From: dks@mit.edu
Date: Wed Aug 2 04:59:17 1995
Straightening Out "The Bell Curve"
[...]
Kalina:
> dks@eagle.mit.edu wrote:
> : ataylor@superior.carleton.ca (alex taylor) writes:
> : > Mark LaRochelle wrote:
> : > >ataylor@superior.carleton.ca (alex taylor) wrote:
> : > >> [...] The most significant indicator of success in
> : > >> the U.S. is the social economic class of the parents.
> : > > Dead False. Parental SES was the strongest correlate to final
> : > > SES in the US in the past, and still is in most other
> : > > countries, but has recently been surpassed in the US by what is
> : > > now a much stronger correlate: cognative ability, as measured
> : > > by IQ tests and the like.
> : > > The statistics are in Charles Murray and Richard Herrnstein,
> : > > *The Bell Curve*
> : > Citing a book that was panned by responsible social scientists
> : > and researchers of other stripes for its poor methodology and
> : > pseudo-scientific conclusions hardly constitutes a refutation.
> : > I've read the book in question. It's vile trash.
> : It is trash.
> : And even if it were not vile, it would still be trash. [...]
>
> Although I have read most of the articles around critiquing _The
> Bell Curve_ (as well as Murray's articles defending it), I have
1 > not read the book, and most of the critiques strike me as long on
> moral outrage but short on scientific criticism. The review
> provided by Dhanesh seems of a kind with these, and I am
3 > sympathetic to Murray's protestations that his work has been
> misrepresented by his critics. I will reserve final judgement
> until I have read the book itself.
1. Aren't we fortunate that Kalina, who has not read the book
itself, can judge how unfair "most" of the *reviews* are?
2. Notice that Kalina talks about all these reviews he's read
but gives no details as to which ones, nor does he justify
his opinion of them; the result is an attempted slur against
the reviews, not justified; and a weak not-quite-defense of
a piece of trash, which Kalina admits he has not read.
3. Why am I not surprised that Kalina is "sympathetic to Murray's
protestations..."? Why am I not surprised that he has read
Murray's "protestations" but not Murray's *book*?
Kalina:
> However, in light of the original klaim, I would ask the
> following:
The original claim can wait a few moments. I was responding to the
previous writer's remark that the book by Murray and Herrnstein is
trash. As I recall, the previous writer did not provide any basis
for the remark. I agree with the remark, and stated my agreement,
so I provided a review that reflects my assessment. I can easily
provide others as well.
It's called substantiating a point of view. I do it all the time.
It's not a trade secret. Other people do it, too. Kalina should
perhaps try it at some point. It's an innovation he could use.
Shell's critique of Murray and Herrnstein is a critique of their
book and their ideas; and of the reception the book received in the
mass media. And that's how it was presented here. Period. Some
may find it useful; others may not.
Whereas one wonders what Kalina is playing at, deriding reviews he
does not name, while not-quite-defending a piece of trash he has
not read.
Perhaps it is Kalina's hitherto unexpressed concern for fairness,
and his silent desire that all authors should be treated fairly.
But if so, why do counter-examples spring to mind that quickly
demolish this thesis?
> While the book may be trash, does it follow that we may discount
> any data contained therein? [ Stuff deleted. -- dks ]
Kalina makes various unsupported assertions in these paragraphs, and
I see no reason to discuss them, least of all with *him*. As for his
*question*, still quoted here, it's a good question, and *should*
be taken seriously. An answer follows.
Most of Murray and Herrnstein's so-called "data" are from one study.
To base a "thesis" as Ambitious as theirs largely on "data" from one
source is -- in itself -- not exactly a sign of great intelligence.
But it's worse: the one study they mostly use -- the National
Longitudinal Survey of Youth -- happens to be based on a sample
that is *not particularly representative* of the population of the
United States. So, even if one has no comment on their argument or
their book, and only wants to use their "data" in some other
context, it is not so simple. The "data" are poorly selected and
not always accurately presented. Instead of trying to use the book
as a reverse-engineered source of "data," it would be much wiser
simply to refer to the original sources.
But that's if we want to try to use the "data" elsewhere. Whereas
looking at the "data contained therein" in the context of the book
itself, the picture is *even worse*. The question is: why do Murray
and Herrnstein contrive to *exclude* data that do not support their
thesis? The authors deliberately provide only the narrowest selection
of the data, and even in doing so they misrepresent and distort.
