Anecdotal evidence
From SkepticWiki
Contents |
[edit] Definition
An anecdote is a narrative, usually brief, explaining an event or an effect. Anecdotes are understood to refer to people that actually exist and events that actually occurred, although sometimes after much retelling they can become myth or urban legend. Anecdotes are generally told by someone who claims to have been there or otherwise speaks with knowledgeable authority.
Anecdotal evidence refers to the attempt to prove a claim by referring to one or more anecdotes, claiming that this constitutes supporting evidence.
[edit] Examples
The following are examples of anecdotal evidence typically used in support of various claims:
- My father and brother both went to chiropractors. They were helped immensely.
- I can tell you we both saw that UFO—it even chased our car as we drove down the street!
- You know, ever since I got this lucky rabbit's foot I haven't had any back problems!
[edit] Anecdote vs. Case Study
A case study is a method of analyzing an effect in a particular instance. Case studies can be beneficial, for example, in medicine; it can be advantageous to closely examine the course of a disease from the earliest possible detection, or examine in detail the effects of a certain drug fighting a disease.
The main difference between an anecdote and a case study is the basic lack of evidence for an anecdote, other than the purported witness testimony (made even weaker in the case of a friend-of-a-friend anecdote). Case studies, on the other hand, are well-documented and supported, even published and peer-reviewed.
However, even case studies make for weak scientific evidence. They can give scientists and researchers valuable information and ideas leading to more discoveries, and can certainly confirm that a particular event occurred, at least in that one case. However, for any useful scientific conclusions to be drawn a full study of a number of cases, properly selected, and enough to be statistically significant, must be performed. For example, just seeing how a single patient reacts to a certain medication is not in itself proof of anything, certainly not compared to a full double-blind placebo-controlled study.
If case studies, rigorously applied and documented, make for such weak evidence, then it's easy to see that anecdotes really don't tell us anything of any scientific value.
[edit] Discussion
Anecdotal evidence tends to encompass many logical fallacies, including confirmation bias, incorrect cause, and post hoc ergo propter hoc. In addition, other possibilities for the effect are rarely examined and the authority of the source is seldom questioned.
Similarly, anecdotal evidence often attempts to generalize from the specific case of the anecdote to everyone everywhere. For example, someone might say, "My cousin had cancer, and we prayed for her, and her cancer went away! Prayer can work for you, too." Even if we forgive the correlation/causation fallacy of the first sentence, using the anecdote to claim a general healing power of prayer simply cannot be inferred.
Anecdotal evidence is most effective when delivered to an audience with the proper bias. For example, an anecdote about a woman raped by an armed man might be very effective when speaking to gun control advocates, while one about a woman scaring off her would-be rapist by brandishing a gun would play well to gun rights advocates. Interestingly, if you give a group the opposing anecdote—i.e., give the defensive gun use anecdote to the gun control advocates—they are likely to dismiss it as unverified or an exception, even though they show no such skepticism towards the anecdote that fits their preconceived conclusions.
Anecdotes can be powerful when stressed and repeated, especially en masse. For example, the widespread anecdote of "Mikey" from the Life Cereal commercial dying from eating Pop Rocks and drinking soda resulted in sales of Pop Rocks plummeting and never recovering, despite the fact that the actor who played Mikey is alive and well to this day and there is no evidence that mixing Pop Rocks and soda will cause anything more than a bad case of gas (as confirmed in an early episode of Mythbusters).
Eyewitness testimony, which is very similar to anecdotal evidence, has been utilized in courts of law for centuries. However, more recently it has been recognized that eyewitness testimony can be unreliable, due to bias, faulty memory, false recognition, and other factors. Also, two eyewitnesses to an event (such as an automobile crash) will often give conflicting reports to the police, even though only minutes have passed since the event occurred. Courts have learned to rely more on physical, tangible evidence, such as that collected by forensic analysts, and use the eyewitness testimony for support.
Anecdotal evidence is often used to challenge skeptics. For example, a proponent of ghost hauntings might give a particularly eerie anecdote and then say, "Now, how do you explain that?" But really, there is nothing to explain, as no real evidence has been presented. Even if the anecdote is true, the skeptic does not have the information needed to properly investigate and evaluate the claim. Also, this is an invalid shifting of the burden of proof.
Another consequence of all of this is that anecdotal evidence tends to be neither reproducible nor falsifiable.
It should be noted that the use of anecdotes as examples of a certain effect, to give the reader a better picture of what is actually being concluded, is perfectly valid and proper. The problem comes when the anecdotes themselves are presented as evidence, and not as examples of something demonstrated with other, valid forms of evidence.
[edit] References
- The Skeptic's Dictionary: Anecdotal Evidence
- Snopes.com page on Mikey's Pop Rocks Death
- The Death of Little Mikey
- Annotated Mythbusters: Pilot 1
