Advocacy by Omission: Richard E. Nisbett’s Intelligence and How to Get It

In his book, Intelligence and How to Get It, Richard E. Nisbett, a social psychologist at the University of Michigan, asserts that cultural factors alone are sufficient to explain all the race differences to be observed in IQ and educational achievement.

Nisbett criticizes the nature + nurture model Arthur Jensen and I presented in 2005 in Psychology, Public Policy, and Law [Thirty Years of Research on Race Differences in Cognitive Ability (PDF)]. Nisbett claims the heritability of IQ is lower than assumed, that Blacks have substantially narrowed the gap on Whites, that any remaining differences can be eliminated through educational and social intervention, and that any assertion of a mean IQ of 70 for sub-Saharan Africans is “desperately wrong.”

The dust jacket blurb puts Nisbett’s book in the tradition of Stephen Jay Gould’s Mismeasure of Man (1981, 1996). What is striking, however, is how much ground the egalitarians have given since Gould’s effort to debunk race, genetics, and IQ.

Nisbett concedes that general intelligence exists, that IQ tests predict success at both school and work, that scores are influenced by genes, and that in White populations, genes contribute to social class differences. He even accepts that IQ is related to brain size and that “Blacks are sometimes found to have smaller brains than Whites.” Gould must be spinning in his grave!

Such is the state of expert opinion today that Nisbett simply had to make these concessions lest his book be disqualified as serious scholarship. But while he admits that genes play a role in accounting for within-group differences, he still maintains they play no significant part in between-group differences.

Nisbett acknowledges that measured group differences exist: Jewish (mean IQ = 113), East Asian (107), White (100), South Asian (87), Hispanic (87), African American (85), and sub-Saharan African (70), although he erroneously claims that South Asians score as highly as East Asians. But Nisbett asserts that family pressure for success leads East Asians and Jews to high levels of achievement, while low expectations and a lack of opportunity lead Hispanics and Blacks to much lower levels of achievement.

Jensen and I have provided a long point-counterpoint review of Nisbett’s book in The Open Psychology Journal [Race and IQ: A Theory-Based Review of the Research in Richard Nisbett’s Intelligence and How to Get It (PDF)]. We found much that was admirable and informative in it, such as the cost-benefit survey of the literature on educational interventions and what can be done to increase school performance. But, sadly, we concluded it was mainly a work of advocacy, not scholarship.

Nisbett misrepresents much of the available information using highly selective reviews of the literature. Especially egregious are his many errors of omission. For example, while complaining of unrepresentative samples in a 1991 review paper by Richard Lynn, Nisbett side-stepped the very much larger compilation of data in Lynn and Vanhanen’s 2006 book. Nisbett completely ignores the adoption studies of East Asians, as well as those on brain size showing a genetic contribution to their high achievement.

I’m going to group Nisbett’s flaws under seven heads:

1. The malleability of IQ scores

According to Nisbett, Blacks closed the IQ gap by 5.5 points (35%) between 1970 and 1992. At the same time, Blacks narrowed the gap in educational achievement by a commensurate 35%. Nisbett argues that educational interventions such as the Milwaukee project, the Abecedarian project, and the Infant Health and Development Program, imply the gap could be eliminated altogether.

Contra Nisbett, however, Jensen and I found that IQ differences between Blacks and Whites have been steady for nearly 100 years, at between 15 and 20 points (about 1.1 standard deviations). After re-analyzing the data on which Nisbett relied, we found that the most optimistic assessment of the Black IQ gain was 2.1 points (14%). Using a wider array of tests, we found no narrowing at all. [See The Black-White Test Score Gap and the New Math Results By Charles Murray, AEI blog, October 15, 2009.]

Nor has there been much (if any) narrowing on tests of educational achievement. Furthermore, the most powerful intervention strategies result, at best, in moving people from the 16th to the 25th percentile at the cost of millions of dollars per person.

2. Adoption studies

Nisbett reviews several adoption studies to demonstrate the power of environmental intervention on IQ. He claims that poor children adopted into wealthier homes made huge gains — between 12 and 18 points. But he omitted to mention the finding that by late adolescence, these effects have dissipated.