Some people reading the book may get the impression that its claims
can be justified. But anyone familiar with the technical literature
in the field knows that the authors have ignored, or dismissed, or
wished away the evidence in studies that do not support them. And
what's more, they have drawn "conclusions" from studies that the
original writers did not see fit to draw. This kind of picking and
choosing from among the available data -- rejecting the inconvenient,
manipulating only that which is expedient -- this is not science.
In short, from assumptions to data to argument to conclusions, the
book is trash. It was written not to contribute anything useful to
science -- the very idea is laughable, because there is absolutely
nothing new in the book. As Shell says in her review, it is but
"a rehashing of an old, worn-out discussion."
No, there is nothing scientific about what Murray and Herrnstein
have produced. It was produced not for scientists, but for the
media. It was a *political act*, and it should be understood and
studied as such. Shell's review is useful in this regard, because
she analyzes the reception that the book received in the mass media.
----
By the way, I _have_ read the book. It is a sly document. Kalina
now comes along and describes the previous review (Shell) as being
"long on moral outrage but short on scientific criticism." This is
abject nonsense, because the review identifies and explains a number
of *critical* mistakes in the book, including:
1. The authors make questionable assertions and then assume them
to be factual, while providing no basis for such assumptions.
2. The authors "commit the cardinal sin of statistical analysis
-- they set out to prove, not disprove, their hypothesis."
3. The authors take "many gleeful leaps of logic."
4. The authors frequently use the "trick of putting a complex
question before the reader and then presenting a possible
solution as though it were the definitive one," while
never mentioning or considering better answers.
5. "[Only] a modest majority of social scientists think genetics
as well as environmental factors influence intelligence. The
Bell Curve, by contrast, contends that genetics swamps
environmental effects..."
6. "Social scientists are not so sure that intelligence can be
measured by any single test, or that it should be."
7. "Many [social scientists] are sure that [intelligence] cannot
and should not be measured [as Murray and Herrnstein claim to]."
8. "In any case, even psychometricians, psychologists who make
their living quantifying intelligence and designing intelligence
tests, would not agree that intelligence based on a single
measure (what The Bell Curve calls "g") can prove one group's
intellectual superiority over another."
These criticisms are not "moral outrage."
Murray has no good answer, so he *whines* about alleged and mythical
"political correctness," in a transparent attempt to garner even more
attention in the media.
Similarly, Kalina thinks he can deride Shell's review, and indeed "most"
reviews, with such phrases as "long on moral outrage," perhaps because
he thinks most people have no capacity to feel sincere moral outrage.
But I believe Kalina is *wrong* about that.
And in any event, "moral outrage" is nowhere in the objections listed
above. They are all based on (i) scientific methodology, and (ii) the
state of research into human intelligence. Each objection points to
one or more mistakes in the book. Any one of these mistakes occurring
alone might have been enough to get the book thrown out of a reputable
graduate seminar. Taken together, they are an *indictment*.
And Shell also cites three other reviewers who have analyzed the book
as if it were science -- they shred it to bits. Here's a line from
Stephen Jay Gould's review in the _New_Yorker_ (19941128, p. 139):
"The central premise is false, and most of the foundations are."
So for Kalina to say that the reviews are "long on moral outrage but
short on scientific criticism" is only to reveal how little he knows
about *both* moral outrage *and* scientific criticism.
But aside from being abject nonsense, Kalina's attempt at slur is an
obvious *device*. Notice that he could not be bothered to refute
even a single statement in Shell's review; he simply deletes the whole
thing, and "spins" it in the abstract -- along with "most" reviews --
using his handy and cynical cliche. Again, what we we see is not an
intellectual assessment, but only *derision* -- a raw *political* act.
But it's no use telling Kalina to be ashamed of himself; no use at all.
----
Shell's review stands. It is devastating. I recommend it.
Here are *additional* observations:
#1. The authors make an elementary mistake in statistics when
they suggest that causal inferences can be drawn from
relationships that are only correlations at best. If
it can be shown that variable X is correlated with -- i.e.,
occurs in a certain relationship with -- variable Y, then we
have a correlation. But it is not evidence that X causes Y,
or that Y causes X -- for example, it may well be that a
third unobserved variable Z is the underlying "control";
such that it is a change in Z that actually *causes* both
X and Y to change in correlation with one another. This
point is elementary, but it does not impress the authors,
who seem to think that making causal statements on the basis
of correlations is "a reasonable thing to do."
