The only way to establish cause and effect with any reliability is to do “controlled” experiments, or controlled trials, as they’re called in medicine. Such trials attempt to avoid all the chaotic complexities of comparing populations, towns, and ethnic groups. Instead, they try to create two identical situations—two groups of subjects, in this case—and then change only one variable to see what happens. They “control” for all the other possible variables that might affect the outcome being studied. Ideally, such trials will randomly assign subjects into an experimental group, which receives the treatment being tested—a drug, for instance, or a special diet—and a group, which receives a placebo or eats their usual meals or some standard fare.

Not even randomization, though, is sufficient to assure that the only meaningful difference between the experimental group and the control group is the treatment being studied. This is why, in drug trials, placebos are used, to avoid any distortion that might occur when comparing individuals who are taking a pill in the belief that their condition might improve with individuals who are not. Drug trials are also done double-blind, which means neither subjects nor physicians know which pills are placebos and which are not. Double-blind, placebo-controlled clinical trials are commonly referred to in medicine as the gold standard for research. It’s not that they are better than other methods of establishing truth, but that truth, in most instances, cannot be reliably established without them.

Diet trials are particularly troublesome, because it’s impossible to conduct them with placebos or a double-blind. Diets including copious meat, butter, and cream do not look or taste like diets without them. It is also impossible to make a single change in a diet. Saturated fats cannot be eliminated from the diet without decreasing calories as well. To ensure that calories remain constant, another food has to replace the saturated fats. Should polyunsaturated fats be added, or carbohydrates? A single carbohydrate or mixed carbohydrates? Green leafy vegetables or starches? Whatever the choice, the experimental diet is changed in at least two significant ways. If saturated-fat calories are reduced and carbohydrate calories are increased to compensate, the investigators have no way to know which of the two was responsible for any effect observed. (To state that “saturated fats raise cholesterol,” as is the common usage, is meaningful only if we say that saturated fat raises cholesterol compared with the effect of some other nutrient in the diet—polyunsaturated fats, for instance.)

Nonetheless, dietary trials of diet and heart disease began appearing in the literature in the mid-1950s. Perhaps a dozen such trials appeared over the next twenty years. The methods used were often primitive. Many had no controls; many neglected to randomize subjects into experimental and control groups.