Kevin Drum is getting some mileage out of this crime-lead article. He’s now making an argument I brought up two years ago (not that I was the only one) when I first read Drum on the topic.
If lead is one factor that influences IQ (and behavior), and if blacks tend to have higher levels of lead exposure than whites (this is the only fact whereas everything else is interpretation), how does this change our genes/environment allocation on the IQ question? Is this just a matter of lead taking up a larger share of the residual left over after the ~50% heritability of IQ?
Drum quotes Rick Nevin who has done the most work on translating lead exposure to group-level outcomes:
In 1960, blacks occupied 15% of central city households and 56% of substandard central city housing…. Average 1976–1980 blood lead for black children ages 6–36 months was 50% above the average for white children….Those children were juveniles when the 1990–1994 black juvenile burglary arrest rate was 60% higher than the white rate, but the black juvenile violent crime arrest rate was ﬁve times higher and the black juvenile murder rate was eight times higher.
More from Nevin on lead exposure and race can be found in his working paper “Lead Poisoning and The Bell Curve“. In it, he wrote:
An increase in blood lead from 1 to 10 mcg/dl (micrograms of lead per deciliter of blood) is associated with a loss of 7.4 IQ points (Canfield et al. 2003). Another one-third IQ point is lost per mcg/dl increase from 10 to 15 mcg/dl, and one-quarter IQ point lost per mcg/dl over 15 mcg/dl. (Schwartz 1994). Therefore, blood lead over 20 mcg/dl can lower IQ to less than 90 among children who would have had average IQ of 100, and blood lead over 40 mcg/dl can lower IQ to less than 75 for children who would have had IQ of 90 (Figure 2). When large scale blood lead screening began in many USA cities around 1970, 25% of city children tested had blood lead over 40 mcg/dl. (Gilsinn, 1972). In addition to reducing IQ, preschool lead exposure is directly associated with an increased risk of violent and delinquent behavior.
Nevin seems to answer the question posed up top about the allocation of lead to differences in intelligence test scores:
USA blood level trends explain 65% of the 1953-2003 variation in average SAT math scores, 45% of variation in SAT verbal scores, and 65% of 1948-2001 MR [mental retardation] prevalence.
Between the 1988-1991 cohort and a 1999-2004 cohort, average childhood blood lead levels for black kids went from 5.2 mcg/dl to 2.8 mcg/dl. White kids’ levels decreased from 3.1 mcg/dl to 1.7 mcg/dl. If we’re comparing average lead levels and average IQ for both groups, these blood lead level differences still don’t explain IQ gaps. If lead were an important environmental explainer we’d expect white IQ (1991) to match black IQ (2004) since blood lead levels for those groups were about the same.
Via Rick Nevin
Another way to look at this is that if the average black in 1991 had about 5 mcg/dl of lead in their blood, they lost about 5 IQ points from their “unleaded potential” IQ level. Whites had lead in their blood as well – about 3 mcg/dl’s worth which matches to about a 3 IQ point loss. The commonly-cited black-white IQ gap is about 12 to 15 points. Lead exposure for the average of both groups seems to only account for a couple of IQ points. This is just taking Nevin’s citations for granted. Studies not cited by Nevin show a less steep IQ-lead slope. Needleman’s study shows a 6.9 IQ point decrement when blood lead levels rise from 2.4 to 30 mcg/dl whereas Nevin seems to cite one study that pegged the decrement to 7.4 points when lead levels rise from 1 to 10 mcg/dl.
If you look back at old murder rate statistics you never see the white murder offense rate come near the level of the black murder offense rate. And this despite the fact that more recent black lead levels are less than white lead levels from a couple of decades ago. In the late 1970s the average black lead level was 20.2 while the white level was 13.7 mcg/dl. Currently average black lead levels are not as high as whites’ were, yet black murder rates and overall crime rates are several multiples larger all the way across the board.
Nevin’s research and Drum’s article are fascinating because of the surprise factor and the potential of being a sort of “theory of everything”. They present legitimate arguments that any Bell Curver has to pay attention to, but the explanatory power, at least as far as IQ is concerned, seems not to be there.
One common refrain is that Nevin is the Rodney Dangerfield of criminology. His research hasn’t gotten the respect it deserves. Perhaps true, but the problem here is that Nevin’s research is heavy on correlation but light on rigorous statistical analysis. He’s done all of Drum’s work for him, but tried to push it through the academic world instead of through a politically active magazine.
For further reading:
Jim Manzi at National Review
Scott Firestone at Discover
Steve Sailer grapples in Drum’s comments section
In NYC, 1 in 4 children who suffer from lead poisoning are Asian