#2. The authors compare various correlations in order to "prove"
their thesis that such-and-so is the "cause" of thus-and-such.
But their comparisons are misleading, because it is not the
case that all variables are meant to do the same work, nor
that all variables are equally precise measures of the
underlying trait. An honest analysis would have depended
less on the poor operationalization of competing measures;
at least it would have discussed the effect of differences
in precision when measures are compared. An even more honest
analysis would have included better measures in the comparison
to begin with.
#3. The authors have a habit of saying one vaguely correct thing
in one usually obscure place, but then saying -- suggesting
-- another misleading thing in a much more prominent place.
The result: readers are invited to draw false conclusions,
but reviewers who point out the mistake are quickly dismissed
in printed "rebuttals," with a "reminder" that the authors
did, indeed, state things correctly -- in footnote 37 of
page 100b or some such.
#4. The authors frame their book as if there were a consensus
on key points in their discussion: (1) that IQ is a real
and distinct thing; (2) that IQ is heritable; (3) that there
are differences in "group" IQ; and (4) that the previous
three points really do matter. Well, there is, of course,
no such consensus. As we saw even in Shell's review, each
of these points is debatable, and is, in fact, debated in
the literature -- but no sign of such debate in this book.
#5. The authors present the results of "analyses" they claim
to have performed, in which they (1) select a number of
economic and social "behaviors"; and (2) compare IQ against
the socio-economic status of parents, as a means of (3)
discerning which of the two is the "cause" (see #1 above)
of aforementioned "behaviors." But *not one iota* of this
"research" has ever been published in in any peer-reviewed
scientific journal. Instead, it is put in print for the
mass media, and for the "ordinary reader," neither of whom
have the training with which to assess the quality [sic] of
the analysis. The reason for this choice of venue seems
obvious: Murray and Herrnstein probably do not care what
(other?) scientists think of this book. It seems they are
more interested in getting their "theory" out into the
"marketplace," where it can have all kinds of happy effects
on political reality. Responsible scientists abhor such
actions. But _The_Bell_Curve_ is not science -- it is the
deliberate disposal of toxic material in the public domain.
#6. The authors are Concerned about what they claim is a downward
trend in IQ in the the United States, which they seek to explain
by positing what they call _dysgenic_pressure_. Presumably,
they mean that people who happen to do poorly on standardized
tests also happen to have more children than other people. Or
rather, they speak not so much about people, but about "low-IQ
women." And therefore since [blah-blah-blah], we can conclude
that the overall level of IQ is decreasing. Unfortunately for
the authors, there are data -- widely known in the field --
suggesting that IQ scores have generally been *rising steadily*.
To "explain" the discrepancy, the authors cook up _ad_hoc_
explanations that make very little sense.
A note on the sources of this supposed _dysgenic_pressure_:
the authors cite and discuss the work of another writer, named
J. Philippe Rushton, who holds the view that "racial" groups
do not all make use of the same so-called "reproductive
strategies." Some "races," according to Rushton, have much
"larger" genital organs than others, and also enjoy a higher
frequency of sexual intercourse. Having cited and discussed
Rushton's views and work in relation to their own, the authors
note that "time will tell whether [Rushton's explanation] is
right or wrong in fact."
#7. Readers ought to realize that "heritability" is not a
simple variable that we can measure once for each trait
and then take for granted as defined forever. For any
given trait, "heritability" is not necessarily the same
for different populations, or even for the same population
at different times. Moreover, an estimate of "heritability"
requires that the underlying trait be defined and measured;
so if the trait is itself not easy to measure, even less
reliable is the estimate of "heritability."
#8. Even for "heritable" traits, we should not make the
mistake of assuming that their appearance in phenotype
is determined "in the genes" and cannot be changed. Many
studies indicate that IQ, for example, can be improved
if we pay attention to the environment in which a child
is challenged to learn. These data are tentative, and
the actual techniques have yet to be worked out, but
-- unless we have political agenda -- it seems unwise to
assume that an estimate of "heritability" is the end of
the story for any given trait.
