Download Lectures on Biostatistics (1971). Corrected and searchable version of Google books edition
Download review of Lectures on Biostatistics (THES, 1973).
honesty
|
I had never intended to write about climate. It is too far from the things I know about. But recent events have unleashed a Palin-esque torrent of comments from people who clearly know even less about it than I do. In any case, it provides a good context to think about trust in science, |
 Earthrise from moon. (click to enlarge) |
My interest in it, apart from little matters like the future of the planet, lies in the reputation of science and scientists.
I have been going on for years now about the lack of trust in science, and the extent to which it is a self-inflicted problem. The latest reactions to the developments at the University of East Anglia and the IPCC may show the nature of the problem with dreadful clarity,
Many of us came into science because, apart from the sheer beauty of nature, it seemed like one of the few honest ways of earning a living. Most scientists that I know still think like that, but recent
events invite some reexamination of honesty in science.
How dishonest is science?
The first thing to say is that I have never come across anything in my own field that would qualify as fraud, or even dishonest. I did once have a visit from a rather distressed postdoc (not in my area of work) who felt pressurised by her boss into putting an interpretation on her work that she did not agree with. In the end, the bit of work in question was left out of the paper. That could be held to be dishonest, in that the omission wasn’t mentioned, but it could also be held that the omitted result was too ambiguous to contribute much to the paper. It was just short of the point where I’d have felt compelled to do something about it. But only just. That is about the worst thing I’ve encountered in a lifetime.
There is, of course, an enormous difference between being wrong and being dishonest. Any research that is worth doing has an outcome that can’t be predicted before the work is done. At best, one can hope for an approximation to the truth. Mistakes in observations, analysis or interpretation will sometimmes mean the announced result is completely wrong, with no trace of dishonesty being involved. But when that happens, others soon fiind the mistake. It is that self-correcting characteristic of science that keeps it honest in the long run.
Of course there have been occasional cases of outright fraud, simple
falsification or fabrication of data. How often it occurs is not really known. There is a recent analysis in PLoS One, about verified cases of misconduct in the USA suggested that 1 in 100,000 scientists per year are to blame, but other ways of counting give larger numbers. For example, if asked around 2 in 100 scientists claim to be aware of misconduct by someone else., The numbers aren’t huge but they are much bigger than they should be.
It isn’t perhaps surprising that the Fanelli study found misconduct was most frequent in “medical (including clinical and pharmacological) research studies”, which are often funded by the pharmaceutical industry, Basic biomedical research and other subjects were better, though sadly that could be only because they are less often offered money.
What gives rise to dishonesty?
It seems obvious that one motive is money, as suggested by the worst rates of misconduct being found in the clinical pharmacological studies, It is well known that studies funded by industry are more likely to produce results that favour the product than those funded in other ways.
The other reason is presumably the human desire to win fame, promotion and to get grants.
It is no excuse, but it is perhaps a reason for misconduct that the pressure to publish and produce results is now enormous in academia. Even in good universities people are judged by the numbers (rather than the quality) of papers they produce and by what journal they happen to be published in. Contrary to public perception, even quite senior people have no guarantee that they can’t be fired, and life for postdoctoral fellows, who do a large fraction of experimental research, is harsh to the point of cruelty. They exist on a series of short term contracts, they work exceedigly hard and have poor prospects of getting a secure job. In conditions like that, the only surprising thing is that there is so little dishonesty.
The pressure to publish in particular journals is particularly invidious because it is known that the number of citations that a paper gets (itself a fallible measure of quality) is independent of the journal in which it appears. Bibliometrists are the curse of our age. (See, for example Challenging the tyranny of impact factors, 2003; and How to get good science, 2007 or its web version; and Peter Lawrence’s article, The mismeasurement of science)
The enormous competitive pressure under which academics work is imposed by vice-chancellors, research councils and other senior people who should know better, It is a self-inflicted wound.
In other words, the authorities provide a strong incentive to do poor, over-hurried and occasionally dishonest science. Perhaps the surprising thing in the circumstances is that there is so little outright fabrication. The very measures that have the aim of improving science actually have just the opposite effect. That is what happens when science is run by people who don’t do it.
For an idea of what life is like in science now, try Peter Lawrence’s Real Lives and White Lies in the Funding of Scientific Research. Or, for someone at the other end of their career, Jennifer Rohn’s account on Nature blogs.
