The First John Maddox Prize

The panel that chose the winners of the first John Maddox Prize — Colin Blakemore, a British psychologist, Tracey Brown (Sense About Science), Phil Campbell (Nature), and Brenda Maddox — deserve a prize for Most Contentious Award. The Maddox Prize is supposed to be awarded to people who have excelled at:

any kind of public activity, including all forms of writing, speaking and public engagement, in any of the following areas:

  • Addressing misleading information about scientific or medical issues in any forum.
  • Bringing sound evidence to bear in a public or policy debate.
  • Helping people to make sense of a complex scientific issue.

The first winners, announced in November, were Simon Wessely, a British psychiatrist, and Fang Shi-min, a Chinese journalist. Criticism of Fang is here. Criticism of Wessely is here (in the comments) and here. One of his papers is here. Wessely is best known for promoting the use of cognitive behavioral therapy (CBT) to treat people with chronic fatigue syndrome (CFS). In particular, “he and his colleagues demonstrated substantial overlap in symptoms between chronic fatigue syndrome and clinical depression. . . . He subsequently developed a treatment approach using cognitive-behavioural therapy techniques, which in many cases brought about substantial improvement.”

The puzzle is that this is considered significant. Maybe people with CFS are depressed because they have CFS? Maybe this is why CBT helps them? A statement explaining the reward does not answer this objection. As for Fang, I have no idea if he deserves the prize. I would be surprised if members of the prize committee could judge for themselves the accuracy and value of his work.

Assorted Links

Thanks to Hal Pashler.

Assorted Links

Thanks to Casey Manion.

Assorted Links

  • Unusual fermented foods, such as shio koji (fermented salt, sort of)
  • David Healy talk about problems with evidence-based medicine. Example of Simpson’s paradox in suicide rates.
  • The ten worst mistakes of DSM-5. This is miserably argued. The author has two sorts of criticisms: 1. Narrow a diagnosis (e.g., autism): People who need treatment won’t get it! 2. Widen a diagnosis (e.g., depression) or add a new one (many examples): This will cause fads and over-medication! It isn’t clear how to balance the two goals (helping people get treatment, avoiding fads and over-medication) nor why the various changes being criticized will produce more bad than good. Allen Frances, the author, was chair of the committee in charge of DSM-4. He could have written: “When we wrote DSM-4, we made several mistakes . . . . The committee behind DSM-5 has not learned from our mistakes. . . .” That would have been more convincing. That the chair of the committee behind DSM-4, in spite of feeling strongly about it, cannot persuasively criticize DSM-5 speaks volumes.
  • The Lying Dutchman. “Very few social psychologists make stuff up, but he was working in a discipline where cavalier use of data was common. This is perhaps the main finding of the three Dutch academic committees which investigated his fraud. The committees found many bad practices: researchers who keep rerunning an experiment until they get the right result, who omit inconvenient data, misunderstand statistics, don’t share their data, and so on.”

Assorted Links

Thanks to Dave Lull and Alex Chernavsky.

Assorted Links

  • Experiments suggest flu shots reduce heart attacks and death. Huge reduction: 50%. The new report (a conference talk, not a paper) is a reanalysis of four earlier experiments. I was surprised to learn that the CDC uses heart attack outbreaks to locate flu outbreaks, implying that the new finding is not a fluke — there really is a strong connection. I already knew heart attacks are more common in the winter, which also supports a connection with flu.
  • Une histoire des haines d’écrivains by Boquel Anne and Kern Etienne. Published 2009. About literary feuds. One of my students was reading a Chinese translation.
  • Correspondences between sounds and tastes.
  • Report on fraudulent Dutch research. “The 108-page report says colleagues who worked with Stapel had not been sufficiently critical. This was not deliberate fraud but ‘academic carelessness’, the report said.” I doubt it. Based on my experience with Chandra, I believe Stapel’s colleagues had doubts but did nothing from some combination of careerism (doing something would have cost too much, for example a lot of time, and gained them nothing), ignorance (not their field), and decency (they saw no great value in ruining someone). I wonder if the report considered these other possible explanations (careerism, ignorance, decency).

Thanks to Tim Beneke.

Assorted Links

Thanks to Dave Lull.