Nisbett’s tendency to omit crucial information is particularly apparent in his discussion of the well-known Minnesota Trans-Racial Adoption Study in which White, Mixed-Race, and Black children were raised by upper-middle-class White parents in Minnesota. This was the largest study of its kind ever undertaken, and the only one with a longitudinal follow-up component, testing the same children at 7 years (in 1975) and 17 years (in 1986).

Nisbett describes how adopted White children averaged an IQ of 112 at age 7, but he omits that on follow-up at age 17, the IQ of these White children had fallen to 106. He did note that the Black children on follow-up had dropped from a mean IQ of 97 to one of 89; which he attributes to the psychological disturbance and “identity problems” they suffered from being raised by Whites!

Compounding his sins of omission, Nisbett neglects to mention three trans-racial adoption studies of East Asian children. In contrast to the Black children, the East Asian children adopted by White parents, despite being malnourished at birth, grew to excel in both intelligence (mean IQ = 108) and educational achievement.

3. Heritability

Nisbett presents a vivid analogy of tossing genetically identical corn seeds into rich soil or poor soil. The average height of the two groups of plants, he says, would differ greatly and would do so entirely because of environmental factors. Nisbett argues that because most heritability studies used middle-class samples (“good soil”), they overestimated the heritability of poorer people and Blacks (“poor soil”).

In support of his analogy, Nisbett cites a twin study by Eric Turkheimer in 2003 in Psychological Science [PDF] that found the heritability of IQ was about 70% for children whose parents were upper-middle-class but only about 10% for children whose parents were of lower social class (over half of them African-American).

But Nisbett omits the studies that have been unable to replicate Turkheimer’s results, including one of over 2,000 twin pairs by Robert Plomin and colleagues published in 2005 in Intelligence. He also omits the studies finding that heritabilities in Blacks were the same as those in Whites, such as Osborne’s study of several hundred pairs of Black and White twins. Osborne found heritabilities of about 50% in both groups on the Basic Test Battery, the Primary Mental Abilities test, and the Cattell Culture Fair Intelligence test.

4. Black-White differences are greater on the more heritable and g-loaded tests

Heritable g is at the core of the debate over how much Black-White IQ differences are due to the genes and can be expected to narrow. However, Nisbett declares that the studies by Jensen and myself finding that Black-White differences are more pronounced on the more heritable and more g-loaded IQ tests were a “red herring.”

To maintain his claim, Nisbett withholds from his readers the data showing that while gand genetic estimates correlated significantly positively with Black-White differences .61 and .48 (P < .001), respectively, they correlated significantly negatively (or not at all) with any secular rise in IQ scores over time (r = -.33; P < .00001) and .13 (ns). The (mostly heritable) cause of the one is not the (mostly environmental) cause of the other.

Most recently, my colleagues and I found the relation between heritable g and Black-White differences in a 2007 study published in the Proceedings of the Royal Society of London [PDF]. We examined data from 152 pairs of twins from the Minnesota Study of Twins Reared Apart and calculated separate heritabilities for each of the 58 diagrammatic puzzles on the Standard Matrices and found they correlated a mean .40 (P < .05) with the pass rate differences (on the same puzzles) between Whites, South Asians, Coloreds, and Blacks in South Africa.

5. The sub-Saharan African IQ of 70

Nisbett cites a critique of Richard Lynn’s compilation of 57 studies showing a mean IQ of only 70 for sub-Saharan Africans by University of Amsterdam psychologist Jelte Wicherts. In the journal <cite”>Intelligence [PDF]. Wicherts found several studies showing much higher IQs for Africans than Lynn provided. His re-analysis raised the African IQ to 80, although that is still a very low 1.25 standard deviations below the White mean.

Lynn has replied to Wicherts (also in Intelligence) [Subscriber link], countering that Wicherts included elite samples in his analyses, which can mislead when estimating the IQ of a general population. After considering all the evidence, Lynn’s best estimate of African IQ remains 70.