Given the high degree of insecurity for young researchers, compounded by well-intentioned but vacuous “training” from daft Robert’s’ "training courses", or the dismaly ineffective Concordat, the only surprise is that so many people remain honest and devoted to good science. Nothing raises the ire of hard-pressed scientists more than the constant emails form HR trying to force people to go to gobbledygook courses on "wellbeing". Times Higher Education recently did a piece on "Get happy", The comments are worth reading.
So what about climate change?
Out of thousands of pages in the IPCC reports, a single mistake was found, On page 493 of the IPCC’s second 1000-page Working Group report on “Impacts, Adaptation and Vulnerability” (WGII) it was said that Himalayan glaciers were “very likely” to disappear by 2035. Glaciers are melting but that date can’t be justified. This single mistake has been blown out of all proportion. Furthermore it is important to notice that the mistake was found by scientists, not by ‘sceptics’. It is a good example of the self-correcting nature of science. Nevertheless this single mistake has provoked something close to hysteria among those who want to deny that something needs to be done.
On the other hand, the hacked emails from the University of East Anglia (UEA) look bad. It simply isn’t possible at the moment to say whether they are as bad as they seem at first sight, We just don’t know whether anything of importance was concealed, but we should know.
One thing can be said with certainty, and that is that the reaction to their revelation by Dr Phil Jones, and by the vice-chancellor of the University of East Anglia, was nothing short of disastrous. Fred Pearce put it very well in Climate emails cannot destroy proof that humans are warming the planet
Most unforgivably of all, UEA refused to comply with requests under the Freedom of Information Act, and there is some reason to think that relevant material was deleted. The deputy information commissioner, Graham Smith, said: in a statement that
“The emails which are now public reveal that Mr Holland’s requests under the Freedom of Information Act were not dealt with as they should have been under the legislation. Section 77 of the Freedom of Information Act makes it an offence for public authorities to act so as to prevent intentionally the disclosure of requested information.”.
That seems to me to be a matter that requires the resignation of the vice-chancellor. On this matter, I think George Monbiot is spot on in his article “Climate change email scandal shames the university and requires resignations“.
There was a big feature about academic freedom in Times Higher Education recently. One of the problems was what happens to someone who brings their own university into disrepute. But when that term is used, it is always used about junior partners in the organisation (you know, professors and the like). It should apply equally to heads of communications and vice-chancellors who bring their own university into disrepute, whether the disrepute is brought about by failing to comply with the Freedom of Information Act, or by promoting courses in junk medicine.
In general, conspiracy theories are wrong. I’m not sure how much of the distortion of climate data results from surreptitious funding of opposition to doing anything by the fossil fuel industry. The Royal Society is an organisation that is not usually prone to conspiritorialist views. That means one must take it seriously the fact that in 2006, the Royal Society wrote to ExxonnMobil to ask them to stop funding climate denialist organisations. This is a bit like the way Big Pharma has been caught funding “user groups” that endorse their products. Some newspapers like to stir up controversies about things that aren’t very controversial. For example there is a good analysis of a recent Sunday Times piece here.
Of course it is often alleged that "quackbusters" are funded by Big Pharma, though in fact the amounts of money involved are far too small for Big Pharma to bother. Climate deniers too like to suggest that there is some sort of conspiracy, arranged between hundreds of labs in the world to conceal the fact that there is no such thing as warming. I guess that shows only that deniers know little about how science works. it is an exceedingly competitive business, and getting hundreds of labs to say the same thing would be like trying to herd cats.
If there is a problem, it is the other way round. Labs are in such intense competition with each other, that it lcan lead to undesirable levels of secrecy.
Blogs in which researches have a direct dialogue with the public are a big help. As always in the blogosphere, the problem is to find the reliable sources. Two excellent sites, in which scientists (not journalists or lobbyists) talk directly with the public are realclimate.org and Andrew Russell’s blog. The post on RealClimate, IPCC errors: facts and spin, is especially worth reading.
Total openness is the only cure
All the raw data and all emails have to be disclosed openly. Everything should be put on the web as soon as possible. By appearing to go to ground, UEA has made enormous problems for itself and for the rest of the world. Some people object to total openness on the grounds that the other side tells lies. In the case of climate change (and in the case of junk medicine too) that is undoubtedly true. The opponents are ruthlessly dishonest about facts. The only way to counter that is by being ruthlessly and visibly honest about what you know, and why.
The UK’s Meteorological Office has, to its great credit, put raw data on line. That policy has already paid off, because a science blogger found a mistake in the way that some Australian data had been incorporated into forecasts. The Met Office thanked him and corrected the mistake. In fact the error makes no substantial difference to the warming trend, but the principle is just great. The more people who can check analyses and eliminate slip-ups the better.
