[Paleopsych] Meme 041: A Complacent AntiRacist

[Premise Checker]

Meme 041: A Complacent AntiRacist
by Frank Forman
sent 5.5.19

I grew up on the myth that all scientific results had been confirmed in 
independent laboratories before being published. Cost alone, of course, 
precludes any such thing, but controversial results do attract 
further examination. Cold fusion is an excellent example: while the 
initial experiments of Fleishmann and Pons are widely rejected as flawed, 
several scientists have conducted later experiments and think there may be 
something out there, though they don't agree on what it might be.

On the other hand, there are a number of experiments that back up what a 
lot of people would like to believe in but others would not that are not 
followed up but get endlessly cited. These results stand alone.

We should develop a list of them.

One is the Pygmalion Effect, where elementary school children were 
randomly assigned to high and low IQ groups but with the teachers knowing 
their supposed IQs. The unduplicated result was that children assigned to 
the high IQ group vastly outperformed those assigned to the low IQ group. 
The book is _Pygmalion in the Classroom: Teacher Expectation and Pupils' 
Intellectual Development_ by Robert Rosenthal and Lenore Jacobson and came 
out in 1968, the original experiment having been done a few years earlier.

Another unduplicated experiment was made by measuring the penis arousal to 
erotic pictures, wherein "homophobes" responded more to pictures of 
homosexual acts than the rest. (I can give no citation for this, perhaps 
because there never was such a study.) And another is Daniel Keren, Jamie 
McCarthy, and Harry W. Mazal, "The Ruins of the Gas Chambers: A Forensic 
Investigation of Crematoriums at Auschwitz I and Auschwitz-Birkenau," 
_Holocaust and Genocide Studies_ 18.1 (2004): 68-103. I know of no other 
article that remotely approaches this one in making a scientific case for 
Nazi gas chambers and welcome citations to other articles.

In the article below, the unduplicated study is that of the IQs of 
children of occupying soldiers in Germany after the end of WW II with 
white German women. The study found almost no difference between the 
children of white and black fathers. The study is Klaus Eyferth, 
"Leistungen verschiedener Gruppen von Besatzungskindern in 
Hamburg-Wechsler Intelligenztest fur Kinder (HAWIK)," Archiv fur die 
gesamte Psychologie 113 (1961): 222-41. I googled the author's name and 
found out that he is now an emeritus professor and had studied cognitive 
modeling and similar such technical (though by no means unimportant) 
specialties, but nothing so sensational as this. One site claims that the 
article was published in Vita Humana (now Human Development) in 1959, 
which would place it before the German and so is dubious. All these 
journals do exist, and maybe the article is genuine, but it is odd that it 
has not been anthologized in English repeatedly and, indeed, not been more 
than fleetingly discussed. And, odd also that similar studies have not 
been conducted or at least not reported, which can happen if studies give 
unwanted results.

I charge complacency here. I know no one who has a material interest in 
differences in innate racial cognitive abilities. On the other hand, there 
is a potentially rich interest we all have in there being differences in 
approaches to the world and in temperament that are rooted in genetic 
differences, which are *presumably* less difficult to eliminate than those 
rooted in local historical circumstance (but reordering whole societies 
has proven to be very expensive in terms of unintended consequences, while 
fiddling with the genome may become very cheap). Those who think 
differently from white, middle-class Americans could offer novel 
approaches and solutions to problems and these abilities will *not* get 
steamrollered away as the world becomes more and more Americanized. Alas, 
when I inquire at gatherings to celebrate diversity about the concrete 
benefits of this diversity, all I get is ethnic cooking (but the Chinese 
are buying up Mexican restaurants in New York City and Mongolian Barbecue 
here has nary a Mongolian in sight. It is now opening a branch in Mongolia 
itself! See http://www.bdsmongolianbarbeque.com .) or ethnic folk dancing 
and folktales.

