Faith healing, self-healing, and related phenomena of the mind's effects on the body present some of the most fascinating puzzles of philosophy, psychology, and medicine. The role of healer or doctor is one of the oldest and noblest, appearing prominently in civilizations throughout recorded history. Yet speculation about the distinction (or lack thereof) between healer and quack has an equally prominent history. It is only during the past century that physicians have had available some scientifically proven treatments for some specific diseases. Most healing in the past relied on successful social influence, quacks and health fraud notwithstanding. And, I will argue, so does much healing today.

A physician who “cures” three very ill patients of the flu by prescribing a new drug is almost sure to believe in the drug's efficacy. The physician thus makes two major errors of inference. First, there is no control group; there is no way to know whether the patients would have recovered without treatment by this drug. Second, the numbers are probably too small to allow reliable inference; in a different set of patients, different results might occur (Kahneman, Slovic, & Tversky, 1982). In brief, this is the problem of distinguishing the artifact from the main effect. Although it appears obvious when presented to experienced scientists, it will trip up more than half a class of medical or psychology students. It will also often trap the individual practitioner. The faulty inference results in the use of many worthless drugs and procedures for many years.