IJE Advance Access originally published online on July 13, 2009
International Journal of Epidemiology 2009 38(4):908-910; doi:10.1093/ije/dyp246
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Commentary: A continued affair with science and judgements
UCL Epidemiology & Public Health, 10–19 Torrington Place, London WC1E 7HB, UK. E-mail: m.marmot@ucl.ac.uk
Accepted 3 June 2009
One Friday evening in 1946, Karl Popper was invited to address the ‘Moral Science Club’, a group of philosophers who met regularly in the room of one of the senior fellows of King's College Cambridge. Among the audience was the now legendary Ludwig Wittgenstein, professor of philosophy at Cambridge University, who is credited with arguing that there are no real philosophical problems only linguistic muddles. Popper had quite different views on philosophy. There were indeed real problems to be sorted out: communism, psychoanalysis and the nature of scientific enquiry.
Accounts of what actually happened next vary. During Popper's seminar, Wittgenstein was said to have become so incensed that he seized the poker from the fire place and started gesticulating with it. Bertrand Russell leapt to his feet and told Wittgenstein to put down the poker. ‘You misunderstand me, Russell’, said Wittgenstein. ‘You always misunderstand me’. Russell's response: ‘You’re mixing things up, Wittgenstein. You always mix things up’. One account has it that Wittgenstein, in anger, challenged Popper to give him one example of a real moral precept. Popper claimed to have replied: ‘Not to threaten visiting lecturers with a poker’. Wittgenstein dropped the poker with a clatter and stormed out.1
If the pen is mightier than the sword, are pokers somewhere in between? My 1976 paper, ‘Facts, Opinions and Affaires du Coeur’,2 reprinted here, eavesdropped on philosophers disagreeing. Such disagreement is hardly surprising, you might argue, philosophers’ ideas do not have to test themselves out in the harsh world of facts. But, if you think that science is not like that, you’ve missed the sight of some of our scientific colleagues in full flight, or fight, but I’ll come to that.
What about lawyers? On the US Supreme Court, regularly, Chief Justice Roberts and Justices Scalia and Thomas vote the same way. They are the conservatives on the bench. The liberals commonly vote differently. All nine Supreme Court Justices have one Constitution, one set of facts and one set of rules of logic, yet consistently they reach different conclusions, closely linked with their prior political positions. (Such disagreement confirms my view that a law court is not a good place to settle issues of whether exposure x causes disease y. The issue of blame should be separated from establishing if there is an exposure–disease relation.)
Philosophers, lawyers, we can add economists to this list of experts who disagree among themselves. Examples abound. Nicholas Stern, a distinguished economist, published the Stern Review on Climate Change. He argued, on economic grounds, that we have to invest now in mitigation and adaptation to avert catastrophic costs in the future.3 In Britain, he was widely praised, elevated to the House of Lords, given honorary degrees and was in great demand in Europe. In the USA, his report almost passed without notice, and distinguished academic economists disputed his analysis.4 The debate turned on the arcane issue of the discount rate. The American economists argued that market behaviour tells us that people value present consumption over future consumption. One hundred dollars now is more valuable than $100 in a generation's time. We therefore need to discount future costs and benefits. Democracy (i.e. people's behaviour in the market) demands it. (I may not have that quite right, but I am not good on theology.) Stern used a discount rate of near zero. He argued that it was unethical to condemn generations yet unborn to the effects of climate change just because we, today, value present consumption more than future benefits to the planet and the costs of dealing with them. Do we really not care about our grandchildren, at least not more than present value discounted at the market rate of interest?
I have little doubt that Stern and his American critics can agree on how to calculate ordered probits, and do ordinary least squares regressions, with country dummies, and they have the same sets of data available to them, yet come up with radically different conclusions. An academic disagreement among economists could decide the future of the planet.
