Case Study Three: Education and IQ

Herrnstein and Murray say (in an appendix, not the main text) that on the Armed Forces Qualifying Test, their basic measure of IQ, every year of education adds about one point to the IQ score. Christopher Winship of Harvard University and Sanders Korenman of Baruch College reran the same data, and--as they note dryly in their paper on the subject--"We found it quite difficult to replicate Herrnstein and Murray's results."

Winship and Korenman found, first, that Herrnstein and Murray had incorrectly described the way in which they had handled the data, and second, that they had severely miscoded the educational levels of seven people in the sample in a way that led the overall effect of education on IQ to be understated. If Herrnstein and Murray had run the numbers accurately and in accordance with their own description, their result would have been that each year of education increased IQ by 2.5 points, more than double the figure they gave in The Bell Curve.

Winship and Korenman further note, "It seems peculiar that Herrnstein and Murray did not think they should control for family background in estimating the effects of education on AFQT." By adding Herrnstein and Murray's own measure of socioeconomic status to the model, the effect of education on IQ rises again.

In the paperback edition of The Bell Curve, Murray has raised his estimate of how much each year of education increases IQ--without comment, and without facing the obvious implication for his central thesis.

{{{{Case Study Four#50885}}#2:CaseStudyFour.asp}}: Socioeconomic Status

Back to {{Case Study Two#50883}}{{#2:CaseStudyTwo.asp}}: Smart Rich People