Greg Mankiw’s self-described “boring” post about SAT scores and income has generated some interest in the econ blogosphere. The following graph, which shows that SAT scores go up with parental household income, appeared in the New York Times.
With good justification, Mankiw called it “the least surprising correlation of all time,” and declared the graph basically meaningless. We can’t simply conclude that more family income directly causes higher SAT scores. After all, the parents’ income partially reflects their IQ, which is partially transmitted genetically to their children.
Mankiw’s discussion went no further than to point out this interpretive difficulty, but his mention of genes and IQ sparked a reaction from Brad Delong and Paul Krugman, who are quick to point out the ways in which parental income and other environmental factors could affect SAT scores.
Their reaction illustrates a bit of a double standard. When someone suggests that genes explain a significant part of the variance in some outcome variable, critics rush in to complain that environmental factors have been ignored. The implication, sometimes not so subtle, is that the person they are criticizing is ignorant and uncaring, and a genetic determinist to boot. This rarely seems to happen the other way around. When a person mentions the importance of an environmental factor that influences an outcome variable, he is usually not bombarded with accusations of being a naïve, blank-slate worshiping bleeding heart.
So what is the precise gene-environment mix for determining IQ and, by proxy, SAT scores? No one knows for sure, but adoption studies can shed some light on the issue. We know that unrelated siblings adopted into the same home will have similar IQs during childhood, but by adulthood their IQs will be almost as different as those of two strangers. The adult adoptees resemble their biological family much more than their adoptive family.
So do genes dominate the home environment as a predictor of IQ? Not so fast, says Richard Nisbett, whose book on raising IQ I recently reviewed. Nisbett says that children are often adopted into environments similar to the ones they came from, and too many adoption studies examine children who are simply transferred from one middle-class family to another. This compresses the range of childhood environments that we observe, reducing the statistical effect of nurture.
Nisbett cites instead specialized adoption studies that compare poor children adopted into wealthier homes with similar poor children who were not adopted. He concludes that the IQ gain for adopted kids is somewhere around 15 IQ points, or one standard deviation.
Though beloved by the New York Times, Nisbett is expressing a minority viewpoint that is overly optimistic. The problem with his argument is that none of the special adoption studies he references show adult IQ scores. It is well-known that the effects of the home environment are significant through early adolescence, and they do not typically fade until the late teens and early 20s. The studies Nisbett cites as “proof” that home environments matter a lot are not inconsistent with the predominantly genetic view.
I hear that Nisbett is attempting to track down participants in one of the poor-children-into-rich-homes adoption studies now that they are adults. Assuming he manages to get a representative sample, it should be interesting to see how much of the IQ gains in the treatment group have been preserved. Let me venture an unscientific, gut-feeling guess—the treatment group will have IQs about 6 points higher than the control group in adulthood. Anyone else want to guess?