“Quotation is a servicable substitute for wit” is a quote frequently attributed to both William Somerset Maugham and Oscar Wilde. And it’s true, it can be. But in the weekend edition of my newspaper the NRC Handelsblad this weekend, there was an article on statistics and results in medical scientific research. An interesting topic, on which XKCD wrote a very funny and to the point comic a while ago: http://xkcd.com/882/. I don’t think you even have to be able to read Dutch to notice the similarities with the illustration accompanying the article in the newspaper.
Not exactly a quote, more like plagiarism if you ask me. Especially considering they don’t mention the source or the name of the author of the original. And I seriously doubt Randall Munroe ever saw a kickback. But of course Rik van Schagen is welcome to drop me a line if I’m wrong (or who knows, the joke wasn’t an xkcd original either, though I was unable to find older versions of it).
Interestingly, they didn’t even seem to quite get the joke. The point of the comic seems to be that, if you repeat a test twenty times and you compare their p-value to a (common) significance level of 0.05, you’re likely to get about one result causing you to reject the null hypothesis. Or, the E-value of the series is 20 x 0.05 = 1. If you’re interested, Wikipedia has an explanation in more detail. By moving the (p<0.05) result to the end instead of having it sit somewhere in the middle like in the xkcd original, they suggest the scientists stop as soon as they get a positive, but the joke is about the total number of colors tested combined with the general misunderstanding (or simple omission) of the scientist's qualification of p>0.05.— tags: nrc, xkcd.