#9. Even if IQ is a good predictor of certain (other) outcomes,
this result need not be an immutable fact of nature. For
example, suppose that IQ is a reliable predictor of
lifetime earnings. We should ask *why*. The answer might
well be that we have devised a system in which it is
necessary to score well on the various standardized tests
in order to obtain the "right" training, so as to pursue a
lucrative career. We may have set things up in such a way
that the system preferentially rewards those who happen to
have test-taking skills that help them to score well on all
these *artificial* standardized tests. This kind of thing
is hardly a necessary consequence grounded in genetics: to
change it, we would only have to change the basis upon which
so many "educational" and professional hurdles are designed.
At that point, the system might begin to reward true
achievement, or a broad spectrum of skills, rather than
those who happen to do well on purely artificial tests.
The following excerpt makes the same point more effectively:
|
| Letter to the Editor,
| _New_York_Times_Book_Review_
| 13 November 1994
|
| By Paul Grobstein
|
|
| [...]
|
| In short, I do not, from my direct experience, have
| any reason to believe standardized examination scores
| are strongly predictive of anything but future
| standardized examination scores. My direct experience
| does, however, say that examination scores, in a
| culture preoccupied with examination scores, are
| likely to be at least weakly predictive of the extent
| to which people feel valued by, and comfortably
| affiliated with, their community. Looked at from this
| perspective, it is no surprise that standard tests
| are not bad predictors of job performance (and,
| inversely, of criminality), and this weak correlation
| provides no evidence whatsoever for assertions that
| "g" is an essential survival characteristic in any
| general sense at all.
|
| To put it slightly differently, the same correlation
| can be read either as an indication that evaluation
| by some single criterion suffices to predict success,
| or as an indication that the use by our culture of a
| single standard for evaluation makes lots of people
| unhappy enough so that they fail to become satisfied
| and productive members of our culture.
|
| [...]
|
For those who wonder who it is that I've just quoted, I should
say that Paul Grobstein is the Eleanor A. Bliss Professor of
Biology at Bryn Mawr College, Bryn Mawr, Pennsylvania; and
author of more than fifty publications on the development of
the nervous system, the organization and function of the brain
in relation to behavior, and theoretical biology.
#10. Which brings me to my last point. I noted earlier that the
"analysis" underlying the "conclusions" in _The_Bell_Curve_ has
never been published in a peer-reviewed scientific journal. But
the story is actually worse. Much worse. Neither Herrnstein nor
Murray has *ever* published *even one* article on the genetic
basis of human intelligence in any peer-reviewed journal. Their
"publications" on this subject -- such as they are -- have all
been in popular magazines, and popular books that seem to be
aimed not at the scientists who could properly judge the work,
but at the mass media.
Herrnstein was a professor of psychology at Harvard University
for nearly forty years. He died recently. While the media
often portray his role in this book as a Culmination of his life's
work in a relevant field, this picture is false. Herrnstein
studied rats and pigeons, not human intelligence; and even then,
he never studied the genetic basis of behavior in these animals.
So his "thoughts" on the link between human intelligence and
behavior, his "thoughts" on the genetic basis of IQ and its alleged
relation to poverty, social class, or the construct we call "race"
-- to the degree that his "thoughts" on these subjects appear in
this book -- they are not the summation of years of research.
What they actually *are* is difficult to say.
We *could* say more about Herrnstein's career, and about one or
two of his previous co-authors -- Jensen, for example, -- but time
is short. We could also say something on the subject of Herrnstein's
intellectual honesty -- and he, *perhaps*, on the subject of ours,
were he alive -- but there's no time for that now, either. For
details, the reader is invited instead to study the publication
by Dorfman that is cited and quoted in the appendix.
Murray, the second author of _The_Bell_Curve_, is not even trained
in psychology, let alone genetics. He is a political scientist,
with degrees from Harvard and MIT. His publications to date have
appeared most often in neoconservative magazines. His agenda are
the usual "neoconservative" agenda. He is by no means a cognitive
scientist. Not once has he conducted research into human (or any)
intelligence; nor has he ever published in any scientific journal
the results of any research on the genetic basis of IQ and poverty.
In short, when it comes to the subject of the book he claims to
have co-authored, Murray does not have the training, nor has he
ever displayed the skills, that might encourage scientists to take
his pronouncements as science.
