Relatively Wrong

by Theodore Dalrymple

I would rather talk to a taxi-driver than to most intellectuals: usually a taxi-driver has more interesting things to say. Indeed, in my brief and undistinguished career as a civil-war correspondent, I realised that most of what appears in our newspapers about such conflicts is garnered from taxi-drivers on the way from the airport to the warring country’s only five-star hotel: and this is as it should be, for taxi-drivers are much less parti pris than Ministries of Information, the other possible source of information for most journalists—apart, of course, from other journalists in the hotel bar, that great echo-chamber of rumour and misinformation. 

I therefore took seriously the opinion of the taxi-driver from the airport in Paris—not, I hasten to add, to a five-star hotel—that, six weeks after the terrorist attacks, many people were still afraid to go out or to frequent crowded places. This was confirmed for me in the Métro, which seemed unusually quiet: though whether my observation was a true one, or was unduly influenced by my conversation with the driver, I cannot tell. No doubt there are trustworthy statistics on the subject, but in any case I would prefer to believe and trust to my own impressions. The Webbs’ book on the Soviet Union was full of statistics that they thought were trustworthy.

Rather grandly, I told the taxi-driver that I was not going to be cowed into altering my frequentations by mere terrorists: I would rather be killed, besides which it was statistically very unlikely that I would be killed even in the event of another attack. After all, I refuse the constant computerised invitations from my doctor for screening procedures to detect the diseases from one of which I am quite likely eventually to die: so why should I avoid so tiny a risk?

Actually, the comparison is a good deal less stark than at first might appear. Since my retirement I have paid closer attention to medical journals than I ever paid before, and I have been surprised by how much that is published in them is deeply flawed where it is not outright dishonest. Previously, I would (like most doctors) skim papers and read their conclusions on the assumption that they followed from data presented; I have since realised that this is far from always the case. But medical practice is highly influence by conclusions that do not necessarily follow from the evidence, and so I have become much less willing to be screened by any doctor than I would once have been.

One of the problems is the nature of so much of the research that is published in the general medical journals. It is not like the heroic days of medicine, when any fool could see that ether anaesthetised a patient undergoing an operation. Nowadays, research is conducted on hundreds of thousands of people, or even millions, and the conclusions rely on complex statistical manipulations of data that not one in a thousand doctors understands and which, in any case, almost always yield completely different conclusions when a different statistical method, equally incomprehensible to doctors, is applied.

The old rhetorical tricks of suppressio veri and suggestio falsi are employed, or at least are present in medical journals surprisingly often. Epidemiological evidence is particularly susceptible to misinterpretation. When factor a is correlated with disease b, for example, the relative risk is often given but not the absolute risk (people with, or who do, or who live in, a are 1.37 times more likely to suffer disease b than those without, or who do not do, or who do not live in, a). Statistical significance is confused with clinical significance: the poor doctor is often left to work out for himself whether the absolute risk to his patients is worth worrying about. Usually it isn’t.

The requirement that patients nowadays be told the truth and make choices for themselves leaves them with insoluble dilemmas. Many patients want to be told what to do, not to solve riddles with no indubitably correct answers. Is it worth taking a tablet every day for five years in order to avoid a one per cent risk of something nasty happening, especially as most people give up taking the tablet before the end of the first year and there is the risk of side-effects? The last time I tried this frank approach on a patient, a man of limited intelligence, he replied, “I don’t know, you’re the doctor”: which struck me as a perfectly reasonable thing for him to say.

No matter how statistically sophisticated a paper in a medical journal, when a correlation is found (and if you search hard enough, you can always find a correlation between something and something else), it is almost invariably considered causative, and speculation offered as to how factor a causes disease b. This is an elementary howler but it is made time and again. When a medical journal published a paper showing that heart attacks were correlated with low levels of selenium, brazil nuts promptly disappeared from the supermarket because they are high in selenium. (A paper in the New England Journal of Medicine not long ago suggested that nut consumption, including of peanuts which are not really nuts at all, was positively correlated with longevity. The elixir of life seems to be boiled oily fish in satay sauce with broccoli.)

You can have great fun with correlations. I once wrote that criminality among white men was caused by a long-acting neurotropic virus introduced by the tattooing needle, since practically all white men in prison are tattooed. Indeed, the association of criminality with tattoos was far closer than that between criminality and any other factor, including what used to be called broken homes.

To my amazement, some people took me seriously and argued in refutation of my hypothesis. Actually, in my heart of hearts I do believe that there is a connection between tattooing and criminality, though not a directly causative one. This does not mean, of course, that the current vogue among the middle classes for tattooing indicates an increase in criminality, only in criminally bad taste. As Aldous Huxley once said, most people will be vulgar if given the opportunity. I think, however, that there is more than this to the recent epidemic of self-mutilation—the fall of the Roman empire, perhaps?—but I can’t quite put my finger on it.

There is one important correlation that has been insufficiently noticed, namely that between European jihadism and rap music. Practically all the jihadis (for whom religion is the continuation of crime by other means) were, until their conversion, fond of rap music, if not actual practitioners of that dark art. Probably failure to become rap stars was to them what failure to obtain entry to art school was to Hitler.

The question is whether rap makes people stupid and aggressive, or whether only stupid and aggressive people listen to rap music.

It is, incidentally, more difficult to criticise modern popular music in print than to criticise Islam. When I wrote in an article for a French publication that rock concerts are like fascist rallies of licentiousness, the editor had to remove it because his staff threatened to resign if he printed it. When I wrote in a Belgian newspaper that modern pop music was responsible for deafness, car accidents and crime, there were innumerable calls for me to be silenced.

But to return to the Parisian taxi-driver. Having solved the problem of terrorism, we went on to the interesting question of why Parisian taxi-drivers are not obliged, like every other driver, to wear seat-belts. Was what another driver (whom I once would have called a lady driver) told me true, that it was so that taxi-drivers could get away more quickly from aggressive or violent customers? She had only once had to take advantage of this wise provision, most passengers offering her sex instead of a fare not having to be taken au sérieux.

No, said, the taxi-driver, this was not true. It was because it was uncomfortable to sit all day with a seat-belt across your front: a less interesting explanation. Poor fellow, he had had to wait three hours for a fare at the airport in a queue of taxis. It had turned him talkative instead of sullen. Apart from anything else, one can learn humility from a taxi-driver.

First published in Quadrant.