Do you remember the debates about genes and IQ? They are still vibrant in behaviour genetics. The debate was particularly virulent in relation to racial differences in intelligence.5 It was a fair bet that the side of the scientific issue with which people aligned themselves related to their political beliefs. The ‘right’ tended to believe that racial differences were human nature, i.e. genetic; and the ‘left’ that they were social and economic in origin. Stephen Jay Gould was so strongly against the ‘nature’ position that he wrote a whole book attacking factor analysis—the basis for coming up with ‘g’ as a measure for general intelligence.6
Recently, I proposed a screening test to detect an economist: if someone comes across the social gradient in health and assumes that health leads to socio-economic position, rather than social circumstances lead to health, then he is an economist. Like all screening tests, there are false positives and false negatives. But the typical economist starting position is that peoples’ health determines what happens to them. The public health starting position is that what happens to people affects their health. When economists analyse birth cohort data, they find evidence that health in childhood influences adult socio-economic position. When public health people analyse the same data they find evidence that childhood socio-economic circumstances influence adult health. Each concludes that their pathway is more important.
This is not just a polite, or even testy, academic debate. The policy implication of these two positions is quite different. If the main causal direction is from health to wealth, the appropriate intervention is to control illness in order to improve an individual's social and economic fortunes or, indeed, eradicate illness to improve the economy of a whole country. If, as I conclude, the main causes of health inequalities reside in the circumstances in which people are born, grow, live, work and age—the social determinants of health—then action to reduce health inequalities must be in those circumstances and the fundamental drivers of those circumstances: economics, social policies and governance.7
We can add social scientists and epidemiologists, then, to our list of disciplines that disagree despite apparently following similar rules of logic and evidence. And politicians ...
As we watched the grisly spectacle of a rush to war in Iraq, the legal justification for which was supposed to be Saddam's weapons of mass destruction, I thought that the politicians were engaged in an extreme form of the intellectual manoeuvres to which scientists are sometimes prone. When the weapons inspectors in Iraq consistently failed to come up with evidence for the existence of weapons of mass destruction, the politicians did not conclude that there were no such weapons. Rather, they said that Saddam had hidden them. The auxiliary hypothesis ‘saved’ the theory. Evidence did not appear to decide it. (Actually, it was worse: lack of evidence was taken as positive evidence that they did exist provided one accepted the hiding hypothesis.) It was precisely this type of observation: that scientists appear to engage in such theory-saving tactics—never readily discard a pet theory just because some data do not fit—that prompted me to write ‘Facts, Opinions and Affaires du Coeur’.
Peter Medawar, in one of his books, tells the story of being taken down a narrow street in Naples—so narrow that the balconies from the houses on opposite sides of the street nearly touch. Two women, on opposing balconies, are haranguing each other. ‘It's always like that’ says the guide, ‘they’re arguing from different premises’.
That surely is part of the answer. Rereading my 1976 piece I would still agree with my observation that people take theoretical positions for a whole variety of reasons, ‘in addition to what the data support’. The data are most certainly relevant, but are by no means the whole story. Prior theoretical positions shape our view of the evidence. These prior positions may have to do with what we wrote in our last paper, our world view, the flash of insight in the bath tub, political positions or professional training.
Naively, I thought the philosophers of science would solve the question of how to resolve scientific disagreements for us. They don’t. That said, I still think Kuhn had an important insight and Lakatos was a relatively unsung hero. But I do not think there is a clearer standard of choosing between competing hypotheses than there was in 1976. To be sure, within epidemiology we now have systematic reviews, and a great deal of thought has gone into revising and improving methods. Periodically, someone will conduct a burial for case–control studies, or observational epidemiology in general, or will celebrate the necessity of controlled trials for everything (Newton? or Darwin?), or will find the answer in genetic epidemiology (what's the question?). This has not made the arguments go away. Because, in the end, the arguments are not principally about methods.
This became clearer to me when I revisited the issues raised in ‘Facts, Opinions and Affaires du Coeur’ a decade later.8 At a dinner in a Cambridge college, I was introduced to a neurosurgeon whose opening gambit was: ‘You’re an epidemiologist. I don’t believe smoking causes lung cancer’.