So, we had Shell's observations, and now we have a few more. I daresay
these observations will also be described as "long on moral outrage and
short on scientific criticism." I trust the reader is able to see such
comments for what they are.
And realize: these are just some of the more general objections I could
have raised. There are many other things in the book that many people
find objectionable. For example:
The authors *themselves* make a *moral judgment* when they
choose to focus on the economic value of individual human
beings, and ignore all other aspects. Proper opposition
to this chosen focus *must also* be presented in terms of
moral values. For those who do not believe that the worth
of a human being has much to do with how well that human
being can be exploited in an economic system, this book by
Murray and Herrnstein is not outrageous -- it is pointless.
This, then, is what a "moral" objection looks like. It speaks of what
is right and wrong in our lives. It asks us to think about the kind of
people we would like to be. *This* is morality.
Meanwhile, there is the book. It exists. It is politics as we know it.
We should read the book, read the reviews, talk to people who know the
subject, and *judge the book for ourselves*. If we can't make a firm
judgment, that's OK. Just *reserve judgment*. There is no law that
says we all have to express an opinion on every subject. Knowing when
to simply be quiet is an extremely rare form of wisdom -- we should
cultivate it.
But above all, whatever we do, we should not let people tell us what
to think who show *no sign of being able to think for themselves*.
That way lies madness and danger.
Dhanesh
The following book review, which suggested a few of the comments that
appear above, also appears in full in the May 1995 (40:5) issue of
_Contemporary_Psychology_; published by the American Psychological
Association. Only an excerpt is presented below, without permission.
(A more favorable review of the book appears alongside in the same
issue of the journal; it concludes: "This is a superbly written and
exceedingly well-documented book." I could review the review, but
I'm not going to bother. Anyone who wishes to see or argue its point
of view now has the citation (above), and is free to pursue it.)
|
|
| Soft Science With a Neoconservative Agenda
| ..........................................
|
| Review by Donald D. Dorfman
|
|
| Richard J. Herrnstein and Charles Murray
| The Bell Curve: Intelligence and Class Structure in American Life
| New York: Free Press, 1994. 845 pp. ISBN 0-02-914673-9. $30.00
|
|
| [...]
|
| _The_Bell_Curve_ is not a scientific work. It was not written by
| experts, and it has a specific political agenda. Still, it is
| possible that the major scientific premises of the book may be
| correct. If two monkeys were put before a typewriter, it is
| theoretically possible for those two monkeys to produce a
| Shakespearean sonnet. Perhaps Herrnstein and Murray produced a
| valid scientific work. I will now evaluate the major premises of
| _The_Bell_Curve_.
|
|
| The rewriting of history: The Burt affair
| .........................................
|
| In 1972, Leon Kamin exposed the empirical unsoundness of the most
| important evidence in support of the IQ hereditarian position, Sir
| Cyril Burt's data (Hearnshaw, 1979). He later published his results
| in a book attacking Burt's data as well as the secondary sources
| who publicized those data (Kamin, 1974). In 1979, Leslie Hearnshaw
| (1979) published a biography of Burt in which he concluded on the
| basis of personal diaries and other material that it was highly
| likely that Burt had fabricated some of his most celebrated data.
| Hearnshaw, distinguished historian of British psychology, delivered
| the memorial address at Burt's Memorial Service and was later asked
| by Marion Burt, Burt's sister, to write a full-length biography of
| Burt. The result was the well-known _Cyril_Burt:_Psychologist_
| (1979). In their discussion of the Burt affair, Herrnstein and
| Murray suggest that some of Burt's "leading critics were aware that
| their accusations were inaccurate" (p. 12), suggesting a possible
| conspiracy against Burt. There is, however, no mention whatsoever
| of Hearnshaw's book in their half-page synopsis of the Burt affair,
| and Hearnshaw's book does not appear anywhere in their 57-page
| bibliography of references. This misrepresentation of the Burt
| affair by omission of important historical facts is not uniquely
| associated with _The_Bell_Curve_. In 1982, Richard Herrnstein
| published an article in _The_Atlantic_Monthly_ in which he attacked
| the media for misrepresenting the evidence in the IQ controversy
| (Herrnstein, 1982). In that magazine article, the Harvard professor
| wrote "that most psychometricians had stopped trusting Burt's data
| years before, partly because of inconsistencies first noted in a
| 1974 article by Arthur Jensen" (p. 70), while omitting any mention
| of Leon Kamin, the psychologist who in reality first noted
| inconsistencies in Burt's data.