Putting raw data on the web is an idea that has been gathering force for a while, in all areas of work, not just climate change. In my own are (stochastic properties of single ion channel molecules) our analysis programs have always been available on the web, free to anyone who wants them, despite the large amount of work that has gone into them. And we run a course. almost free, on the theory that underlies our analyses. Within the last couple of months we have been discussion ways of making public all our raw data (in any case, we would always have sent it to anyone who asked). Digitised single channel records are big files (around 100 Mb) and it is only recently that the web has been able to deal with such large amounts of raw data. There are also problems of how to format data so other people can read it, The way we are all heading is clear, and the fact that some people in climate science appeared to be hiding raw data is a disgrace.
Public relations is not the cure
,
It is not uncommon to read that science needs better PR. That is precisely what is not needed. PR exists to put only one side of the story. That makes it an essentially dishonest occupation. Its aims are the very opposite of those of science. The public aren’t stupid: often they recognise when they are getting half the story.
It is particularly unfortunate that many universities have developed departments with names like "corporate communications". Externally they are seen as giving information about science, and indeed some of the things they do are successful public engagement in science. Only too often, though, it is made clear internally that an important aim of these departments is to improve the image of the university.
But you have to choose. You can engage the public in science or you can be a PR image-builder. You can’t be both.
The matter came to a head in 2008 when, according to a report in Times Higher Education, the University of Nottingham issued a memo that defined public engagement as: “The range of activities of which the primary functions are to raise awareness of the university’s capabilities, expertise and profile to those not already engaged with the institution”.
The mainstream media and political blogs
The biggest problem of all with climate change is that it has become more about politics than about facts. It has become an essential credential for any conservative to deny that climate is changing. It is part of their public image, and most conservatives neither know nor care about evidence. Like Sarah Palin, they just know. In the USA especially, the argument is not really about climate at all. It is really about discrediting Barack Obama -a sort of swift-boat treatment that uses whatever lies are needed.
Just as with the great MMR fiasco and the promotion of its false link to autism, reports in newspapers and blogs must bear much of the blame for failing to inform readers of the actual underlying facts and, just as important, the uncertainties. Of course some papers have done a pretty good job, particularly the Guardian and the Independent in the UK, and the New York Times.. The political blogs, by and large, haven’t. The Huffington Post has made little effort (and publishes some appalling nonsense about medicine too).
The problem with political blogs and tabloid newspapers is that they are much more interested in sensation and circulation than they are in giving accurate news and information. Take, for example, the Guido Fawkes blog. To be fair, the blog itself says "The primary motivation for the creation of the blog was purely to make mischief at the expense of politicians and for the author’s own self-gratification. Its writer", so you know not to expect much, Paul Staines, was at the Westminster Skeptics event, Does Political Blogging Make a Difference? He makes no pretence of taking the news seriously, which, I guess, is why I don’t read his blog. After the talks I asked why his blog did little about climate change. His answer was "where are your sandals?". On the way home I tweeted, from a very overcrowded train (most trains from Euston being cancelled that night),
"On way home from #sitp political blogging. Learned that Guido serious about nothing but Guido. Narcissist not journalist."
At least one other person there agreed (thanks, Dave Cole).
It was good to hear Sunny Hundall of Liberal Conspiracy (the only one I read), but I found myself agreeing mostly with the chair, Nick Cohen. It would be a tragedy if the great national and local papers were to vanish. Guido Fawkes and Huffington Post are not remotely like proper newspapers.
Specialist blogs like this one are fine if you are interested in the topics we write about, but we don’t begin to supplant proper newspapers. Bloggers can and do occasionally get good stories. Those that are written by scientists can analyse more critically than most journalists have either the knowledge or the time to do. Bur they don’t come close to supplanting the detailed reporting in good newspapers of local events, what happens in law courts or in parliament. That’s why it is vital to buy newspapers, not just read them free on the web.
Follow-up
James Hayton, who is in nanoscience has posted his thoughts obout trust in science on his blog. I discovered this via Twitter (@James_Hayton). He also posted a beautiful clip from the Ascent of Man, in which Jacob Bronowski speaks, from Auschwitz, of the consequences of irrational dogma. I’m old enough to remember Bronowski on a 1950s radio programme, the Brains Trust, though James Hayton clearly isn’t. Now I enjoy equally his daughter, Lisa Jardine‘s talks about science and history.
