I was all set to post as previously threatened another screed about golden rice in the wake of my spat on Steve Savage’s website with some of his commenters, when all of a sudden Steve releases a new post on the somewhat related issue of scientific evidence, which is perhaps of more general interest. So I think I’ll hold off for now on the golden rice and go with the science/evidence theme. In other news, I’ve been tangling with the former poet laureate on the Guardian letters page and with proponents of the pig swill ban among other things over at the Food Climate Research Network. Goodness, am I really that argumentative? Probably, alas. What a good thing I’m confined to this little window in the blogosphere (click x, top right).
Anyway, Steve’s argument is that science is a conversation which only begins with publication in a peer-reviewed journal, and that the system is hijacked when scientists aggressively move their findings into the mainstream public conversation before the scientific conversation has reached a consensus. The basic lines of his argument are hard to fault, I think, except that the tendency for scientists to grandstand their conclusions for personal or political reasons is hardly new (think Edison vs Tesla), and ‘scientific consensus’ can often be an elusive destination. But the funny (actually, quite predictable…) thing is that all Steve’s examples of this deplorable practice are ones that have emphasised the negative effects of the mainstream food and farming system he champions. For many of us more sceptical of this system than Steve, the deplorable practice runs at least as much in the opposite direction, as for example in aggressively favourable public prejudging of golden rice by folks that Steve happily links from his blog.
Part of the problem, I think, is that because science has been so successful at unteasing causalities and informing technological developments we invest unreasonable expectations in it to arbitrate between different views of how the world should be which are ultimately rooted in politics and philosophy and which therefore cannot be resolved by scientific experiments. Steve wants there to be scientific conversations, but he doesn’t want Séralini’s study linking GM maize to cancer in rats to be published in a peer-reviewed journal, even though it’s apparently made it through the peer review process twice and was retracted in circumstances that were opaque at best.
That doesn’t seem very conversational of him – surely it’s better for these things to be available in the publicly-accredited scientific domain so that the conversation can truly begin. No doubt the study is flawed – almost all studies are flawed somehow or other. But the Séralini affair and others of its ilk does make me wonder whether there’s some publication bias going on in the world of GM research. If one or a few studies suggest a link between a GM crop and disease, it doesn’t mean that the case against GM crops is closed. A single study rarely proves anything. But you might expect to find the odd study in the scientific literature linking a GM crop to a negative health outcome of some sort even if only on the grounds of simple probability – the fact that there seem to be none (and the fact that those like Séralini and Pusztai who’ve attempted to suggest one have been so relentlessly hounded in ways quite alien to disputes in less politicised scientific arenas) is to me suspiciously redolent of publication bias, or worse. And not just to me – a study published in Environmental Sciences Europe argues that there have been ‘critical double standards’ in the evaluation of Séralini’s study as compared to the feeding studies conducted by Monsanto on their maize.
From publication bias to confirmation bias – one accusation among several levelled at me on Steve’s site by David Röll. My exchanges with Röll have led me to think that he’s basically a wind up merchant and I’m probably taking his comments way more seriously than I should, but hey let’s try to derive something useful from his windy rhetoric. So I’ll admit it, yes, I suffer from confirmation bias. And so, manifestly, does Steve Savage. And everyone else, surely. We all come to particular views over a period of time as a result of various direct and indirect influences and experiences, but the world’s complexity generally exceeds the neat lines with which we seek to organise it. When we encounter scientific research that appears confirmatory of our worldviews we latch on to it gleefully, again I’d argue in part because of the somewhat excessive cachet of science-as-truth in our culture. And when, inevitably, we encounter plausible research that challenges aspects of our worldviews, we look for flaws and rationalisations. And why not – that’s surely all part of ‘the conversation’. Nobody abandons a slowly accreted worldview overnight. Though hopefully addressing its contradictions and contrary evidence allows us to get more nuanced in our understandings.
The Berkeley physicist and philosopher Thomas Kuhn built an influential theory about the history of science around the notion that confirmation bias is part and parcel of the scientific process – a philosophy that can be summed up by the old cliché that you can determine the eminence of a scientist by the length of time they obstruct progress in their discipline. But the great thing about science – almost uniquely among human endeavour – is that its procedures ultimately enable it to overcome confirmation bias and the passing opinions of influential savants. As someone trained in social science rather than natural science, the misery of my discipline is that we just don’t have the same procedures available for escaping ideological blinders. On the other hand, the joy of it is that – economists aside – for the same reason we’re not so prey to the hubris of supposing that our convictions exist above the messy world of politics and argument, issuing instead like some fount of sweet water from the uncorrupted well of pure knowledge. Which is why I consider misguided the shrill appeals to ‘reason’, ‘science’ and ‘logic’ for deciding in favour of agribusiness-as-usual as a solution to contemporary problems promulgated by the likes of Graham Strouts and David Röll and, albeit less aggressively and more informatively, Steve Savage (though that’s not to say that there’s no role for these qualities in addressing such problems).
Science can overcome confirmation bias, but the process of this overcoming is neither fast nor simple. What particularly worries me is the apparently growing use of the label ‘junk science’ to summarily dismiss from consideration research or analysis that isn’t consonant with the supposed consensus asserted by the person deploying the term – the surest way for science to forget the radical questioning that gives it its edge over other modes of thought and to become just another church intent on dispatching the heretics. George Monbiot has shown how the junk science label arose out of corporate efforts to deny the scientific evidence on the consequences of tobacco and, more recently, on climate change. On a much smaller stage, the way that David Röll sought to dispatch my scepticism over golden rice was cut from the same cloth – it’s so much easier to dismiss your opponent for junk science, Gish gallop, conspiracy theory or whatever than actually engage with their arguments.
Well, there are those I’ve accused of Gish gallop myself – time is pressing, and why work through a foot-thick tissue of questionable assumptions and dodgy evidence (especially when the person concerned is only likely to respond with ad hominem abuse). But if you don’t engage with the specific arguments, it opens the door to your own confirmation bias and certainly gives you no right to consider your case proven. In many situations, science does not speak with one voice, and cases of outsider science becoming mainstream are legion. Much as I despair when someone says exactly this in justification of earth vibrational essences, perpetual motion machines or other nonsense (I’m thinking more of things like plate tectonics, the Alvarez hypothesis and the symbiosis in the eukaryotic cell), that fact remains. And in any case, all this talk of ‘the science’ in relation to essentially political commitments on food, farming and society is misleading – the advantages and disadvantages of different farming futures do not only or even mainly lie in what ‘the science’ tells us, but in what kind of social worlds we wish to inhabit. Which was pretty much my position in the FCRN debate, and is a recurrent theme on this blog. So thanks for coming back for more.