Peter Medawar on IQ
I've been rediscovering the erudite and rather fabulous Sir Peter Medawar. The two bits below are from a piece that he calls "further footnotes" to a book review of his called "Unnatural Science"*
1. The question of the single-number valuation of IQ. Several correspondents have spelled it out to me that both athletic prowess and a human being's body temperature, for example, are influenced by a host of variables; yet an athlete's performance in the hundred-yard dash and a patient's temperature are both recorded in single values on one-dimensional scales. This is very true, but a patient's temperature is not taken by any physician to measure his state of health, nor is an athlete's speed in the hundred-yard dash used as a measure of athleticism or any other physical analogue of intelligence. The one measures a patient's body temperature only, and the other how long a man takes to run a hundred yards in a hurry. In just the same way an IQ test measures a candidate's prowess at the particular, kinds of intellections which are measured by such tests.
[…]
5. An anecdote about wicked old Sir Francis Galton will illustrate very clearly the difference between an IQ score and intelligence. Lewis Terman, God alone knows how, estimated Galton's IQ as 200—a figure of which he said that it was not equaled by more than one child in 50,000 of the generality—but at the same time we know from Galton's own memoirs that when at age eight he was issued with Caesar's Commentaries for class use he was vastly surprised to find his copy so new-looking and shiny, considering that, having been written by Caesar, it must be getting on for 2,000 years old. Yet is it not absurd to ask oneself—as strictly speaking one should if one is to acquiesce in the illusion of single-value mensuration—how much must be deducted from an IQ score of 200 to make allowance for so egregious a mistake?[Peter Medawar, "Unnatural Science, Cont'd.", The New York Review of Books 24, Number 11, 23 June 1977.]
———-
*P. B. Medawar, "Unnatural Science", The New York Review of Books 24, number 1, 3 February 1977; reviewing The Science and Politics of IQ, by Leon J. Kamin, and The IQ Controversy, edited by N.J. Block, edited by Gerald Dworkin. The piece begins this way:
If a broad line of demarcation is drawn between the natural sciences and what can only be described as the unnatural sciences, it will at once be recognized as a distinguishing mark of the latter that their practitioners try most painstakingly to imitate what they believe—quite wrongly, alas for them—to be the distinctive manners and observances of the natural sciences. Among these are: (a) the belief that measurement and numeration are intrinsically praisworthy activities (the worship. indeed, of what Ernst Gombrich calls idola quantitatis); (b) the whole discredited farrago of inductivism — especially the belief that facts are prior to ideas and that a sufficiently voluminous compilation of facts can be processed by a calculus of discovery in such a way as to yield general principles and natural-seeming laws; (c) another distinguishing mark of unnatural scientists is their faith in the efficacy of statistical formulas, particularly when processed by a computer — the use of which is in itself interpreted as a mark of scientific manhood.
2 Responses
Subscribe to comments via RSS
Subscribe to comments via RSS
Leave a Reply
To thwart spam, comments by new people are held for moderation; give me a bit of time and your comment will show up.
I welcome comments -- even dissent -- but I will delete without notice irrelevant, rude, psychotic, or incomprehensible comments, particularly those that I deem homophobic, unless they are amusing. The same goes for commercial comments and trackbacks. Sorry, but it's my blog and my decisions are final.
on Wednesday, 8 March 2006 at 09.10
Permalink
The entire edifice of the statistical approach to scientific analysis begins to wobble if one considers the most fundamental limitations imposed by the theory of numbers:
1. The final result of any computation involving multiplication, division and the extraction of roots can have no more significant figures than the input parameter with the fewest significant digits.
2. For computations involving addition and subtraction, no power of 10 can have significance in a result if it enters the calculation as a non-significant digit.
3. The mean of a series of measurements can have no more significant figures than the measurements themselves. (Follows from 1 & 2).
Therefore most statistical tables are worthless.
I also agree with Medawar's criticism of the whole discredited farrago of inductivism — especially the belief that facts are prior to ideas and that a sufficiently voluminous compilation of facts can be processed by a calculus of discovery in such a way as to yield general principles and natural-seeming laws;
Rita Levi Montalcini (Nobel Prize for Medicine 1986) said "I don't believe there would be any science at all without intuition".
on Thursday, 9 March 2006 at 21.00
Permalink
Points #1 & #2 are true, but trivial. Significant figures aside, point #3 is misleading: a series of measurements that represent a statistical sample of the measurement population can and do produce a mean value with greater precision than the individual measurements. What this has to do with the breathtaking leap to the idea that "most statistical tables are [therefore] worthless" is beyone my comprehension.
I wouldn't go nearly so far as to equate Medawar's criticisms as discrediting the "farrago of inductivism", but it's easy enough for me to go along with the idea that intuition is essential to science.