1 March 2010. Phil Jones, and the vice-chancellor of the University of East Anglia, appeared before a parliamentary committee. I found their responses to questions very disappointing. The evidence submitted by the Institute of Physics was strongly worded, but spot on.
“The CRU e-mails as published on the internet provide prima facie evidence of determined and co-ordinated refusals to comply with honourable scientific traditions and freedom of information law. The principle that scientists should be willing to expose their ideas and results to independent testing and replication by others, which requires the open exchange of data, procedures and materials, is vital.”
7 March 2010. Thanks to some kind remarks from Michael Kenward (see first comment). I sought wider coverage of this item in the mainstream media. Consequently, on Thursday 4 March, a much shortened version of this article appeared on the Guardian environment site. That piece has accumulated so far, 230 comments. The discussion of it has spread to the two blogs that I recommended, Andy Russell’s blog and RealClimate.org, though it has been diverted onto the side-issue of the letter from the Institute of Physics. The seemingly innocent idea that total openness would increase trust has, to my real astonishment, resulted in hysterical accusations that I’m a crypto-denialist. The constant politically-motivated attacks on climate science seem to have induced a paranoid siege mentality in some of them. There is a real danger that such people will harm their own cause, and that would be tragic.
Being interested in science communication, I was pleased when the BMJ asked me to review Unscientific America , by Chris Monney and Sheril Kirshenbaum.
The BMJ provides a link that allows you access to the whole review. They have made very few changes from the submitted version, which is reproduced below (with live links in the text. [Download pdf of print version]
I very soon discovered that the book had already caused ructions in the USA, as a result of its advocacy of appeasement of religious groups. In particular there was all out war with P.Z.Myers, whose very popular blog, Pharyngula. documented the battle in detail).
It is an American book through and through, and in the USA the biggest threat to reason comes from the far-right religious fundamentalists who preach young-earth creationism. It is said that 46% of US citizens believe that the earth is less than 10,000 years old. The same far-right religious groups also preach that carrying guns is good, that Iraq was responsible for 9/11, that climate change is a socialist conspiracy and that health care for everyone is a communist plot. And they never hesitate to lie in the promotion of their ‘religious’ views. The US situation is totally different from that in Europe, where religion is all but dead, and young earth creationism is the preserve of a few cranky used-car dealers (and possibly Tony Blair?)
|
Review of the Week Trust me, I’m a scientist David Colquhoun, professor of pharmacology, University College London Unscientific America sounds like a fascinating topic, not least because the book is a follow-up from Mooney’s The Republican War on Science. It is written entirely from a US perspective (the USA sequenced the genome and invented the internet, apparently unaided). It’s reported that 46 percent of Americans believe that the earth is less than 10,000 years old. That’s certainly cause for alarm and Mooney & Kirshenbaum are certainly alarmed. They think that the public needs to be educated in science. They identify the obvious problems, evolution, climate change and quackery and ask what can be done. The problem is that they propose no good solutions, and some bad ones. Their aims are worthy but sometimes the book reads like an over-long and somewhat condescending whine about why science and scientists are not sufficiently appreciated. I simply don’t think that it’s true that the public are not interested in science, nor that they can’t understand it at a level that is sufficient to be useful. It’s true that they have been let down badly by some sections of the media. Think particularly of the “great MMR hoax”1. The disastrous fall in vaccination is more attributable to talk show presenters and air-headed celebrities than to lack of interest from the public. People are systematically deceived by anti-vaxers, climate change denialists, vitamin pill salesmen and a horde of crackpot alternative therapists. There is one problem that Mooney & Kirshenbaum don’t talk about at all, yet it seems to me to be one of the biggest problems in science communication. It isn’t lack of interest by the public, nor even lack of understanding, but lack of trust. The tendency of real science to indulge in hyperbolic self-promotion is one reason for the lack of trust. Sometimes this descends into outright dishonesty2,3. That is a tendency that is promoted by government and funding agencies by their insistence on imposing silly performance measures. The public is quite sensible enough to take with a pinch of salt the almost daily announcements of “cancer cures” that emanate from university press offices. On the face of it, one should be encouraged that ‘public engagement in science’ is the mode du jour. It isn’t