• Theoretically Speaking by Mike Fowler

    I'll use this forum to post my ideas about work I'm doing, work I've read, or things that pop into my head; hopefully to raise discussion and help me learn more about this crazy little thing called science.

    • A load of balls

      Wednesday, 28 Jan 2009 - 13:43 UTC

      I don’t know whether to laugh or cry about this painful musical article from the BBC – a letter in the BMJ in 1974 describing a patient with a medical condition known as cello scrotum 1, has turned out to be as much of a sham as Piltdown man.

      Baroness Ana Chronism Elaine Murphy has just admitted to the jolly jape, but the original letter has been cited at least seven times elsewhere!2 It’s in danger of entering the collective knowledge of the field as a holy truth (a worrying proportion of medical students don’t believe in evolution). One citing article even stated “That the names for the disorders are interesting, however, should not trivialize their importance.3 (my stress). Am I missing something here? Is this a joke that’s snowballing out of control? A Frankensteinianismical catastrophe in the making? Or is it simply a genuine, long running joke that all in the medical community were aware of?

      Commenting on the original letter, a BMJ spokesman quoth “it all adds to the gaiety of life”. They’re now considering a retraction, 25 short years after the letter was first cited. How gay, indeed, especially for all those poor, crippled cellists…

      Now, as a (lapsed) drummer and viola fiddler, I’m relieved that another potentially career threatening injury has been wiped off the musicians’ danger list, and I’m all for stories about spaghetti trees on the 1st of April, but just how seriously should we treat the scientific record?

      This story reminded me of a more “serious” hoax perpetrated in the hallowed pages of the journal Social Text as recently as 1996, by Alan Sokal, a physicist who wanted to illustrate how easy it was to publish an utterly nonsensical article in a postmodern cultural studies journal. While the article was not peer-reviewed, the Editors did ask Sokal to make various changes to the text, which they found difficult to read. He refused, and they published anyway. Sokal didn’t sit back on his laurels, chuckling at his own wit and the vapidity of the journal editors, however. He immediately published a corresponding article elsewhere describing exactly what he had done. Two for the price of one, not a bad scam!

      In the past, I’ve proof read (before submission) an article written by a gaggle of medical doctors, where they highlighted 3 unusual diagnostic case studies of a normally non-fatal liver condition which unfortunately turned out to be something else quite fatal. The sample size of 3 individuals left me astonished – this would never even be sent out to review by a serious ecological journal, although Nature seems to have a different standard 4 so probably isn’t taken seriously by serious ecologists. The main message I took from the medical article, that a fatal disease was being misdiagnosed as non-fatal in humans, was a pretty convincing reason to highlight these cases. I was also surprised by the difference in the weight of evidence that’s required in different biological fields.

      So, back to the question: just how seriously should we treat the scientific record? And just how carefully do we check our references?


      1 Murphy, J.M. (1974) Cello Scrotum British Medical Journal

      2 A web of science citation search shows 7 articles. A pubmed search added one extra that wasn’t on WoS.

      3 Mouse joint—another manifestation of an occupational epidemic?
      L. A. Norman
      West J Med. 1991 October; 155(4): 413–415.
      PMCID: PMC1003031

      For some reason Norman forgot Brewer’s droop and Wanker’s cramp though.

      4 Poole et al. (2005) Elephants are capable of vocal learning. Nature 434, 455-456 (24 March 2005) doi:10.1038/434455a

      “Robust” evidence from a couple of elephants making various grumbling, parping noises.

      Supplementary Audio 1 Mlaika imitating truck.wav
      This audio file starts with the sound of a truck followed by Mlaika’s imitation.

      I guess the truck was broken down the day they recorded this… Sceptical? Moi?

      Last updated: Wednesday, 28 Jan 2009 - 13:43 UTC

      • Comments

        • Date:
          Wednesday, 28 Jan 2009 - 15:02 UTC
          Chris Surridge said:

          The way I heard it Dr Murphy sent the piece in after reading a similar case study on Guitar Nipple in the BMJ earlier in the year. She assumed that had to be a hoax and wanted in on the joke.

          P. Curtis Guitar nipple. Br. Med. J. 2, 274 (1974).

          That is probably related to Cellist Chest first reported in 1962.

