Somehow, one of the annoying qualities of one Andrew Sullivan is this compulsion to serve as an apologist for those whose stock and trade is race-based research on IQ. I guess the old man can't help himself, since he is at it once again. This is a topic that I have come back to on occasion, mainly because it serves as a reminder that the ugly side of IQ research has its origin in a peculiarly Eurocentric worldview, going back centuries. Long after the pseudoscience of eugenics lost its much undeserved credibility during the middle decades of the 20th century, its practitioners and apologists have continued to find outlets for their particular bile - primarily in "serious" newspapers and magazines. I've written plenty in the past about eugenics and IQ, if you care to read it. Much of what Sullivan wishes to defend is "empirical reality" is such only in the loosest sense of the term. Yeah, you can find data collected, but given the shoddiness of the work, and the racist agenda driving the "research" in question, it's more science fiction than science. The sad thing is that it tars some legitimately good research on intelligence in the process. Suffice it to say, it's a phenomenon that is considerably more complex than measurable with IQ tests (which are themselves of still questionable validity), and one in which the real differences of interest are not group differences.
I'll also link to a few others who express their disagreement with Sullivan quite eloquently: Brad DeLong, Lawyers, Guns, and Money, Dana Goldstein, and Ana Marie Cox.