Testimonials and vivid anecdotes are one of the most popular and convincing forms of "evidence" presented for beliefs in the transcendent, paranormal and pseudoscientific. Nevertheless, testimonials and anecdotes are of near zero value in establishing the probability of the claims they are put forth to support. A sincere and vivid account of one's encounter with an angel, an alien, a ghost, a Bigfoot; or purple auras around dying patients, a miraculous dowser, a levitating guru, or a psychic surgeon, is of little empirical value in establishing the reasonableness of believing in such matters. Why? Well, one reason is that such accounts are inherently unreliable and biased. They are of no more value than the televised accounts of satisfied customers of the latest weight loss program. In fact, the advertisers' obviously biased testimonials are superior to most testimonials of those claiming to have seen angels: we can at least imagine a way to test the claims of the advertiser. There is no conceivable test for angel sightings.
The testimonial of "personal experience" has no scientific value. If others cannot experience the same thing, there is no way to determine the reliability of the experience. If there is no way to test the claims made, there is no way to tell if the experience was a delusion or was interpreted correctly. If others can experience the same thing, then it is possible to make a test of the testimonial and determine whether the claim based on it is worthy of belief. For example, a reader sent in the following example of the placebo effect:
As a child I vomited every time I got on an airplane, until a stewardess told me to try drinking alka-selzer before the flight. It worked! Unfortunately, it also tasted horrible, so I tried half a glass. That worked too. After a while just putting the alka-selzer in the glass was enough...This reader was wise enough to recognize that, while testimonials to the wonders of alka-seltzer would be scientifically worthless, it would be possible to test the effects of the stuff. On the other hand, a testimonial to having been transported to heaven during the night to dance with the angels is one which could never be tested. Such testimony is scientifically worthless, not only because it is a testimonial but because it can't be tested.
[Delano DuGarm]
If testimonials are worthless, why are they so popular and why are they so convincing? There are several reasons, I think. Testimonials are often very vivid and detailed, making them appear very believable. They are often made by enthusiastic people, who seem trustworthy and honest, with no apparent reason for wanting to deceive us. They are often made by people with some semblance of authority, such as Ph.D. in psychology or physics. Finally, testimonials are believable because people want to believe them. Nevertheless, testimonials which are of untestable claims are worthless on their face. And testimonials of testable claims are worthless until tested.
further reading
How to Think Straight About Psychology, 3rd ed., Keith E. Stanovich (New York: Harper Collins, 1992).