What's the harm?

It's important not to confuse anecdotes with evidence. When your friend tells you that he has personally seen a ghost, this is not a datum: it is an unverifiable, subjective experience that when you dig deep will almost certainly fall afoul of the argument from ignorance: "I can't explain what I experienced, therefore it was a ghost." When an acquaintance tells you that her sister was diagnosed with Guallain-Barré syndrome after receiving an influenza vaccine, you are not justified in claiming that vaccines are Big Pharma's way of keeping us sick.

But, as Tim Farley has said many times, anecdotes can be useful. (That's why they're so popular with all manner of hucksters!) Most people find it more easy to identify with a personal story. I am reminded of the ever popular quotation, "The death of one man is a tragedy, the death of millions is a statistic," often attributed (erroneously, perhaps) to Joseph Stalin.

And some anecdotes benefit from extensive documentation, which helps. Like this one:

You might have read the case of Ying Wu this week: a fully qualified traditional chinese medicine doctor operating out of a shop in Chelmsford who for several years prescribed high doses of a dangerous banned substance to treat the acne of senior civil servant Patricia Booth, 58, reassuring her that the pills were as safe as Coca-Cola. Following this her patient has lost both kidneys, developed urinary tract cancer, had a heart attack, and is now on dialysis three times a week. Judge Jeremy Roberts gave Wu a two-year conditional discharge, saying she did not know the pills were dangerous and could not be blamed, because the practise of traditional Chinese medicine is totally unregulated in Britain, a situation which he suggests should be remedied.

Wow. Feel that impact? A two-year conditional discharge? She "could not be blamed"? How about practicing medicine without a license? The substance was banned, after all! Can't something be done?

Okay. So how about this:

A survey of 122 otolaryngologists identified 21 ear injuries resulting from ear candle use.

See the difference? The first is an anecdote, the second, data. Both involve personal injury due to fraudulent medical practices, and the second involves more people (one of whom punctured an eardrum—more interested now?). But it's "just a statistic".

Perhaps anecdotes can prove useful when dealing with quacks and other purveyors of woo—so long as they are examples backed by the evidence.

What do you think?

Tip o' the rubber ducky to Ben Goldacre.