Now, it's been alleged that "racist" interests are served by keeping the 
black man down and exploiting him, but this goes against the logic of 
free market economics. On the other hands quite easy to see material 
interests in fostering the belief of equality of innate racial mental 
capacities: it gives resources to educators and other social planners. My 
charge is that these educators and planners, while operating out of their 
own idealistic visions at first, became dominated by those who just want 
to hold on to jobs.

In other words, they became complacent. If they were in fact concerned, 
while they might not want to hear any bad news that equality of outcome 
(almost always measured by money income, whence Mickey Kaus's term "money 
liberals") will be only narrowed not finally achieved, what would really 
be upsetting would be a finding that what they get paid to do has very 
little impact and that they are not needed. So let's not undertake any 
study to factor out genetics and BREAK UP THE VARIOUS ENVIRONMENTAL 
FACTORS INTO THEIR COMPONENTS. It may turn out that schooling is much less 
important than nutrition or removing lead paint. It is far, far safer to 
moan about bad schooling, bad nutrition, lead paint, etc., at equal 
volumes and to forestall any study that would break down the causes of 
racial differences in achievement into environmental components, 
hereditary components, and (let us not forget) free will, though the 
latter is elusive, at least as far as multiple regression studies go. You 
can claim that none of the independent variables (like IQ) means anything 
and that the populations studied (races) are meaningless aggregates, and 
(this never happens) that the dependent variable (money) is also 
meaningless.

But you remain complacent about any actual results.

The article below is the first one I have been able to locate to cite the 
J. Philippe Rushton and Arthur Jensen, "Thirty Years of Research on Race 
Differences in Cognitive Ability," Psychology, Public Policy, and Law 
(2005 July), but the author said the article was forthcoming and evidently 
had a preview copy.

[I am sending forth these memes, not because I agree wholeheartedly with 
all of them, but to impregnate females of both sexes. Ponder them and
spread them.]

--------------------

William T. Dickens: Genetic Differences and School Readiness
Future of Children 15.1 (2005) 55-69
http://muse.jhu.edu/demo/future_of_children/v015/15.1dickens.html

Abstract

The author considers whether differences in genetic endowment may
account for racial and ethnic differences in school readiness. While
acknowledging an important role for genes in explaining differences
within races, he nevertheless argues that environment explains most of
the gap between blacks and whites, leaving little role for genetics.

Based on a wide range of direct and indirect evidence, particularly
work by Klaus Eyferth and James Flynn, the author concludes that the
black-white gap is not substantially genetic in orgin. In studies in
1959 and 1961, Eyferth first pointed to the near-disappearance of the
black-white gap among children of black and white servicemen raised by
German mothers after World War II. In the author's view, Flynn's
exhaustive 1980 analysis of Eyferth's work provides close to
definitive evidence that the black disadvantage is not genetic to any
important degree.

But even studies showing an important role for genes in explaining
within-group differences, he says, do not rule out the possibility of
improving the school performance of disadvantaged children through
interventions aimed at improving their school readiness. Such
interventions, he argues, should stand or fall on their own costs and
benefits. And behavioral genetics offers some lessons in designing and
evaluating interventions. Because normal differences in preschool
resources or parenting practices in working- and middle-class families
have only limited effects on school readiness, interventions can have
large effects only if they significantly change the allocation of
resources or the nature of parenting practices.