Assorted Links

  • You can major in Fermentation Science. No joke. When I was eight, I learned the concept of college major. I asked my mom, “What did you major in?” “Extracurricular activities,” she said. I failed to get the joke. She later explained she had spent more time working on the school paper than on her classes.
  • In a famous paper, the statistician Ronald Fisher accused Mendel of faking his data. Fisher wrote: “the data of most, if not all, of the experiments have been falsified so as to agree closely with Mendel’s expectations.” This is not terribly consistent with the fact that Mendel’s highly improbable conclusions were correct. It’s as if Fisher had said “Person X used false info to claim he is worth $10 billion” and (b) in fact Person X is worth $10 billion. You can see that (a) and (b) may both be literally correct but that the term “false info” (Fisher’s “falsified”) probably conveys the wrong impression. This paper (“A Statistical Model to Explain the Mendel–Fisher Controversy”) has a more plausible explanation of the pattern in the data that Fisher noticed.
  • Conflict of interest in the Nobel Prize in Literature. The conflicts of interest underlying the Nobel Prize in Physiology and Medicine — which are given out for “pure” science, thus justifying more funding — remain unnoticed by journalists.

Thanks to Bryan Castañeda.

Assorted Links

Thanks to Anne Weiss.

How Common is Dishonesty in Medical Research?

Richard Smith, former editor of the BMJ, writes about Peter Wilmshurst, a British cardiologist, an unusually brave and honest man:

He was the coprincipal investigator on a trial funded by NMT, an American company, to see whether closing a hole in the heart of patients with migraine would cure their migraine. It didn’t. He refused to agree to be an author on a paper published in the journal Circulation because the paper was misleading, and he gave an interview to a journalist in the US pointing out the problems in the study. NMT sued him for libel, not in the US, where proving libel is difficult, but in England, where the onus is on the defendant to prove his innocence. NMT probably assumed (rightly in the case of most people) that the financial risk would cause Wilmshurst to cave in. They were wrong, and the case collapsed when NMT went bust.

The interesting thing about the stories Smith tells about Wilmshurst is the high rate of misconduct they imply:

[In 1996] we invited him to come to the BMJ and give a talk—behind closed doors—to our staff and advisers and colleagues from the Lancet. He reeled off case after case of misconduct, many of them involving prominent people. The audience listed intently, but I was unsure of the reaction. Might somebody leap up and say “How dare you accuse x of misconduct. He is one of the great men of British medicine”? In fact in my memory the reaction was the opposite. People said things like “Actually, it’s worse than you know…” . . .

Many of [Wilmshurst’s stories] involve doctors who are guilty of misdemeanours but who sit in judgement on others. He told the story of Peter Richards who decided to bury the fact that Clive Handler, a doctor, at Northwick Park Hospital, was found guilty of using NHS research funds to subsidise his private practice at a time when Richards was medical director of the hospital and chair of the professional conduct committee of the GMC. Previously he had been dean of St Mary’s Medical School, prorector for medical education at Imperial College, and chair of the Council of Deans of UK Medical School and Faculties. When Handler eventually appeared before the GMC, the GMC’s lawyers ask that Richards stand down from chairing the committee. As Wilmshurst said, it’s as if a judge at the Old Bailey were to say “I’ll have to excuse myself from hearing this case as I helped the accused bury the body.” After having to stand down from this committee Richards continued to chair other conduct committees. Wilmshurst told several stories of doctors who had been found guilty of research misconduct but [had] gone on to be deans and others in charge of researchers.

Smith does not point out what this means. Doctor X is found guilty of research misconduct. Everyone knows this. Doctor X is still appointed dean or whatever. Maybe the people who make these appointments don’t care. Or maybe research misconduct is so common it cannot be a disqualification.

At the end of the article Smith points out his long friendship with Wilmshurst:

RS has known Wilmshurst for 16 years . . . the BMJ was sued for libel over an article by Wilmshurst that was published when RS was editor of the journal. The article has not been retracted but is not available on the BMJ website.

Presumably Smith was forced by BMJ lawyers to be this vague. I have not been able to locate the article. From a talk by Wilmshurst. “Eventually Rubin got his report published in [The New England Journal of Medicine, under the editorship of Dr. Arnold Relman], because he threatened that unless his report was published he would go to the press and point out the collusion between the journal and Sterling-Winthrop.”

Wilmshurst’s conclusion: “Dishonesty is common in medical research.”

Thanks to dearime.