 The same reason it was so difficult to study cancer among coffee drinkers is the same reason it's so difficult to study cancer among meat eaters. Even if they found lower cancer rates among those eating vegetarian, maybe it's just because vegetarians exercise more, or smoke less, or inhale less diesel fumes because they all own a Prius. So the way you get around that is you study a group of healthy meat eaters, who, for example, smoke just as infrequently as the group of vegetarians you're trying to study. To equal things out, control for non-dietary factors. So you don't just classify people into meat eaters, fish-only eaters, and vegetarians. You adjust for smoking, past smoking, current smoking, the amount of smoking, cigarette smoking, cigar smoking, pipe smoking, alcohol consumption, body mass index, physical activity level, for women how many children they've had, which can be protective against breast cancer, whether or not they're on the pill, anything they could think of to factor out everything else aside and just focus in on what they were interested in, whether or not one eats meat. Now, controlling for obesity is not really fair to the vegetarians. We know that vegetarians are significantly more likely to be thin, which we know is protective against cancer. So by comparing vegetarians only to thin meat eaters, it undercuts one of the benefits of eating vegetarian. It effectively erases one of the reasons why eating vegetarian may reduce cancer rates. But they weren't interested in indirect ways in which meat might cause cancer, like meat leading to obesity leading to cancer. They wanted to study meat and cancer more directly. And to do that, you have to handicap the vegetarians even further. Maybe the reason vegetarians are so healthy is not because they eat less meat, but because they eat more plants. So vegetarians were compared to meat eaters who on average ate about the same amount of fruits and vegetables every day. May not have been easy, but they were able to dig up thousands of meat eaters who ate four to five servings of fruits and veggies a day about as much as the vegetarians were eating. Again, this puts the vegetarians at a comparative disadvantage by removing one of the key benefits of more plant-based diets, which is more plants. By comparing vegetarians to omnivores who don't eat a lot of meat and have a high fruit and vegetable intake, this could reduce the chance of observing lower cancer rates in the vegetarians, but they wanted to isolate out the meat component, so they did. They compared vegetarians only to healthy meat eaters with healthier diets, and still, though there was one surprising finding I'll cover later, bottom line, they found the incidence of all cancers combined was lower among vegetarians.