Sunday, 16 February 2014

On scientific fraud

A comment on the impending legal proceedings between Mark Steen and Michael Mann over at Barry Bickmore's Climate Asylum site set me thinking.  It is this from Brandon Shollenberger:
David Appell, your responses make you appear unhinged. You begin tamely, saying other people getting the same results means Mann’s results are trustworthy. In saying this, you completely ignore what I said about this argument in my comment. The reality is it is a silly claim. If I falsified data to “prove” a point, my results would be bogus no matter what. They wouldn’t become okay if other people after me actually proved that point.
 It is rather boring to heave to remind everyone that every time an independent examination of Michael Mann is done, no fraud is found. Even the obvious should become obvious to some people in the end. Obviously not, it would seem.

But what of actual scientific fraud? Well, it does happen and some high profile cases have been uncovered, but the more famous the claim, the more chance there is that your fake discovery is going to be rumbled. And in spite of something like ten years of trying, Mann hasn't been rumbled. Yes, he made some errors but that wouldn't be surprising because he was inventing the techniques as he went along.

Anyway, the way that scientific fraud gets uncovered is twofold. Firstly, the original paper can be scrutinised and an egregious error spotted. It can happen at any stage before publication and by anyone with the right skills after publication.  I wouldn't say that all the examples of papers that are retracted are examples of deliberate scientific fraud but I wouldn't be surprised if some are.

Secondly, replication doesn't happen. Famously, the cold fusion fiasco of 1989 unravelled for Fleischman and Pons when replication didn't happen. They made a bold claim which was unsupported by the well established theory surrounding cold fusion, and it would have come as little surprise that attempts at replication were made and failed. Was there fraud here? Only those making the initial claim can say and one of those is dead.

There have been some high profile cases of scientific fraud. Andrew Wakefield's paper on the MMR vaccine is one such example given in Wikipedia. By concealing his huge conflict of interest, not being straightforward over his results and more, Wakefield should be a pariah. Instead, there are still some who claim he was a martyr.  One thing that caught Wakefield - his results could not be replicated.

Marc Hauser's example possibly does not cross the line into deliberate fraud but his results could not be replicated either.  This opened the door to greater suspicions and colleagues and former students gave evidence that data was falsified. Hauser's reputation is low as a result. His scientific research, in evolutionary psychology, didn't have the same political impact as Wakefield's.

Perhaps mostly forgotten now is the scandal of Sir Cyril Burt who researched the heritability of IQ using twins who were separated at birth. His data is suspect since the study group he was interested in is always going to be small but he managed to find them. Or not.  It is hard to prove fraud in Burt's case but to me the evidence seems overwhelming.

The climate science deniers are only interested in one target - the Michael Mann hockey stick. This reconstruction of the temperature record of the Earth going back centuries has withstood some of the greatest scrutiny of all scientific papers and still holds up. There has been much huffing and puffing but that darn hockey stick graph just keeps on keeping on. And replication has strengthened its case. If, and I don't think for one second that Mann has committed any fraud, the whole thing or a crucial part of it were made up, it would be odd indeed that Mann got the right result. After all, the chances of that are slim.

It was easier for Mann to find ways to make different bits of science fit together rather than make the whole thing up.

The motive for shouting fraud is plain. Anyone can see from the hockey stick graph that the temperature is going up and quickly. So it is essential for deniers to discredit it. That replication reproduces the shape is incidental. Get Mann and the science collapses. It is difficult not to assign a malicious motive to the attempts both to discredit the science and the scientist. Mark Steyn has found out that accusations can bring consequences. In Steyn's case I suspect he has just followed the baying hounds on various internet blogs and is ill informed as a result. He needn't be, of course, as there are plenty of sites that could put him straight.

I saw one denier saying that the global warming thing was all about banning fossil fuels and that was why climate change was invented. As conspiracies go, it would make a poor one. After all, scientists keep undermining it by saying that individual weather events cannot be blamed on climate change. And the effects are all a long way off, even if they are happening now. Sea level rise is slow and temperature rises over a human lifetime might not be noticeable, especially since human memories are poor, and remembering the feeling of a temperature isn't something we are programmed to do (compare that with smells). Wouldn't smoke be something scientists could use to help ban fossil fuels? Well, not since the introduction of smokeless fuels after the deadly smogs of the 1950s in London.

Scientific fraud does happen. But in the case of Michael Mann it hasn't.

1 comment:

  1. I am not sure whether Mark Steyn is just ill informed. If I understand it correctly, his previous accusations have resulted in a number of investigations that have all cleared Michael Mann. A a consequence he should really know better and is thus more likely to be found guilty of libel as someone just stupidly parroting WUWT & Co.

    If there were a conspiracy of climatologists against humanity, I would have voted for claiming it gets colder. That produces much more fear. Unfortunately, no one asked me.