You are viewing shpalman

shpalman [userpic]

Golden Balls

14th June 2009 (11:30)

[BPSDB] I don't want to spend too much time picking apart Lionel R. Milgrom's1 reading of Sir Michael Rawlins's speech in his J. Alt. Complement. Med. editorial which has very little to do with Otto Weingärtner's2 recent defence of the attempts of Shang et al. and Maddox et al. to teach homeopaths about doing experiments properly3,4 instead of craply.5,6 Holfordwatch have already taken apart Patrick Holford's attempt at quote mining it, and Badly Shaved Monkey introduced the subject at JREF and badscience.net, and I've tried to explain how the DBRCT is just the most reliable way of working out if your intervention is actually doing anything or not, to minimize the errors and converge on the right answer in the way which Weingärtner2 describes (and it wouldn't be necessary to be scrabbling about in the statistical noise if homeopathy worked as well as some of these people claim it does). More recently (I admit it's taken me a while to get around to finishing this post), David Colquhoun has highlighted the criticism which the National Institute of Clinical Excellence (NICE) and the Medicines and Healthcare products Regulatory Agency (MHRA) have drawn, whose jobs it is to make sure that medicines work and are worth using, over recommendations related to quackery. The NICE guidelines related to lower back pain are especially important in the light of the British Chiropractic Association's attempts to sue Simon Singh, but what's most relevant to this post is the possible illegality of a label on a bottle of “Arnica 30C” pills which says “a homoeopathic medicinal product used within the homoeopathic tradition for the symptomatic relief of sprains, muscular aches, and bruising or swelling after contusions” when (a) the pills contain no Arnica, what with it having been diluted by a factor of 1060, and (b) there is no evidence that these pills will do anything at all; in fact there is positive evidence that homeopathic Arnica pills do nothing.7

It's also true what Milgrom says about science being “more in hock to powerful interest groups” such as the French homeopathy company Boiron, who were paying two of the co-authors on the discredited4 Nature paper from the Benveniste group.5 Except he said “science” when he meant “quackery” (and see also “Homeopaths in sacka with Big Quacka”).

What remains is the erroneous8 belief that Lüdtke and Rutten and Rutten and Stolper9,10 have discredited Shang et al.3 when in fact Rutten and Stolper10 merely make a bunch of obvious8 false statements about Shang et al.3 while Lüdtke and Rutten9 (having been published in a proper journal rather than the Faculty of Homeopathy's house fansheet)don't say much at all; both of them confirm that rubbish trials make homeopathy look more likely. Milgrom writes that

Randomness of experimental reproducibility, however, is not the sole preserve of homeopathy. This phenomenon is exhibited during studies of parapsychology and psi phemomena, and Weingärtner's arguments are general enough include these. Indeed, it is just conceivable such arguments might be usefully applied to other areas (e.g., the known reduction in effect sizes obtained from RCTs on drugs, compared to effect sizes obtained in real-life practice,11 or tackling another important area of science where random reproducibility exists [i.e., in systems close to chaos]).12

Listen, Lionel: we know that if we're doing a noisy experiment with a weak effect, that we need to do it well and repeat it many times to build up good statistics and see if the effect is really there. That's all this means. Bernoulli knew it and Shang et al. knew it. There's no magical new insight here. We've already seen that the effects of homeopathy disappear into the noise as the experiments are done better.3 You can't just take the badly-done studies with positive effects as proof that homeopathy works and dismiss the negative ones as flawed or subject to (completely unphysical) “non-local” effects. Science would be impossible in a world as you see it. Clearly, science is possible otherwise I wouldn't be able to write this post and you wouldn't be able to read it, and this is objectively true despite whatever post-modern harbles-du-jour13 you subscribe too, so give up.

The other editorial related to this article,2 from Alex Hankey,14 will be next.

.pdf of Sir Michael Rawlins's presentation

Keep libel laws out of science

(and see also Fusion is a dish best served cold).free hit counter javascript

  1.  L. R. Milgrom, J. Alt. Comp. Med. 15, 205 (2009).
  2.  O. Weingärtner, J. Alt. Comp. Med. 15, 287 (2009).
  3.  A. Shang, K. Huwiler-Müntener, L. Nartey, P. Jüni, S. Dörig, et al., The Lancet 366, 726 (2005).
  4.  J. Maddox, J. Randi, and W. W. Stewart, Nature 334, 287 (1988).
  5.  E. Davenas, F. Beauvais, J. Amara, M. Oberbaum, B. Robinzon, A. Miadonna, et al., Nature 333, 816 (1988).
  6.  P. B. Hill, J. Hoare, P. Lau-Gillard, J. Rybnicek, and R. T. Mathie, Vet. Record 164, 364 (2009).
  7.  E. Ernst, and M. H. Pittler, Arch. Surg. 133, 1187 (1998).
  8.  P. Wilson, Homeopathy 98, 127 (2009).
  9.  R. Lüdtke, and A. L. B. Rutten, J. Clin. Epidemiol. 61, 1197 (2008).
  10.  A. L. B. Rutten, and C. F. Stolper, Homeopathy 97, 169 (2008).
  11.  M. C. Michel, and M. Goepel, Eur. Urol. 38, 40 (2000).
  12.  N. Hall, and ed, The New Scientist Guide to Chaos (Penguin, 1992).
  13.  A. Sokal, and J. Bricmont, Intellectual Impostures (Economist Books, 2003).
  14.  A. Hankey, J. Alt. Comp. Med. 15, 203 (2009).

This document was translated from LATEX by HEVEA.

No HTML allowed in subject

  
 
   
 

Notice! This user has turned on the option that logs your IP address when posting. 

(will be screened)