The effects of most interventions on cognitive ability resemble the
effect of exercise on physical conditioning: they are profound but
short-lived. But if interventions make even small permanent changes in
behavior that support improved cognitive ability, they can set off
multiplier processes, with improved ability leading to more
stimulating environments and still further improvements in ability.
The best interventions, argues the author, would saturate a social
group and reinforce individual multiplier effects by social
multipliers and feedback effects. The aim of preschool programs, for
example, should be to get students to continue to seek out the
cognitive stimulation the program provides even after it ends. [End
Page 55]

In national tests of school readiness, black preschoolers in the
United States are not doing as well as white preschoolers. Researchers
find black-white gaps not only in achievement and cognitive tests, but
also in measures of readiness-related behaviors such as impulse
control and ability to pay attention. Could some of these differences
in school readiness be the consequence of differences in genetic
endowment? In what follows I will review research evidence on this
question.^1

Evidence on the Role of Genetic Differences

To evaluate the research findings on the role of genetic differences
in cognitive ability, I begin by drawing a clear distinction between
evidence that genetic endowment explains a large fraction of
differences within races and evidence that it explains differences
between races and ethnic groups. There can be little doubt that
genetic differences are an important determinant of differences in
academic achievement within racial and ethnic groups, though the size
of that effect is not known precisely. Depending on the measure of
achievement used, the sample studied, and the age of the subjects,
estimates of the share of variance explained by genetic differences
within racial and ethnic groups range from as low as 20 percent to
upward of 75 percent. However, most estimates, particularly those for
younger children, seem to cluster in the range of 30 to 40 percent.
The fraction of variance explained by genetic differences in a
population is termed the heritability of the trait for that
population.^2

But the heritability of academic achievement within racial or ethnic
groups says little about whether genes play a role in explaining
differences between racial groups. Suppose one scatters a handful of
genetically diverse seed corn in a field in Iowa and another in the
Mojave Desert. Nearly all the variance in size within each group of
seedlings could be due to genetic differences between the plants, but
the difference between the average for those growing in the Mojave and
those growing in Iowa would be almost entirely due to their different
environments.

If researchers were able to identify all the genes that cause
individual differences in school readiness, understand the mechanism
by which they affect readiness and the magnitude of those effects, and
assess the relative frequency of those genes in the black and white
populations, they would know precisely the extent to which genetic
differences explain the black-white gap. But only a few genes that
influence cognitive ability or other behaviors relevant to school
readiness have been tentatively identified, and nothing is known about
their frequency in different populations. Nor are such discoveries
imminent. Although genetic effects on several different learning and
school-related behavior disorders have been identified and many
aspects of personality are known to have a genetic component, genes
have their primary effect on school readiness through their effect on
cognitive ability.^3 Experts believe that a hundred or more genes are
responsible for individual differences in cognitive ability. Many of
these genes are likely to have weak and indirect effects that will be
difficult to detect. It could be decades before enough genes are
identified, and their frequencies estimated, to make it possible to
determine what role, if any, they play in explaining group
differences.

So it is necessary to turn to less direct ways of answering the
question. Much has been written on this topic in the past fifty years.
James Flynn's Race, IQ, and Jensen, published in 1980, remains the
most thoughtful and thorough [End Page 56] treatment available.^4 More
recently Richard Nisbett wrote a shorter review of this literature.^5
Both Flynn and Nisbett take the view, as do I, that genetic
differences probably do not play an important role in explaining
differences between the races, but the point remains controversial,
and Arthur Jensen provides a recent discussion from a hereditarian
perspective.^6 Here I will review the major types of evidence and
explain why I think they suggest that environmental differences likely
explain most, if not all, of the black-white gap in school readiness.
I will concentrate entirely on the evidence on cognitive ability, as
it is the most studied trait that influences school readiness, and
genetically induced differences in cognitive ability account for the
vast majority of genetically induced differences in school readiness
within ethnic groups. Almost no studies have been done of racial
differences in other traits that might influence school readiness. And
I choose to focus on the black-white gap rather than to consider the
role of genetic differences in determining the academic readiness of
disadvantaged groups more generally, again, because it is a topic that
has been more thoroughly studied.
[Box: Clearing Up a Confusion]

Direct Evidence on the Role of Genes: European Ancestry and Cognitive
Ability

Blacks in the United States have widely varying degrees of African and
European ancestry. If their genetic endowment from their African
ancestors is, on average, inferior to that from their European
ancestors, then their cognitive ability would be expected to vary
directly in proportion to the extent of their European ancestry. Some
early attempts to assess this hypothesis linked skin color with test
scores and found that lighter-skinned [End Page 57] blacks typically
had higher scores. But skin color is not strongly related to degree of
European ancestry, while socioeconomic status clearly is. Thus the
differences might reflect environmental rather than genetic causes.
Nearly all commentators agree that these early studies are not
probative.

