Thursday, April 16, 2009

Pundit, smarten thyself

Kristof, in his predictably inane column on "how to raise our IQ":
The implication of this new research on intelligence is that the economic-stimulus package should also be an intellectual-stimulus program. By my calculation, if we were to push early childhood education and bolster schools in poor neighborhoods, we just might be able to raise the United States collective I.Q. by as much as one billion points.

I wonder if he realizes that by this logic we ought to be engendering enormous quantities of children, mentally retarded or not.

Mark Liberman at LanguageLog periodically grumbles about the fact that no one ever cites standard deviations, and a mean without a standard deviation -- or some other measure of the width of the distribution -- contains no useful information. Kristof's editorial provides a good example of what this is about.

By age 5, the children in the program averaged an I.Q. of 110, compared with 83 for children in the control group.

Depending on the sample size and the variance this could be marginal, massive, or entirely insignificant.

1 comment:

andy said...

Thanks Sarang. I read that column yesterday and I thought "reason sure isn't Kristof's strong suit."

I also take issue with his assumption that the stimulus package can simply "bolster public schools in poor neighborhoods." If that were so simple, then why hasn't anybody done it already?