The Magic, Necessity, and Illusion of Statistics

Individuals wait for quick tests for COVID-19 at Hospital Franciscano Nossa Senhora das Graças, in Lagoinha, São Gonçalo, Brazil, on Nov. 25, 2020. (Luis Alvarenga/Getty Images)

Commentary

Newly literate people tend to have a reverent attitude towards the printed page, for they suppose that if the truth can be printed, what is printed must be true. Statistics have a similarly magical quality for many, perhaps for most, of us nowadays, as if by their very nature or existence they conveyed a superior insight into reality.

Statistics exert an oddly mesmerising influence on us. The other day, for example, I ventured by chance on to a Brazilian news website which reported that, according to the latest statistics, Brazil had fallen three places in the list of countries in the world with the highest death rates from Covid-19 per million of the population. It was now slightly behind the United Kingdom, but slightly ahead of the United States, whereas before it had been ahead of Italy and Spain—ahead in this case being undesirable, and behind being desirable.

A cause for rejoicing then, if not for national pride! If I had been Brazilian, I would have been much heartened by the latest figures.

But hang on a moment! The numbers of deaths per million were still rising in Brazil, as they were bound to do if people continued to die from the disease: it was just that they were not rising as rapidly as they were in some other countries.

This should have been at most a very minor consolation for Brazilians unless, that is, they took particular pleasure in the misfortunes of others—which, alas, many of them, being human, probably do.

Was it not La Rochefoucauld who said that there is in the misfortune of our friends something not entirely unpleasing? We may be dying in numbers, but at least people are (proportionately) dying even more frequently elsewhere. That is a consolation to us.

Tablets of Stone

But there is an objection to such statistics other than that they reveal our surreptitious delight in the misfortunes of others, and that is that they immediately turn into tablets of stone in our minds, as if they were indubitable and unalterable truths sent down from Sinai. Often our critical faculty evaporates when a statistic is cited: we bow down and worship it, even though it may be little more than a vague surmise.

The person who read the article on the website to which I have referred uncritically would not realise just how flimsy were the claims made in it.

He would not be aware, for example, that the differences in the death rates for the various countries cited were so small that they fell well within the margin of error; that the criteria by which a death was ascribed to Covid-19 might not be the same between countries, or that it might vary even within countries; that the ascertainment rate of the cause of death might be different according to the efficiency or otherwise of the medical and statistical services of the various countries; or that all such statistics are subject to subsequent revision.

Nor would such a person be aware that he would reach precisely the opposite conclusion had he read a different source of information, if you can call anything so doubtful information. (In fact it was not difficult to find statistics that contradicted those of the article on the website.)

And if two people, each armed with his own set of uncritically-accepted but contradictory statistics met and discussed the matter, their tone might soon mount and they would fall out as if on a matter of deep principle.

The problem is compounded by the increasing inability of people to distinguish between denial and refutation. It is often reported these days that a man has refuted an accusation when he has merely denied it. The confusion between the verbs to deny and to refute is significant because it suggests a culture in which it is accepted that each person has his own truth.

Eternal Vigilance

On the day I came across the Brazilian website I read an article in my local newspaper in England reporting on the increase in the rat population of the nearest large town to my own. It reported that, thanks to the lockdown, there were now twice as many rats—literal rats, not the metaphorical human variety—in the town as there used to be, and that they now outnumbered the humans by two to one. It gave their number as 358,616.

How many readers would have laughed at this preposterously precise figure, and how many would have thought that they now knew the exact rat population of the town? I would imagine—but I have no statistics on the matter—that the latter would have far outnumbered the former.

I am not myself immune from the psychological effects of statistics whose accuracy and meaning must be doubtful. When I read or hear the latest economic figures—growth, unemployment, and so forth—my spirits rise or fall accordingly.

Growth pleases me, even though, in the abstract, I do not believe that ever rising levels of consumption are the purpose of life or the secret of happiness, and even if that growth neither affects me personally nor is immediately visible around me.

I know that the unemployment figures are so often manipulated for political purposes that (except in times like the present, when it is obvious that the number has risen) they mean very little. All the same, doubtful statistics affect my mood and rational considerations do not help me to eliminate the effect.

And yet I also recognise the necessity for statistics. We cannot do without them. The world is so large, so various and complex that one cannot hope to seize its reality by means of personal experience alone. It is a long time since doctors, for example, could simply assert as evidence their own experience that such-and-such a treatment works in such-and-such an illness, and expect either to be believed or respected.

You cannot know the number of years by which life expectancy has increased simply because your Aunt Jemima lived to be 103 (and smoked fifty cigarettes a day for eighty-five years).

A French friend who worked in the French government statistical office said that it was known to many of those who worked in it as the department of lies. This may be an exaggeration (I don’t have the statistics on the proportion of statistics that are falsified rather than merely mistaken, or are actually true). Lies were more often by omission than by pure invention.

In short, the price of believing true statistics and disbelieving false ones is eternal vigilance, and even then success is not assured. Moreover, eternal vigilance is both difficult and tiring. I have lapses myself.

Theodore Dalrymple is a retired doctor. He is contributing editor of the City Journal of New York and the author of 30 books, including “Life at the Bottom.” His latest book is “Embargo and Other Stories.”

Views expressed in this article are opinions of the author and do not necessarily reflect the views of The Epoch Times.