          H.N. Mandell Cellist’s Chest N. Engl. J. Med. 15, 348 (1962).

        • Date:
          Wednesday, 28 Jan 2009 - 15:23 UTC
          Mike Fowler said:

          Yep, that was made clear in the BBC article. I was wondering how easy it is for these things to pass from being jokes, to being taken seriously, one or two citations down the line.

          Mandell seems to be an amusing type though, with an article called “Frivolity in medicine. Is there a place for it?” (1988, Postgrad Med 83(8):24-8). I can’t access it, unfortunately, so perhaps the joke’s on me.

        • Date:
          Wednesday, 28 Jan 2009 - 16:14 UTC
          Bob O'Hara said:

          I’m going to be very suspicious about anything you publish in Oikos from now on.

        • Date:
          Wednesday, 28 Jan 2009 - 16:25 UTC
          Eva Amsen said:

          The things with small sample sizes in medical papers is that there often are only small sets of patients. There are debilitating genetic mutations that have only been reported in maybe 6 people, but if they are the only 6 and they all have the same problem, that is the entire sample. I do find case studies sketchy, though, and want to see some molecular biology to back things up, but that’s why I’m not a doctor. Well, I am, I guess – but of the philosophy kind.

        • Date:
          Wednesday, 28 Jan 2009 - 16:59 UTC
          Cristian Bodo said:

          Just for the record, Sokal didn’t stop after publishing the second article describing his scam, but went on to write an entire book, basically chastising postmodernist social scientist for borrowing concepts from the hard sciences…leading to entirely non-sensical (but ver scholarly pieces). It was called “Les Impostures Intellectuelles”, and I think it was translated into English as “Fashionable Nonsense”.

        • Date:
          Wednesday, 28 Jan 2009 - 17:03 UTC
          Chris Surridge said:

          Guitar Nipple has n=3 which is quite a lot for a case study.

          At one time in my life I had to come up with an opinion of whether Medical Case Reports came within the remit of a very wide remit science journal. My opinion (which is not necessarily that of the journal in question) was that on balance they weren’t science at all. If pushed, I take the philosophical position that science is the presentation of data to invalidate hypotheses until what you have left is the truth. Case studies are rather a single (or very small n) observation which consequently don’t have the power to invalidate anything. The medical equivalent of the lists of exotic sightings in a Bird Watching journal.

          Of course a meta-analysis of lots of individual case studies might prove or disprove something but these won’t be affected by jocular inclusions like “Cello Scrotum” aren’t going to mess with those too much as there is only a single recorded case.

        • Date:
          Wednesday, 28 Jan 2009 - 21:00 UTC
          Mike Fowler said:

          Bob, I’ve got one lined up for you already Unfortunately, it’s all serious science. Nothing like the whining Danish or jovial, Czech beer drinking, sociological nonsense Oikos usually publish.

          Eva, medicine was a whole new world of scientific methodology for me. I was trained as a zoologist, where experiments and statistics are designed to try and disentangle a host of unmeasured variables from the one you’re actually interested in. The three "r"s (replicate, replicate, replicate) were drummed into me from an early age.

          Christian, I loved the Sokal story when I first heard about it in a Philosophy of Biology course a few years ago. As Bob hints at, I’m often tempted to try this one myself sometimes. How easy is it to present a model in a theoretical paper that has no relation to the results presented? I don’t think a lot of people actually check… Theoretical ecology is one field where it is often cheap and easy to check maths or simulation results. I try to where feasible.

          Chris, I don’t think meta-analytical studies on a bunch of case studies that have no power to void null hypotheses are ever going to tell us anything more useful. And “Scrotello” now has a sample size of 0. Sounds like a new chocolate bar though.

        • Date:
          Thursday, 29 Jan 2009 - 12:20 UTC
          Cristian Bodo said:

          I’d actually suggest that “Scrotellos” should be a revamped version of Ferreros, but only available in bags of two.

        • Date:
          Thursday, 29 Jan 2009 - 12:37 UTC
          Mike Fowler said:

          Youch! There’s only one way to answer that…


Search blogs

web feed Want a blog?

Submit this post to

Advertisement