I countered by asking if it would be helpful if I told him of some of the evidence that should lead to the conclusion that smoking causes lung cancer and he should tell me why he did not find that evidence compelling. I laid out the evidence but this proved not to be helpful. Neither of us shifted from our position. He then threw in, as if it were relevant, that, as a neurosurgeon, he did not think that motorcyclists should be forced to wear crash helmets.
Of course, I thought that I was a model of sweet reason and that he was a right-wing hyena whose libertarian views were clouding his sight of the evidence. He, no doubt, thought that I was one of those typical public health/social engineers who, in a misplaced attempt to fiddle with society and improve things, would over-interpret any scrap of data and use it as evidence; and that he was the last defender of liberty.
All of this is rather important. Resolution of the scientific debates has profound policy implications. It is, I suggest, precisely these policy implications that account for some of the apparent scientific disagreements. To illustrate, I chaired the World Cancer Research Fund Panel of Diet and Cancer.9 After several years of work, when we published our sober recommendations, we were labelled in the British national press as ‘health Nazis’. ‘Fat lies of the food fascists’, screamed one headline. The vigour of this negative reaction was a bit surprising. Why Nazis? And, more particularly, in the present context, why ‘lies’?
The review on diet and cancer was painstaking. An international body of experts deliberated on the results of systematic literature reviews commissioned from a number of leading scholars in universities. After years of work, we published careful conclusions on what the evidence showed, and cautious recommendations that followed: stay lean, avoid processed meat, limit red meat, increase fruit and vegetables. In the original report, we said nothing about who should do what with our recommendations.
I think the epithets, Nazis and fascists, apart from being insulting, are lazy, as is ‘lies’. In the fevered intelligence of these commentators, conclusions such as ours might lead to the state interfering with people's freedom to kill themselves with unhealthy food. Rather than debate the actual evidence, it is easier to dismiss us as fascists and liars. It saves the trouble of seeing whether they might actually agree with our careful reading of the scientific evidence.
None of this reflection in any way dents my respect for the evidence. When we published the report of the Commission on Social Determinants of Health6 one government, apparently, labelled our report as ‘ideology with evidence’. It was meant as an insult. I took it as a compliment. We were, on the Commission, quite explicit about our ideology. We said: ‘social injustice is killing on a grand scale’. The current way we run our affairs is killing people and is quite wrong. If that is ideology I am unrepentant.
The next important question is what to do about it. A lot of useless, and even harmful, things have been done in the name of good intentions. Here the evidence matters. If we are not frequently surprised that things we thought to be true turn out not to be so, or that changes that we thought would improve health equity turn out not to, then it is unlikely that we are doing our jobs properly. We need, continually, to be seeking the best evidence that we can. But let us recognize the ideological debates for what they are.
Conflict of interest: None declared.
 |
References |
|---|
 Top References |
|---|
1 Edmonds D, Eidinow J. Wittgenstein's Poker: The story of a ten minute argument between two great philosophers (2001) London: Faber and Faber.
2 Marmot M. Facts, opinions and affaires du Coeur. In: Am J Epidemiol (1976) 103–519. Reprinted Int J Epidemiol; doi:10.1093.ije/dyp234.
3 Stern N. The Stern Review: The Economics of Climate Change (2006) London: HM Treasury.
4 Deaton A. Letter from America: on transatlantic vices, or Stern in America. R. Econ. Soc. Newsl (2007) 130:3–4.
5 Hernstein RJ, Murray C. The Bell Curve: Intelligence and Class Structure in American Life (1994) 1st. USA: Simon & Shuster Inc.
6 Gould SJ. The Mismeasure of Man (1981) New York: W.W.Norton.
7 Commission on Social Determinants of Health. Closing the gap in a generation: health equity through action on the social determinants of health. In: Final Report of the Commission on Social Determinants of Health (2008) Geneva: World Health Organization.
8 Marmot MG. Epidemiology and the art of the soluble. Lancet (1986) 327:897–900.[CrossRef]
9 Panel on Food, Nutrition Physical Activity, & the Prevention of Cancer. Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective (2007) London: WCRF.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ebrahim Ideology with evidence: global warming, maps and ethics Int. J. Epidemiol., August 1, 2009; 38(4): 895 - 896. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