quite that simple though. Only too often, universities regard public engagement as a branch of their own PR machine4. They even instruct you about what tone of voice to use when talking publicly. One reason why scientists need to talk to people outside the lab is precisely to counteract this tide of nonsense from PR people, who are paid to deceive. The problem for academics is usually time. We already do three jobs, teaching, research and coping with HR bollocks. How can we find time for a fourth job? That’s not easy, especially for the best researchers (those that do research themselves, not just lead a team). Mooney & Kirshenbaum suggest that the solution is to create a “cadre of communication and outreach experts”. I don’t think this would work. They would, by and large, be outsiders, writing uncritical paeans, dictated by big name scientists. A new cadre of PR hangers-on does not sound like a great idea. A better, and very much cheaper, solution would be to provide a course in free blogging software and we’ll do it ourselves. The two chapters that I looked forward to reading, on religion and on “The bloggers cannot save us”, proved deeply disappointing. The authors are firmly in the camp of what Richard Dawkins called the “Neville Chamberlain school of evolutionists&rdquo.; They maintain “if the goal is to create an America more friendly to science and reason, the combativeness of the New Atheists is strongly counterproductive”. They are particularly critical of P.Z. Myers5, the University of Minnesota developmental biologist who is splendidly clear in his views. Of the communion wafer, he famously said “It’s a frackin’ cracker”. But he, and Dawkins, are right. When it comes to young earth creationists we have a war on our hands, and nowhere more than in the USA. What’s more it’s a winnable war. Mooney & Kirshenbaum are all for appeasement, but appeasement won’t work. It might please the more moderate wings of the church, but they already believe in evolution and are regarded by fundamentalists as being just as big an enemy as Myers and Dawkins. And, one must ask, who has done best at getting a wide public readership? P.Z Myers’ blog, Pharyngula, has up to two million page views a month. Dawkins’ book The God Delusion has sold three million copies. In comparison the bland and often rather condescending corporate science web sites get tiny numbers of hits. In Europe in general, and the UK in particular, young earth creationists are not the major problem that they are in the USA, despite being supported by Tony Blair6. Perhaps the nearest analogy in Europe is the threat to reason from various sorts of crackpot medicine. The appeasers are widespread. The Royal Colleges and the Department of Health are at the forefront of the Neville Chamberlain approach. But appeasement hasn’t worked there either. What has worked is the revelation that university courses are teaching that “amethysts emit high yin energy”7. Or, in a lecture on herbal approaches for patients with cancer, “Legally, you cannot claim to cure cancer. This is not a problem because: ‘we treat people, not diseases’ “8. This is shocking stuff but it has not been unearthed by the corporate media, but by bloggers. I think Mooney and Kirshenbaum have it all wrong. They favour corporate communications, which are written by people outside science and which easily become mere PR machines for individuals and institutions. Such blogs are rarely popular and at their worst they threaten the honesty of science. More and more individual scientists have found that they can write their own blog. It costs next to nothing and you can say what you think. A few clicks and the world can read what you have to say. Forget corporate communications. Just do it yourself. It’s fun. And think of the money you’d save for doing science if the PR people were just fired. (1) Goldacre, B. The media’s MMR hoax. 2008 http://www.badscience.net/2008/08/the-medias-mmr-hoax/ (2) PLoS One. Ghostwriting documents now fully available on PLoS Medicine website. 21-8-2009 http://speakingofmedicine.plos.org/2009/08/21/ghostwriting-documents-now-fully-available-on-plos-medicine-website/ (3) Colquhoun, D. Universities Inc. in the UK. The Corporate Corruption of Higher Education: part 2. 6-12-2007 https://www.dcscience.net/?p=193 (4) Corbyn, Z. Nottingham raises eyebrows over definition of ‘public engagement’. 21-8-2008 http://www.timeshighereducation.co.uk/story.asp?sectioncode=26&storycode=403234 (5) Myers, P. Z. Pharyngula. 2009 http://scienceblogs.com/pharyngula/ (6) Pyke, N. Revealed: Blair’s link to schools that take the Creation literally (Independent 13 June 2004). 13-7-2004 http://www.independent.co.uk/news/uk/politics/revealed-blairs-link-to-schools-that-take-the-creation-literally-732032.html (7) Colquhoun, D. Westminster University BSc: "amethysts emit high yin energy". 23-4-2008 https://www.dcscience.net/?p=227 (8) Colquhoun, D. Herbal approaches for patients with cancer. 10-8-2009 https://www.dcscience.net/?p=2043 |
Follow-up
P.Z. Myers has posted about thie book review, on Pharyngula, as Is this book dead yet? There are a lot more comments there than here, though few of them address the question of science communications..
Butterflies and Wheels is generating a lot of hits