Verified by 
Intelligence
The Equality of the Sexes I: Fact or Artefact?
Did the original IQ test constructors bias the tests in favor of men?
Posted January 20, 2009
"The one exception to the general rule that different groups or populations usually differ in average IQ is that both sexes have approximately the same average IQ on most tests. This is not, however, a true empirical finding but a consequence of the manner in which the tests were first constructed...the two sexes were defined to have equal intelligence rather than discovered to have equal intelligence." (Evans and Waites, 1981, 168).
It is widely believed that the equality of general mental ability differences between men and women is purely an artefact of the way the tests were initially created. Indeed, researchers often treat this idea as fact. But is it?
Like many good scientists, I don't like it when political correctness gets in the way of the truth. If such political correctness influenced the original IQ test constructors, this was not good science and these researchers should rightfully be exposed for the charlatans they are accused to be. If, however, they are wrongly accused as having been politically correct when in fact they were thoughtful, reasonable fellows, and the outcome just happened to be something that appears politically correct, then I think the record should be set straight. I believe in giving blame where blame is due and giving respect to those who deserve respect. However, just because an outcome seems politically correct certainly doesn't mean that individuals intentionally manufactured that outcome. It is indeed possible that politically correct outcomes can come about of their own natural accord. So let's go back in time for a moment and take seriously this accusation that has been placed on the original IQ test makers by critically examining what really happened.
Luckily, I have the help of a great intelligence researcher-Professor of Experimental Psychology at the University of Cambridge and one of my lifelong mentors-Nicholas J. Mackintosh. I was fortunate enough to do a Masters degree with him in 2004-2005 where we investigated sex differences in cognition. The history of IQ testing I will report here is included in his comprehensive, balanced, and reflective book IQ and Human Intelligence. So what is Mackintosh's conclusion on the matter? In his words, "...[A]lthough it contains a small grain of truth, it is a serious misrepresentation of the history of IQ testing (p.182)." Let's see how he arrived to this conclusion.
One of the earliest theorists of the nature of human intelligence was Francis Galton, who lived from 1822-1911. Like many men of his Victorian generation, he didn't have much confidence in the mental abilities of women. In fact, Galton thought that women's primary biological function was to bear children and do labor. "It always seems to me that a hard-worked woman is better and happier for her work. ... A nursery maid, who is on the move all day, seems the happiest specimen of her sex...", Galton said in his book Hereditary Genius in 1869.
So right off the bat, the history of thinking on intelligence began with less than favorable assumptions as to the intelligence of women. Thank goodness, however, that Galton's original tests of intelligence aren't the primary measurement of intelligence in school today. Indeed, Galton thought the best way to measure intelligence was through sensory-discrimination tasks such as grip strength.
As it turns out, two of the earliest IQ test constructors, Cyril Burt and Lewis Terman did not share Galton's view on the mental abilities of woman. Or if they did, they apparently didn't let that get in the way of arriving at the truth. Indeed, they both felt that the issue could be resolved by careful empirical enquiry. In the early 20th century, Burt administered a wide range of tests to schoolchildren of both sexes and found a variety of differences, some favoring boys, others favoring girls, but most not as large as the differences between men and women in physical characteristics.
In fact, the one consistent trend they found was that the larger the sex difference on a test, the less that test related to teachers' assessments of general intelligence. This led Cyril Burt to conclude "The higher the process and the more complex the capacity, the smaller, on the whole, become the sex differences (Burt & Moore, 1912, p. 379)".
Likewise, Lewis Terman empirically found few differences between the sexes in general intelligence. As Terman wrote in his introduction to the Stanford-Binet test, a widely used test even today:
"Many hundreds of articles and books of popular or quasi-scientific nature have been written on one aspect or another of this question of sex differences in intelligence; but all such theoretical discussions taken together are worth less than the results of one good experiment. Let us see what our 1,000 IQs have to offer towards a solution of the problem... When the IQ's of the boys and girls were treated separately there was found a small but fairly constant superiority of the girls up to the age of 13 years, at 14 however the curve for the girls dropped below that for boys..however the superiority of girls over boys is so slight...that for practical purposes it would seem negligible." (Terman, 1916, pp. 60-70).
So, if anything, Terman found a slight superiority for girls. This was based on a standardization sample of approximately 1000 boys and girls, aged 4-16. That's considered a large enough sample to detect a real difference if it truly exists.
As Mackintosh notes, "It is quite clear that neither Terman, nor Binet before him, had given any thought at all to the question of possible sex differences in deciding what items to include or exclude from their tests (p.183)."
Now, this does not mean that the question of sex differences wasn't ever given consideration by the test constructors. In their introduction to the first revision of the Stanford-Binet, Terman and his collaborator Maude Merrill eliminated a few tests in their trial batteries that yielded the largest sex differences. There is evidence that David Wechsler, another IQ test contructor did the same. But this was after they did their first standardization.
And even by doing so, the effect didn't wash away all sex differences. In one of the earliest IQ tests, the Wechsler-Bellevue test (which still has some items that are on the widely administered modern day WAIS IQ test), there remained small differences in overall scores in favor of women. This led Wechsler to express: "We have more than a ‘sneaking suspicion' that the female of the species is not only more deadly but also more intelligent than the male (Weschler, 1944, p.107)".
Therefore, it should be crystal clear that IQ tests were not in fact designed from the very beginning to produce equal scores for the two sexes. Terman genuinely believed in 1916 that he was making an empirical discovery, and he genuinely did.
As for Terman and Wechsler, Mackintosh states: "...[U]nless he and Wechsler were carefully concealing what they were up to, their initial tests discovered that there were no more than trivial differences in overall IQ between the sexes, and that such differences as there were favoured women."
The natural next question however is: Were Terman's and Wechsler's original conclusions correct? Are men and women really equally intelligent? To answer this question requires an understanding of the fact that the equality of the sexes is a consequence of the way IQ tests are constructed. What? But you thought I just demonstrated this wasn't true? More on that in Part II.
© 2009 by Scott Barry Kaufman
References
Burt, C.L., & Moore, R.C. (1912). The mental differences between the sexes. Journal of Experimental Pedagogy, 1, 273-284 and 355-288.
Evans, B.. & Waites, B. (1981). IQ and mental testing: An unnatural science and its social history. London, UK: Macmillan.
Mackintosh, N.J. (1998). IQ and Human Intelligence. Oxford, UK: Oxford University Press.
Terman, L.M. (1916).The measurement of intelligence. Boston, MA: Houghton Mifflin.
Wechsler, D. (1944). Measurement of adult intelligence, (3rd edm). Baltimore, MD: Williams and Wilkins .
