Live by the Numbers, Die by the Numbers

The following article was called to my attention by Dirk van Putten.

“Odds Are, It’s Wrong: Science fails to face the shortcomings of statistics”

The author claims that “…a mutant form of math has deflected science’s heart from the modes of calculation that had long served so faithfully. Science was seduced by statistics.”

Of course the author is wrong. It is not the statistics that lied, it is the misusers and abusers of statistics who have spread fallacious ideas, intentionally and unintentionally.

The fascination with numbers goes back a long way in human culture. Numerological correlations with observed events have been integrated with sacred mythologies as far back as we can see — the Egyptians, the Druids and the Mayan’s, for example. — but it was not until the secular revolutions of the 17th and 18th century that this numerology came to be accepted as the principal instrument of truth-saying with respect to the nature of “reality”.

The explosion of technological achievement realized by Newton’s mechanics for example, marked the rapid emergence of a widespread belief of an independently verifiable clockwork reality and that view came to be seen as triumphant over, for example, sacred experience and philosophical inquiry. Thereafter, to be scientific, inquiry in all realms of human experience aspired to numerological certainty. Facts are facts, and all we need to do if find them. Paradigmatically, the ascendance of this view marked the beginning of the age of discovery in which the creative powers of humans beings become subordinated to a world of discoverable external constraints — the “natural laws” that govern our human aspirations. These may be useful when applied to building automobiles, but maybe less so when building communities of human beings. (Case-in-point — The current healthcare debate.)

While the instrumentality of Newtonian mechanics was and continues to be, compelling, the evidence that, given enough data and computational power, the world can be reduced to a set of tractable equations though which, at least hypothetically, true value and perfect prediction become possible, has not been born out by observation. But the belief in this reduction continues to hold sway.

So statistics, in the view of some, is viewed as a work-around to deal with problems of insufficient data and computational power,  but this was not Deming’s meaning. No amount of data and no amount of computational power can result in perfect prediction, but as a basis for action, from an economical standpoint, statistics can help us to predict better, given our aims. But if our aims are not agreed upon, the numbers become nonsense. “Facts do not and cannot speak for themselves.

Little advancement in the teaching of statistics is possible, and little hope for statistical methods to be useful in the frightful problems that face man today, until the literature and classroom be rid of terms so deadening to scientific enquiry as null hypothesis, population (in place of frame), true value, level of significance for comparison of treatments, representative sample. There is no true value of any concept that is measured. There may be, of course, an accepted operational definition (questionnaire, method of measurement) and an accepted value—accepted until it is replaced with one that is more acceptable to the experts in the subject-matter.

W. E. Deming

Here’s a discussion of the problem by Deming.

On Probability As a Basis For Action