Eric Turkheimer of the University of Virginia has a new study out showing that the genetic heritability of IQ is much smaller among very poor kids than it is among middle-class kids. [No on-line version available, as far as I can tell; its forthcoming in Psychological Science.] The obvious interpretation is that middle-class environments are similar enough so that most of the child-to-child differences are innate, while the environments facing some poor kids are so deprived that their “natural” intelligence doesn’t have a chance to show.
The more general point is that the “heritability” of a trait is not a constant across environments, and that the attempt to parse all differences into an environmental component and a genetic component whose coefficients sum to unity — the basic project of The Bell Curve — is incoherent.
The finding strengthens the case for public policies to enrich the environments faced by poor kids. (Among the best of these might be making their parents less poor.) It does not, of course, say anything about the merits of individual programs.
Kevin Drum, who spotted the study in the Washington Post [*] and from whom, in turn, Tapped picked it up, seems to think that the finding shows that the Bush plans to change Head Start are a bad idea. [*] I don’t know the details of the Bush proposal, and I’m more than willing to believe that they’re bad, but Kevin’s conclusion doesn’t follow from the evidence presented.
Head Start was a nice idea, but the execution was lousy and the measured results have been uniformly disappointing. The emphasis is social rather than cognitive, and staffing tends to be “community” members rather than the skilled teachers needed to to do the job of helping kids from poor families overcome the heavy handicaps they carry as they enter school.
Everyone knows about the Perry Preschool Project, which was proven to be successful in terms of improved high-school graduation rates and reduced criminal activity.[*] But that project, though it’s often described as Head Start-like, was in fact much more intensively academic and much, much more expensive than Head Start itself: it cost about $4000 (in 1967 dollars) per kid, or about $20,000 per kid in today’s money.
The reported results in terms of crime alone, if they were real, make spending that money look like a bargain. But that’s not to say that Head Start as it stands is worth much of anything. Instead of defending Head Start, liberals should be proclaiming that it is inadequate and insisting on extensive and expensive fix.
Liberal conservatism — the tendency to defend existing programs rather than clamoring for change — is among the heaviest burdens we bear in attempting to displace the existing conservative hegemony.