|
|
| Does the distribution of IQs follow a bell curve?
| .................................................
|
| The distribution of IQ test scores cannot be expected to follow a
| bell curve unless it is constructed by the tester to do so
| (Dorfman, 1978). The shape of the distribution of IQ test scores
| will depend on the average difficulty of the test items as well as
| their intercorrelations. The high item intercorrelations in IQ
| tests imply that the IQ distribution can take a variety of shapes.
| The central limit theorem does not apply to random variables with
| positive intercorrelations (Lamperti, 1966). Frederic Lord (1952),
| one of the fathers of modern test theory and former president of
| the Psychometric Society, gave results on this question: "The
| results given are sufficient to show that the distribution of test
| scores cannot in general be expected to be normal, or even
| approximately normal. The question naturally arises as to what
| possible shapes the frequency distribution _fs_, as given in (76)
| [Lord's Equation (76)], may assume. The answer is that this
| function may assume any shape whatsoever, provided the item
| intercorrelations are sufficiently high" (Lord, 1952, pp. 32-33).
| The symbol _fs_ refers to the distribution of test scores.
|
|
| Does cognitive ability consist of a single general factor?
| ..........................................................
|
| The book uses factor analysis to infer the existence of a single
| hypothetical general factor of cognitive intelligence that is
| presumed to account for most of cognitive performance. One of the
| problems with factor analysis as a tool for determining the
| underlying structure of a system is that neither the factors nor
| the loadings are uniquely defined if you have more than one factor
| (Lawley & Maxwell, 1963), and it is difficult to determine if you
| have only one factor. In experimental cognitive psychology, factor
| analysis is virtually never used as a tool to determine the
| underlying cognitive structure. It is a tool for correlational
| cognitive psychology, not experimental cognitive psychology. I
| inspected the subject index of some well-known texts in
| experimental cognitive psychology and found that the term factor
| analysis never appears in the subject index (e.g., see Anderson,
| 1985; Matlin, 1994; Reed, 1982). Why not? Kendall and Stuart (1966)
| may provide the answer: "Application of the same technique [factor
| analysis] to physical systems very often results in weighted sums
| of variables to which no clear interpretation can be given" (p.
| 310). In short, "The main difficulty, as a rule, is to know what
| the results mean" (p. 310), Kendall and Stuart point out.
|
|
| Can you measure the heritability of IQ?
| .......................................
|
| The most direct way of estimating heritability is from data on
| monozygotic twins reared apart (MZA) and separated in early infancy
| (Bouchard, Lykken, McGue, Segal, & Tellegen, 1990). This MZA design
| allows for the estimation of heritability if the following major
| assumptions are met: (a) environments are a random sample from the
| population of environments, (b) genotypes are a random sample from
| the population of genotypes, (c) there is no genotype-environment
| correlation, and (d) there is no genotype-environment interaction.
| If the pairs of MZAs differ in age, then these assumptions will not
| be met. If these assumptions are met, then the intraclass
| correlation between IQ scores of MZA twin pairs directly measures
| heritability. Sir Cyril Burt's (1966) study of 53 MZAs appears to
| have met the first three assumptions. Unfortunately, Burt's data
| appear to have been invented (Hearnshaw, 1979). Bouchard et al.'s
| (Minnesota) survey of MZAs provides the next best data set.
| Unfortunately, to the best of my knowledge, the detailed case-study
| records of the Minnesota MZAs have never been released and have
| therefore not been subjected to public scrutiny to determine the
| degree to which assumptions have been met and the degree to which
| the MZAs told the truth to the Minnesota group. Finally, if there
| is genotype-environment interaction--then the fourth assumption is
| not met--and heritability is undefined. But this is the most
| controversial assumption underlying the MZA design. Herrnstein and
| Murray present no convincing evidence to justify the fourth
| assumption.
|
|
| Does high within-group heritability of IQ imply between-group
| heritability of IQ?
| ...................
|
| The authors have made a fundamental error well-known by
| professional geneticists. It is sometimes called "Jensen's error."
