Nisbett's latest book serves several purposes. On one level, he is arguing with his fellow IQ researchers that "schools and culture matter" more than the field suspected (e.g. a 1987 survey by Stanley and Rothman found high estimates of individual heritability and plurality support for the view that both genetic differences and environment played a role in American IQ gaps). Nisbett marshals a persuasive case, and while his thesis is forcefully stated he is scrupulous in pointing out contrary arguments and evidence. Occasionally the main text seemed to overstate a point, but in the vast majority of such cases a turn to the notes section or appendices revealed that the complications had been mentioned there (with a few exceptions related to the views of hereditarians on dysgenics and the causes of low IQ in Africa), and the book could serve as a good introduction to the field for laypeople.
Another major element of the book is a call for bringing scientific rigor to education. Rigorous controlled experiments, research that takes our knowledge of IQ into account, genetics-aware studies such as Eric Turkheimer's (which measure environmental effects on phenomena such as substance abuse using twins to control for genetic effects), and clear thinking have the potential to greatly improve the effectiveness and cost-effectiveness of education and improve cognition. Fear of hereditarian views and associations with racism often hinder educational policymakers from taking our scientific knowledge of IQ into account, with negative effects on educational quality, and we may hope that this text will help to diminish that stigma and help to drive further improvements in cognitive ability.
At a higher level, the book is in some ways an ode to the randomized experiment, taking a strong stand against regressions that are helpless to identify causality and easily manipulable by motivated researchers. A reader gets the sense of a smart and honest empirical scientist who is eager to have questions resolved by hard evidence, and ready to change his mind in accord with those experiments, even when he comes to the table with strongly held initial views on a controversial topic of policy importance. This is something we need to see more of in scientific and policy debates and I give this book my strong recommendation. If you enjoy this book, I would suggest a follow-up with James Flynn's "What is Intelligence?" and Arthur Jensen's "The G-factor."
I will discuss the individual chapters below:
Chapter 1 provides a brief introductory account of IQ, and its importance for life outcomes, while mentioning other important traits such as motivation and self-control. It is quite satisfactory and puts IQ in context, without diminishing its role or getting embroiled in terminological disputes over the description of other abilities or discussion of their measurement.
Chapter 2 attempts to place the high heritability of IQ scores among middle class families (as much as 70-80% of variation in IQ can be explained by genetics within that group) into context, noting that adoptive households tend to be relatively good environments for children (confirmed by across-the-board boosts in IQ for adopted children, although these generally fade with time), and this consistently high quality means that variation in genetics plays a greater relative role. He also cites Turkheimer's research finding lower heritability of IQ among famlies with lower socioeconomic status.
Chapter 3 reviews the literature on the IQ-increasing effects of education, drawing on natural experiments of differential phase-in of education, or disruption due to strikes and disturbances. It is undeniable that schooling increases IQ relative to its absence, and these gains can be lasting. A discussion of the Flynn Effect, the secular increase in IQ scores over the last half century and more, follows James Flynn's "What is Intelligence?" and emphasizes that the particular IQ subtest scores that have increased cover skills that modern education and entertainment place much more emphasis on than in the past. One particularly devastating application is to the critique of 'culture-fair' nonverbal tests such as Raven's Progressive Matrices. These tests, which require subjects to identify patterns in abstract shapes, have shown extremely high culturally-driven increases over time, but they are also used extensively by researchers such as Lynn and Vanhanen to estimate developing country IQ levels. This application is typically defended along the lines that the within-country correlations of life outcomes and Raven scores are strong, but such correlations would not be disrupted by an across-the-board change such as the Flynn Effect (and have not been disrupted in rich countries) or its absence, and these tests are clearly 'culture-unfair.'
Chapter 4 covers the state of educational research, and is one of the best parts of the book, although the findings have been discussed elsewhere. Nisbett urges a scientific approach to educational research and policy, relying on randomized experiments, and critiques the ludicrous excesses of schools of education that reject the experimental method. Reviewing the research on funding levels, teacher quality, and class size, he rejects common myths and focuses on evidence. For instance, teacher quality matters quite a lot for student achievement, but it is not strongly related to seniority after the first two years or to Master's degrees, the criteria according to which most union contracts allocate raises. Instead, direct measurements of student improvement are the best predictors of future performance. Federal policy, the D.C. school system under Mayor Fenty and Michelle Rhee, and the Gates Foundation are all moving in this direction, and I hope that the new wave of experimentally-informed educational practices is able to displace ineffective policies and their political support.
Chapter 5 reviews the large differences in parenting between social classes, e.g. the number of words upper middle class parents speak to their children is much greater than the number spoken by lower class parents, and upper class parents are more encouraging of thought by their children. These results are confounded by genetics, which may simultaneously affect parenting style and child behavior, but they are very interesting nonetheless, and the gains of children raised in adoptive households and some parenting interventions suggest that they do have some causal importance. Importantly, such parenting practices do not change instantly when offspring rise into the middle class, which may be important in explaining the poor test performance of the children of first-generation middle class families of all groups.
Chapter 6 addresses the IQ gap between blacks and whites, arguing for a genetic contribution of zero, and must be read in tandem with Appendix 2, which provides more detailed arguments. Nisbett discusses a substantial partial closing of the IQ gap among children over the last 50 years, and corresponding improvements in school achievement. These gains weaken among teenagers and further among adults, but even on their own they represent a major advance, and suggest that efforts like KIPP could go much further.
Nisbett places a particular focus on evidence from European admixture in African-Americans (on average about 20%). For instance, he notes that skin color, facial features, and blood groups show only tiny correlations with IQ. These arguments turn out not to work because of the sample sizes and the genetics of those traits, but a similar analysis with DNA ancestry testing should now be feasible. More persuasively, Nisbett reviews the mixed evidence from studies of adopted children with one or two black parents. Each of the several studies has confounding effects at play, such as lack of parental IQ information, but collectively they leave it in doubt whether biracial children really do show an advantage over children with two black parents, and pose a challenge for hereditarians selectively focusing on the one (of three) major study showing an advantage for biracial children.
Nisbett presents a number of other lines of argument, with one particularly telling point being the fact that that among African-Americans women predominate in the upper IQ echelons, while among whites the pattern is reversed, and I expect that they will indeed shift scientific opinion towards a lower estimate of the importance of genetics. At the least, we should be confident that environmental interventions, provided that they are validated by randomized trials, can further increase black IQ and that fatalism is mistaken.
Chapter 7 discusses IQ-boosting programs such as the Perry Preschool Program, the Milwaukee Project, and the Abecedarian Program. Overall, there are many disappointing results and some gems, but those gems have shown in randomized experiments that they produce benefits that exceed their costs. IQ gains are often temporary, but a gain that lasts for several years after a program enables increased academic achievement and learning during that time (with lasting benefits), and some gains are significantly more long-lived. More importantly, life outcomes such as employment, use of public assistance, and crime can be dramatically affected. The Obama administration has promised to and is likely to roll out extensive programs along these lines, but the key to success will be rigorous and continuous evaluations that are taken into account in expanding some programs and paring back others.
Chapter 8 covers the high achievement of people of Northeast Asian descent (Japanese, Chinese, Korean).
Read more ›
