Hilleboe in 1957, in the American Journal of Public Health. Keys was aggressive in dismissing the criticism. He was influential enough to make it stick. He later chaired the committee that wrote the American Heart Association dietary guidelines. The guidelines blamed fat. The guidelines became government policy. The government policy became the food pyramid. The food pyramid became what your grandmother was told to eat. It began with a man who had twenty-two data points and published six. Everything downstream: the statins, the low-fat products, the LDL obsession, the dietary guidelines your doctor is still referencing, traces back to that selection. Cherry-picking doesn't usually have this kind of legacy. This one did.

It is important to understand this.