| Jensen made that error in his famous 1969 _Harvard_Educational_
| _Review_ article. The critical importance of that error was first
| clearly illuminated by Roger Milkman, a professor of biology at the
| University of Iowa and a world authority on population genetics and
| evolutionary biology. The article, "A Simple Exposition of Jensen's
| Error," was published in the _Journal_of_Educational_Statistics_ in
| 1978 (Milkman, 1978). Melvin Novick was editor of that journal when
| Milkman's article was published. Novick, professor of statistics
| and education at the University of Iowa at the time, later became
| president of the Psychometric Society. What is Jensen's error? It
| is that within-race heritability has no implications for
| between-race heritability. _The_Bell_Curve_ is therefore flawed
| regard to inferring between-race heritability in IQ from
| with within-race heritability in IQ.
|
|
| Does IQ or SES cause socially undesirable behaviors?
| ....................................................
|
| Herrnstein and Murray use logistic regression to determine which is
| more important--IQ or SES--in determining socially undesirable
| behaviors. Logistic regression is a form of regression in which the
| dependent variable is binary. In all of their analyses, they assume
| a simple additive model in which the logit (a transform of the
| sample proportion) is assumed to equal B0 + B1IQ + B2SES + B3 age +
| random residual [numbers after Bs should read as subscripts]. They
| assume no IQ-SES interaction. They use the standardized beta
| weights to determine the relative importance of IQ and SES in
| determining the probability of various undesirable or desirable
| behaviors. Unfortunately, IQ and SES are highly intercorrelated
| (collinearity).
|
| There are two major problems with Herrnstein and Murray's attempts
| to determine whether IQ or SES is more important. First, there is
| the collinearity problem. Weisberg (1985) describes the
| collinearity problem in linear regression: "When the predictors are
| related to each other, regression modeling can be very confusing.
| Estimated effects can change magnitude or even sign depending on
| the other predictors in the model" (p. 196). Next, there is the
| problem of deciding that the predictor with the largest
| standardized beta weight is the most important. Weisberg describes
| why this approach is faulty: "Unfortunately, this logic is faulty
| because the scaling depends on the range of values for the
| variables in the data" (p. 186). Perhaps these are the reasons why
| Herrnstein and Murray never published their logistic analyses in
| peer-reviewed journals.
|
| Were Herrnstein and Murray as lucky as the proverbial monkeys at a
| typewriter? That depends on your point of view.
|
|
| References
| ..........
|
| Anderson, J. R. (1985). _Cognitive_psychology_and_its_implications_
| (2nd ed.). New York: W. H. Freeman.
|
| [...]
|
| Bouchard, T. J., Lykken, D. T., McGue, M., Segal, N. L., &
| Tellegen, A. (1990). "Sources of human psychological differences:
| The Minnesota study of twins reared apart." _Science_, 250, 223-228.
|
| [...]
|
| Dorfman, D. D. (1978). "The Cyril Burt question: New findings."
| _Science_, 201, 1177-1186.
|
| [...]
|
| Herrnstein, R. J. (1982, August). "IQ testing and the media."
| _The_Atlantic_Monthly_, 68-74.
|
| [...]
|
| Kamin, L. J. (1974). _The_science_and_politics_of_I.Q._
| New York: Wiley.
|
| Kendall, M. G., & Stuart, A. (1966). _The_advanced_theory_
| _of_statistics_ (Vol. 3). New York: Hafner.
|
| Lamperti, J. (1966). _Probability:_A_survey_of_the_
| _mathematical_theory_. New York: W. A. Benjamin.
|
| Lawley, D. N., & Maxwell, A. E. (1963).
| _Factor_analysis_as_a_statistical_method. London: Butterworth.
|
| Lord, F. M. (1952). "A theory of test scores."
| _Psychometric_Monographs_ (Whole No. 7).
|
| Matlin, M. W. (1994). _Cognition_ (3rd ed.).
| Fort Worth, TX: Harcourt Brace.
|
| Milkman, R. (1978). "A simple exposition of Jensen's error."
| _Journal_of_Educational_Statistics_, 3, 203-208.
|
| [...]
|
| Reed, S. K. (1982). _Cognition:_Theory_and_applications_.
| Monterey, CA: Brooks/Cole.
|
| Weisberg, S. (1985). _Applied_linear_regression_ (2nd ed.).
| New York: Wiley.
|
|
![[Three for All]](http:../gif/trilogo-c.gif)
![[Mens et Manus]](http:../gif/mitlogo-r.gif)