the g factor: a statistical myth

An academic of the Ludwig-Maximilians-University Munich sent me a link today leading to an essay:
— g, a statistical myth.

G is a so called “global intelligence factor” and was suggested by of Charles Spearman in 1904*, and it is still widely believed to be true.

Spearman’s idea was that a student’s grade in, say, English was the sum of two factors, a general factor, common to all subjects, and a specific factor unique to English, plus random, noisy, test-to-test variability. Similarly grades in math would be the sum of the general factor, a math-specific factor, and noise. The specific factors were, Spearman thought, completely uncorrelated, so all the correlations between math and English grades would be due to the general factor.

The article presents reasons to doubt a global intelligence factor.

I have to say that I find it an extraordinarily silly inference, and I’m astonished that anyone who understands how to calculate a heritability has ever thought otherwise.

[...]

This was, again, first demonstrated by Thomson — in 1914. I’ll go over a slight variant of his original model, in the hope that it will lessen the odds that we have to spend the next 93 years debating what ought to be a closed issue.

It’s funny to read, it’s good to children, it votes for Obama, and you can even understand most parts of it without knowing how to spell the word statistics properly (well, to be honest, that last part was most probably not entirely the truth).

– - – - – - – - – - – - – - – - – - – - – - – - – - -
* American Journal of Psychology 15, 201-293

(thanks to Dr. Heene)