Charles Murray

Doug Henwood dhenwood at panix.com
Wed Feb 3 16:42:46 PST 1999


curtiss_leung at ibi.com wrote:


> Glad you mentioned Gould. I had read _Mismeasure of Man_ way before
> _Bell Curve_ came out, and I was curious to see how Murray and
> what's-his-name answered Gould's arguments against the sophisticated
> statistical methods used by the new IQ racists. (In a nutshell: if I
> add my waist size to the room temperature in degrees Kelvin and then
> multiply by the number of Jaegermeister <uuukkk!> shots consumed by
> high school teens within a five block radius of my house last night, I
> will not have a meaningful measure of anything, and Gould argues
> persuasively that the fancy IQ metrics are much the same stuff as
> that.)

Gould reviewed the Bell Curve for the New Yorker, and pointed out that even by Murray & Herrnstein's own evidence, IQ explains only about 2-3% of social outcomes. That is, they claimed that IQ explained 50% of social outcomes and their regressions had r^2's of around .05 (the only time I've ever seen an r^2 quoted in the New Yorker!). So 5% of 50% = 2.5%. It's no wonder they presented only the regression trendlines in their texts, since if they'd shown the full scattergrams they'd have looked only slightly less than random. And that, of course, is on the basis of their own crackpot models.

Doug



More information about the lbo-talk mailing list