Anecdotes Do Not Make A Science

Anecdotes - stories recounted in support of a claim - do not make a science. Without corroborative evidence from other sources, or physical proof of some sort, ten anecdotes are no better than one, and a hundred anecdotes are no better than ten. Anecdotes are told by fallible human storytellers. Farmer Bob in Puckerbrush, Kansas, may be an honest, church-going, family man not obviously subject to delusions, but we need physical evidence of an alien spacecraft or alien bodies, not just a story about landings and abductions at 3:00 A.M. on a deserted country road. Likewise with many medical claims. Stories about how your Aunt Mary's cancer was cured by watching Marx brothers movies or taking liver extract from castrated chickens are meaningless. The cancer might have gone into remission on its own, which some cancers do; or it might have been misdiagnosed; or, or, or.... What we need are controlled experiments, not anecdotes. We need 100 subjects with cancer, all properly diagnosed and matched. Then we need 25 of the subjects to watch Marx brothers movies, 25 to watch Alfred Hitchcock movies, 25 to watch the news, and 25 to watch nothing. Then we need to deduct the average rate or remission for this type of cancer and then analyze the data for statistically significant differences between the groups. If there are statistically significant differences, we better get confirmation from other scientists who have conducted their own experiments separate from ours before we hold a press conference to announce the cure for cancer.

[From chapter 3 of Micheal Shermers book "Why People Believe Weird Things"]