By Andrew Gelman in his blog.
James Coyne wrote to me a couple weeks ago:
This time I’m critiquing a horrible mediational analysis. The larger context is that the authors have refused all requests to share data that would be needed to make an independent evaluation of their interpretation. I am now in what will be a highly visible confrontation with them because I’ve asked for related data that they published in PLOS One. PLOS One has strict data sharing policies and we are awaiting a Thursday midnight deadline for a response from their university.
He sent me a copy of the paper (Rehabilitative therapies for chronic fatigue syndrome: a secondary mediation analysis of the PACE trial, by Trudie Chalder, Kimberley Goldsmith, Peter White, Michael Sharpe, and Andrew Pickles) that he’d criticized. I glanced at the paper and his comments and wrote back to him:
Maybe it’s worth just ignoring this stuff. There’s so much bad research out there! Or is the topic itself important, and you’re concerned that the bad research will have a malign effect on policy?
Before going on, that you should not consider my above statement as a diss of the Chalder et al. paper, nor, for that matter, should you consider this disclaimer as a statement of support. I looked at the paper only very quickly, and I have a skepticism of pretty much any of these mediation models (as illustrated, for example, by figure 7 from Chalder et al.), so I responded to Coyne with a generic statement that sometimes it’s better just to ignore bad research.
I agree it’s an absolute disaster of a paper. It otherwise should not be the object of my attention but there is some nastiness associated with its dissemination. Peter White, one of the authors is paid for working with a group trying to get social welfare payments denied to people with chronic physical illness. He promotes the idea that chronic fatigue syndrome, whatever its origins, is perpetuated by fears of exertion. The deconditioning model that is the basis for these claims has been discredited. So he is forced to rely on dodgy research like this to promote the view that chronic fatigue syndrome is largely “cognitive” and psychosomatic. I’ve seen horrible testimony from him that patients who “appear” to be disabled should be denied handicap parking passes because that will force them to walk more. He has personally testified in a number of hearings where patients with chronic fatigue syndrome have had the benefits revoked.
To read the rest of this blog, click on the link below: