Bell Curve: Intelligence and Class Structure in American Life (A Free Press Paperbacks Book) 1st Free Press pbk. ed Edition

Since you're reading this, I assume you're thinking of buying — or at least reading — this book. That being so, you'll probably want to read other reviews than mine. This is in principle a good idea; but having just read all of them (147 at the time of writing) I should warn you that you'll need both considerable stamina and a strong stomach: there are indeed thoughtful and informative reviews, but they are islands in a sea of drivel. By "drivel" I mean the following:

1) Reviews consisting entirely (or almost entirely) of expostulation rather than information ("racist garbage", "most important book of the 20th century")

2) Asserting what the book doesn't deny and denying what it doesn't assert.

3) Distortions of the book's content, and other disinformation, for instance:

- "the panel criticized the authors for not explaining what intelligence is" (intelligence is defined on page 4 (!) ).

- "The Bell Curve ignores bad diet" (Nutrition is explicitly dealt with on pp. 391-3).

And so on.

Many of the critics appear not merely to have misunderstood the book, but not even to have read it; amusingly, this is actually admitted in one review ("Although Head has only browsed through the book, she has seen this kind of pseudo-science before"). Some appear to be basing their argument upon the Moralistic Fallacy: if different groups had different average IQs for even partially genetic reasons, it would be a Bad Thing, and therefore that cannot possibly be the case.

(It also follows, naturally, that there's no need to examine the evidence, because the correct outcome is known in advance; and also that those who think otherwise must be making factual and/or logical errors, either unconsciously or consciously, making them fools in the former case and fascists in the latter.)

For myself, I found this a strange book in some ways, but only one other reviewer (Jennifer Kerns, I think the name was) touched on the reason. And that is that the book falls logically into three parts, which by their very nature are of varying reliability.

The first, and by far the largest, covers the available evidence on IQ and heredity. The second and third parts extrapolate present trends to the future (with unpleasant consequences) and make policy recommendations to deal with these projected consequences. Thus almost by definition these are on shakier ground.

- The first section, which excited by far the most controversy, is (ironically), easily on the firmest ground scientifically — as confirmed (for example) by an American Psychological Association task force explicitly set up to investigate it*; and by a letter to the Wall Street Journal by fifty-two leading psychometricians, a copy of which can be found on the Net ("Mainstream Science on Intelligence", also reprinted as an appendix in H.J. Eysenck's Intelligence: A New Look).

It seems to me a very able summary: it defines its terms, states its assumptions, produces its evidence and argues the merits of the various theories purporting to explain it. So there's no need for you to take my word (or anyone's) as to whether the thesis is justified; the evidence and the arguments are both there; if you're capable of rational thought, you should be able to decide for yourself. And this is what I advise you to do.

- The second part envisages the potential stratification of society by intelligence into a hereditary élite and underclass. Here the authors start to part company with some (at least) of the aforementioned psychometricians. Eysenck, for instance — certainly in the "hereditarian" camp as regards IQ — writes of an earlier article in Atlantic Monthly:

"Here Herrnstein is definitely beginning to run off the rails in his predictions (...) he disregards the importance of regression, the genetic factor which causes children of very bright and very dull parents to regress towards the mean of the whole population (...) [R]egression makes it quite impossible that castes should be created which will breed true — that is, where the children will have the same IQ as their parents. Within a few generations, the differences in IQ between the children of very bright and very dull parents will have been completely wiped out." (The Inequality of Man, pp.213-219)

Richard Lynn, however, disagrees, pointing out that if regression operated in all cases, then dog-breeding, and indeed evolution as a whole, would be impossible.

- The third part, the policy recommendations, is well outside my area of competence, so I offer no comment.

I should, however, like to make one further comment on other reviews, those containing the recommendation: "People wanting an honest scientific analysis of the claims of racial superiority should read Stephen Jay Gould's The Mismeasure of Man".

Gould's writing certainly has many admirable qualities, but honesty and scientific impartiality are not conspicuous among them — for specifics, see (for example) Chapter 3 of John L. Casti's Paradigms Lost. Or see J. Philippe Rushton's review of "Mismeasure", or Arthur Jensen's review ("The Debunking of Scientific Fossils and Straw Persons"), or John B. Carroll's, all of which you can find on the Web.

I've been following the debate over IQ for 40 years, and The Mismeasure of Man has more factual errors per page than any book I've ever read.

For a critical but still rational review of Herrnstein & Murray, I suggest Thomas Sowell's from American Spectator, which can also be found on the Web ("Ethnicity and IQ").

If you want a balanced account of the IQ field, try Intelligence: The Battle for the Mind, half of which is written by H.J. Eysenck and half by Leon Kamin, with a final rejoinder from each. The best summary I'm aware of remains, despite its age, H.J. Eysenck's The IQ Argument (Race, Intelligence and Education in the U.K.); but good luck getting hold of it!

*UPDATE 2007

I should have said that although the APA report could not (or at least did not) explicitly rebut any of Herrnstein & Murray's data, or their logic, it refused to endorse their conclusions.

I haven't changed the body of the review because that would make nonsense of the discussion in the Comments.

For a more detailed factual account of the tactics of Gould et al, I recommend Ullica Segerstråle's Defenders of the Truth, although I'm not sure I'm convinced by her psychological diagnosis.

Probably the best survey of the various issues and viewpoints that I'm aware of is The Race Gallery by Marek Cohn (1995).

In reading the synopses above and the few negative reviews below, I have to wonder if their authors read the book in question, or merely the media hype. This book is not about "ethnicity and intelligence." It's not racism, nor even about race.

This book tries to show that A) people are becoming stratified according to intelligence (you go to school with, work with, and largely socialize with people of similar ability) and B) many of our social problems can be explained in terms of differences in intelligence (ie, in blunt terms, dumb people are more likely to commit crimes, etc.) They provide a huge base of data to support their thesis.

The authors have bent over backwards to try to avoid any hint of racism in their studies; the only place the issue even arises is when they report that blacks and Latinos have historically scored lower in IQ tests than have whites (Asians have scored higher), and that the claims of "cultural bias" are not supported by any data or studies. These details alone are enough to inflame the politically correct among us, unfortunately.

To portray this book as some type of white supremecist manifesto, you would have to have a strong agenda of your own, and totally disregard the content of the book.

I thought I knew a lot about the topic. I stand corrected.

IQ is the most scientific and mathematical concept in all of psychology.

Every test that somehow measures cognitive processing capacity strongly correlates with g, general intelligence.

Studies with sample sizes of tens and hundreds of thousands clearly showed that IQ is the best predictor of success in basically all professions. Attempts to devise other tests to predict job success did not work out, as IQ is the best predictor and explains more variance then all other factors together.

The book meticulously cites all research in foot notes, explain scientific concepts such as correlation and factor analysis. Still manages to be understandable to the educated layman.

The authors were aware that their scientific findings would be met with strong political resistance. They were right.