26 March 2012
ByAppeared in BioNews 650
Newspapers overcook mediocre science all the time. You know this, I know this. All the same, a story in the Sunday Times a couple of weeks ago (1) deserves special mention for its utterly uncritical reporting.
The title smells sweetly of hogwash, but I'll let that slide - slapping bombastic headlines on sensible copy is pretty much what subeditors are paid to do.
And, in fact, 'Found: Genes That Make Brits Free-thinkers' is an understatement. The research the article relays (2) didn't identify several genes, working across complex networks, to make Britons the maverick renaissance men and women that we undoubtedly are, but a single gene.
As for the article under the heading, it contains no quotes from interviews with any scientists. Instead, it presents a sentence from the relevant paper as a quote from the lead researcher - Dr Joan Chiao of Northwestern University in the USA. We are told that the study's 'findings will need further confirmation'. We are not told that such scrutiny has already been applied and that the paper received a response, published in a later issue of the same journal (3). We are not told that this response concluded that the study's findings are 'not supported by current data'. We are not told that the research was published two-and-a-half years ago. These facts can all be ascertained by a five-minute web search.
Why such scornfulness when scientific fluff is hacked up daily by journalists the world over in search of copy? Because the research in question isn't the kind of inconsequential nonsense that can be blown around and nobody gets hurt. It is rather a contemporary contribution to an old and dangerous myth that has lain low in science recently, but now may be in the ascendant. Of course, you wouldn't think that from the perky headline.
The study, the Sunday Times relates, 'suggests that the individualism seen in western nations, and the higher levels of collectivism and family loyalty found in Asian cultures, are caused by differences in the prevalence of particular genes'.
As I say, the paper actually posits variations in a single gene as the root cause of the East-West cultural divide, but enough criticism of the write-up. One merit that the Sunday Times' account does have is that it reflects the mistaken assumptions of its source material very well.
To start with there's the placement of individualism and collectivism, together with family loyalty, in binary opposition. Unless you accept the cliché of there being no 'I' in 'team' as a sociological law, this is clearly nonsense. Then there's the related assumption that because Western societies are, on the whole, less family-oriented they must be more individualistic, that's also very dubious unless individualism is taken to mean something close to what the sociologist Emile Durkheim described when he spoke of 'the disorganised dust of individuals' emerging with modern capitalism at the turn of the last century (4).
But I'm getting ahead of myself and the paper itself addresses such concerns in only the most facile manner. Essentially, the study examines the influence of a gene which codes for a protein regulating serotonin in the brain. Serotonin is a neurotransmitter - a chemical which brain cells use to communicate with each other - and is particularly implicated in mood control. The antidepressant Prozac (fluoxetine) acts on the gene's protein product, the serotonin transporter.
The gene coding for this transporter exists in several variants and these mutations probably affect the transporter's function. In particular, the gene contains a region that can exist as a short (S) or long (L) allele. Studies suggest that which version you carry may have a subtle influence on mood and behaviour.
The paper's main claim is that collectivistic values flourish among populations where the S allele predominates whereas individualism is associated with the L allele. The centrepiece is a graph on which a measure of individualism-collectivism (vertical axis) is plotted against S allele prevalence (horizontal) for 29 nations (5).
You might think that quantifying a false dichotomy as abstract as 'collectivistic versus individualistic values' would be impossible. Not so, apparently. The researchers use an index based on 'employee value scores' conducted around the world by IBM between 1967 and 1973 (6). That's obviously beyond reproach as a dataset.
Most of the data plots on the resulting graph run in a more or less straight line up the vertical axis. In other words these national populations show considerable variability in collectivism-individualism without any impact of the S allele. There are, however, four tightly grouped countries just off this line in the top right. These are China, Singapore, Taiwan and Korea which all have both high collectivistic values and a predominance of the S allele in their populations.
Bearing in mind the vertical line that contains the other data plots, this looks like a coincidence to me. Then again, I have only a rudimentary understanding of statistics. I was therefore unable to follow the method that enabled the researchers to show a significant predictive relationship.
However, the quantum chemist Dr Istvan Mayer is less of a slouch with stats than I am. By chance, the paper fell into his hands and he decided to check the correlation out and found it lacking. His plain-English commentary, published online, met with a predictably obfuscating response from the authors (7).
Now the fun really begins. With their correlation in the bag, the researchers have to show some degree of causation. This is a real poser because the S allele tends to be associated with depression and anxiety - not the kinds of mental states normally associated with people with a tendency to cohesively bound family structures.
But, hang on - East Asian countries enjoy lower levels of diagnosis of these mental disorders. Armed with this knowledge the researchers construct the 'culture-gene co-evolution hypothesis' of human behaviour - a fabulously circular and occasionally impenetrable theory vainly described in the text as 'parsimonious'. It says that East Asian cultures have structured themselves collectively to counter the negative effects of the S allele. They have done this very well, which accounts for the lower rates of diagnosed mental disorders.
This is not a hypothesis. It is a 'Just-So' story concocted to fit known data. But by this point you will not be surprised to learn that the study ekes out another significant correlation; this time between S allele frequency and low mood disorder diagnoses.
Taken as a whole, the study is a groaning smorgasbord of bad science. All the classic ingredients are present: arbitrary quantification of reified concepts; confusing statistics; correlation assumed as causation; attention to only a few of myriad confounding variables; conflation of phenomena at the individual and population levels.
Such low scientific calibre is enough to make a transfer onto the pages of the national press regrettable, but that isn't what got my hackles up. Nor is the study's reductio ad absurdum of the gene-centred view of human evolution where the sweep of human history is explained as a few genes pulling the strings on human puppets. No, it's the contribution to the dangerous myth that really irks.
The myth has a name that has been carefully omitted from the paper. The Sunday Times, though, nails it in the first sentence. Also further down, like this: 'If confirmed, the findings would suggest that races may have a number of inherent psychological differences - just as they differ in physique and appearance'.
'Race' is a word that has been more or less banished from science, but that doesn't mean the myth of its existence isn't alive and well. Or even, on the evidence of Dr Chiao's paper, flourishing. The problem is that while scientists came close to consensus at the end of the last century on race being a 'biologically meaningless concept' and 'predominantly socially defined' they whispered these phrases among themselves and left the wider public in the dark.
For the record, the imagined 'races' do not differ very much at all in physique or appearance. They really only differ considerably in skin pigmentation, possibly the most superficial trait going. Curtis Mayfield was no scientist but 'human people and there's only one kind' (8) is a biological truism.
The hypothesis that the Sunday Times claims is awaiting confirmation represents not an extension of the race myth but its essence. Physical differences are ultimately all superficial; psychological differences are not. That's why it is so important to say that this study is scientifically destitute.
Which it is. And although it might be less horrifying, it is just as false as any study you could imagine locating the gene for laziness among the blacks, or miserliness among the Jews.
If you think that is an overreaction, consider the following - after reading the Sunday Times' article I plugged phrases from it into Google in an attempt to locate the press release that, I imagined, inspired the story. I didn't find the press release. But I did find the article being bandied about approvingly on the message boards for a well-known white-supremacist website. You'll forgive me, I hope, for not providing the reference.