I'd much rather write about Christmas (or even the EM Forster novel, come to think of it) than climate change, but something has been troubling me and I thought I'd share my concerns.
I'm worried that 'debate' about climate change has become pointless. Rapid polarisation has turned discussion about the quality of evidence, and the impact of that evidence on one's views about potential AGW, into a football match. Leftwing bloggers write sneering twitter comments about rightwing bloggers' inability to 'get' 'the science'; rightwing bloggers respond with what might be thought of as the written equivalent of 'I'm not listening, nah nah nah, it's cold this week so there can't be any warming at all'. Intelligent people attempt to move the discussion onto terms of economic consequences of specific actions; but I fear they are going to be disappointed.
I'm interested in why this has happened, and what it might tell us about the role of science in politics; science in general, and in particular what I think of as 'inference' but which is usually called 'the evidence'. In fact I could take my pick of issues on which to base this article. From David Nutt and the role of pharmacology in determining drugs policy, to Liam Donaldson and the role of epidemiology in determining alcohol advice to families, it's been a depressing year for people who cling to the irrational hope that there could be a role for rationalism in politics. I doubt this will surprise you, but I blame everyone involved, a bit.
A few months ago I got fed up feeling like a complete ignoramus about climate research. I'm not a physicist but I am a statistician, and I spent two years on a postdoc grant working in an atmospheric physics unit which modelled the spatio-temporal dispersion of atmospheric pollutants, and so I thought there must be some of the literature which I could read, critically, and come to a slightly more informed opinion. So I started reading about the 'hockey stick', a graph which appeared in the 2001 UN IPCC report, which powerfully showed a sharp increase in temperatures around the time of industrialisation in the northern hemisphere. The most famous version of this graph is due to a 1998 publication by three climate change scientists, Michael E. Mann, Raymond S. Bradley and Malcolm K. Hughes ('MPH'). You can read their paper here.
The trouble is, the statistical methods by which this graph was constructed are - I'm going to choose my words very carefully - sub-optimal. Put it this way: if I were a reviewer for a statistical journal, and this paper was sent to me, I would send it back for rework by the authors.
Two quantitative scientists, McIntyre and McKitrick, attempted to reproduce the graph, and failed: hence the 'hockey stick controversy'. (Their website is here). The controversy reached such a level that the US Congress became involved, and commissioned a report led by Edward Wegman of Johns Hopkins university to investigate the original MPH paper. You can read their report here: Download WegmanReport. It's very technical and I found it very readable. I have to say it is a devastating indictment of the MPH methodology. If you set aside the technical discussion of the misapplication of the statistical methods in MPH, the component of the Wegman review which I found most troubling was their finding that MPH had little or no contact with academic statisticians during their work. Since their work relies on statistical modelling, this is more than a little worrying.
OK now we need to pause. Because I cannot underline enough that this initial reading did not make me come to the conclusion that the AGW hypothesis is some vast, leftwing conspiracy, designed to drive us back to the Stone Age, living off lentils (though I do that anyway). I did, however, twitter my negative thoughts about that MPH paper - and what was the response?
I was inundated with comments to the effect that I didn't know what I was talking about, I should have read papers by X, Y and Z and so on. One scientist told me that my reasoning was flawed because Wegman itself wasn't peer-reviewed (it's a report for the US congress), and another told me my conclusions about MPH were groundless because I had said that their graph was on the cover of the 2001 IPCC report (it's not, I was mistaken about that - it appears inside the report).
Now this is all fine, and the same people also sent me some very interesting references which led me to believe that, substantively, the shape of the MPH hockey-stick is as good as it gets in terms of our understanding of the historical progress of temperatures in the northern hemisphere; other researchers have used proper statistical methods and come to the same conclusion as MPH. So (and forgive the repetition) I remain firmly in the group which thinks there is an issue with climate change and we ought to plan how to mitigate it while reducing our dependency on fossil fuels (there are so many other good reasons for this anyway).
But the reaction of the scientists to my very mild criticisms - criticisms based on my statistical reading of a source paper (I have a Ph.D. in statistics, it's not something I've just 'picked up' as I wander through life) - shocked me, I admit. If this is how the climate science community reacts to a fellow scientist's Twittered thoughts, what must be the atmosphere be like within that climate change community? It felt like I was being accused of letting the side down. Essentially, the message was: either you spend your career working in this area, or you can't have an opinion other than to uncritically accept what you are told by us. Having an opinion, of course, is not the same thing as being willfully irrational. There are nuances in science, just as there are in the rest of life. To ask a critical question does not make you a fool, rushing in. To pretend otherwise is as irrational as the modus operandum of the crystal-botherers.
And then came 'climategate'. I will just say that this incident didn't make me revise my opinion about the need for greater transparency over source data, and more visible attempts to reproduce results by different teams of researchers.
I get annoyed when lay-people have a go at the area I work in (we try and develop drugs to treat psychiatric illness). I get annoyed; but I don't try to shout them down (I'd be sacked if I did). I get annoyed; but I don't attempt to substitute my thinking as a single grand theory which only a fool wouldn't support. I get annoyed; but I don't ascribe malignity of intent to those who might not agree with me. I'm very, very careful with the words I use when I speak at Congresses and so on: I don't ever want anyone to feel like I'm trying to beat them up with my opinion. Apart from anything, this is usually politically counter-productive.
What might be some lessons for climate change scientists, in the aftermath of what feels like a largely pointless meeting in Copenhagen? First, I think that openness and transparency have to be re-prioritised. Secondly, I'd calm down the celebrity-support thing. My views about evidence aren't altered one iota for knowing that an actress has turned up to say we've all got to do something, like now. Actually I'd extend that definition of celebrity to include people like Colin Blakemore. I've got nothing but awe when it comes to Professor Blakemore's research in neuro-biology, and the sense he brings to the discussion about the use of animals in medical research, but I thought his Guardian article at the weekend, in which he claimed that Copenhagen was a success for science, a little weak. I'm still trying to work out why he was there. He doesn't work in climate change research and he's not a politician.
A few sentences to finish, about a proper role for science in politics. It's bad enough that the important debate on global warming has become polarised (I can guess the comments this piece might attract). We can't let that polarisation extend to all scientific input (David Nutt might agree). We've got to a stage where it feels like the Left will automatically support the view of any formal group of scientists, almost to the extent of saying that the scientific view must become policy, without any modulation. We on the Right, of course, are doing what we were evolutionarily developed to do: we're reacting, in the formal definition of reaction. This will get us nowhere good. It is perfectly possible to think that Alan Johnson's sacking of David Nutt was a disgrace, without thinking that Professor Nutt's views should automatically become criminal policy. It is perfectly possible to think that Liam Donaldson oversteps his mark, without wanting to give free alcopops to five-year-olds. The Left isn't the voice of science, and we mustn't become its anti-voice.
So my suggestion to the next government: appoint scientists to provide informed evidence about matters of policy. Never, ever sack them for speaking as their integrity dictates they must. Similarly, do not ever give succour to the willfully irrational, or use their prejudice as a reason for ignoring any particular piece of scientific advice. Treat the electorate as adults, and don't worry too much about consistency along any single dimension (yes, alcohol is more damaging than ecstasy, but it need not follow that the penalties for access to either must be altered), and implement advice as and when it seems appropriate, in the wider context of the society we share.
I'm aware this proposal is hardly rocket-science; but then, I'm only a statistician.