Lynn corroborated an African IQ of 70 with tests of educational attainment such as the Trends in International Mathematics and Science Study (TIMSS). This test is standardized with a mean of 500 and SD of 100 on representative samples of 14-year-olds with Ns = 2,000 to 10,000. On the 2003 TIMSS for England, Botswana, Ghana, and South Africa, the results were 503, 355, 278, and 268, respectively. Thus sub-Saharan Africans are 2 standard deviations below Whites in both IQ scores and tests of educational achievement.

6. Race Differences in Brain Size

Nisbett agrees that the correlation between brain size and IQ could be as high as .40 and that Blacks are “sometimes” found to average a smaller brain than Whites. However, he argues there were too many anomalies to take the race differences in brain size seriously as an explanation of IQ differences.

He points out that Blacks do not always have smaller brains than Whites, that the causes may be environmental, that the relation between brain size and IQ may not be causal, that men and women differ in brain size yet have equal IQ, that in Ecuador, there is a sample of people with very small head sizes who do better in school than normal individuals, that the evolutionary direction of brain size over the last few thousand years is down, and that Einstein had a smaller brain than the average reported for Blacks!

But, contra Nisbett, we found that each and every one of these points was either false or misleading. The race differences in brain size have been corroborated in dozens of studies from the 1840s to the present using MRI, endocranial volume from empty skulls, brain weight at autopsy, and external head size. Averaging all the data: East Asians = 1,364 cm³; Whites = 1,347 cm³; and Blacks = 1,267 cm³. Thus, East Asians averaged a 17 cm³ larger brain than Whites and 97 cm³ more than Blacks.

Nisbett omits all the brain size data on East Asians, who average a larger brain than Whites at birth, 1 year, 7 years, and adulthood. Since the East Asian samples are the smallest in stature and lightest in weight, and the Black samples the tallest in stature and heaviest in weight, these data contradict Nisbett’s claim that the smaller brain size in Blacks is due to lower birth weight and poorer nutrition.

Brain size is highly heritable. One MRI study in 2002 in

Nature Neuroscience examined 112 extended twin families and found heritabilities of 82% for whole-brain gray matter volume, 87% for whole-brain white matter volume, 86% for IQ, and 100% for the relation between them. Each cubic centimeter of brain tissue contains millions of brain cells and billions of synapses. Hence, race differences in brain size help to explain the IQ differences.

7. Racial admixture studies

Nisbett agrees that the U.S. Black population has about 20% European admixture and that Blacks with more European genes should have higher IQs. However, he argues that the correlation between IQ and skin color (about .20) is too low to support the genetic hypothesis. Any such small relationship, he says, would be due to the social advantages that accrue to Blacks with lighter skin.

Nisbett omits several more recent results. For example, a study in 2002 by Richard Lynn examined a National Opinion Research Center (NORC) survey of 442 Black adults who were asked whether they were “very dark”, “dark brown”, “medium brown”,“light brown”, or “very light.” The correlation between these self-ratings and a 10-word vocabulary test score was .17 (P < .01).

Data from South Africa also supports the hypothesis. The Mixed-Race population known as the “Coloreds” (their preferred term) have an average IQ of 85, intermediate to the respective African and White means of 70 and 100.

The brain weight data on mixed-race populations also support the racial admixture hypothesis. Robert Bean found (in 1906, in the American Journal of Anatomy,) [PDF] as did Raymond Pearl (in 1934, in Science) [Subscriber], that the greater the amount of White admixture in Blacks (judged from skin color), the higher the mean brain weight at autopsy. A 1997 study in Intelligence found mixed-race Eurasian children fell intermediate in brain size and IQ to the two parental groups.

In 2006, Donald Templer and Hiroko Arikawa (also in Intelligence) found a negative -.92 correlation between IQ and skin color for 129 countries. The darker the skin, the lower was the IQ. Skin color was measured by expert ratings and by skin reflectance.

Subsequently, Templer found IQ and skin color correlated with geographic latitude and mean winter temperature (r = .80). He conceptualized skin color as a multi-generational adaptation to colder climate, which selected for higher intelligence and larger brain size because of the greater cognitive requirements for survival.

Our conclusion: Predictions that the Black-White IQ gap will narrow are acts of blind faith. A much stronger dose of skepticism is required than Nisbett manifested in regard to the power of educational and social interventions.