More recent studies have looked at measures of European ancestry, such
as blood groups or reported ancestry, that are not visible. Such
studies have found little or no correlation between the measure of
ancestry and cognitive ability, though all are subject to
methodological criticisms that could explain their failure to find
such a link. Thus although these studies do not provide evidence for a
role for genes in explaining black-white differences, they do not
provide strong evidence against it.

Direct Evidence on the Role of Environment: Adoption and Cross-Fostering

If there is no direct evidence of a role for genes in explaining the
black-white gap, perhaps there is direct evidence that environment can
or cannot account for the whole difference between blacks and whites.
Several studies have shown that environmental differences between
blacks and whites can, in a statistical sense, "explain" nearly all of
the difference in cognitive ability between black and white
children.^7 But because the studies do not completely control for the
genetic endowment of either the child or the parents and because many
of the variables used to explain the difference are themselves subject
to genetic influence, the effect being attributed to environment may
in reality be due to genetic differences.

What is needed is a way to see the effect of environment without
confusing it with the effect of genetic endowment. For example,
randomly choosing white and black children at birth and assigning them
to be fostered in either black or white families would ensure that the
children's environments were not correlated with their genetic
potential and would show how much difference environment makes. No
existing study replicates the conditions of this experiment exactly,
but some come close. The strongest evidence for both the
environmentalist and hereditarian perspectives is of this sort.

After the end of World War II both black and white soldiers in the
occupying armies in Germany fathered children with white German women.
Klaus Eyferth gathered data on a large number of these children, of
mainly working-class mothers, and gave the children intelligence
tests.^8 He found almost no difference between the children of white
fathers and those of black fathers. The finding is remarkable given
that the black children faced a somewhat more hostile environment than
the white children. Hereditarians have challenged these findings by
appealing to the possibility that the black soldiers who fathered
these children might have been a particularly elite group. Flynn has
researched the plausibility of this explanation and concludes that
such selection did not play more than a small role.^9 Thus Eyferth's
study suggests that the black-white gap is largely, and possibly
entirely, environmental.

A study similar to Eyferth's found the cognitive ability of black
children raised in an orphanage in England to be slightly higher than
that of white children raised there.^10 Again, critics have raised the
possibility that the black children were genetically advantaged
relative to other blacks, and the whites disadvantaged relative to
other whites. And again, Flynn finds it unlikely that this contention
explains [End Page 58] much of the disappearance of the black-white
gap.^11 This study, too, suggests that the black-white gap is mainly
environmental.

If the black-white gap is mainly genetic in origin, children's
cognitive ability should not depend on the race of their primary
caregiver, comparing those of the same race. Yet two studies comparing
the experience of black children raised by black or white mothers
suggest that it does.^12 Here too, because the children were not
randomly assigned to their caregivers, it is possible that the
children raised by black mothers were of lower genetic potential, but
it would be hard to make such a selection story explain more than a
small fraction of the apparent environmental effect.

Another transracial adoption study provides mixed evidence, but some
of the strongest that genes play a role in explaining the black-white
gap.^13 A group of children, some with two black parents and some with
one white and one black parent, were raised in white middle-class
families. When the children's cognitive ability was tested at age
seven, the children with two black parents scored 95, higher than the
average black child in the state (89) and only slightly below the
national average for whites, while the mixed-race children scored 110,
which was considerably above it.^14 On the one hand, this finding
suggests a huge effect of environment on the cognitive ability of the
adopted black and mixed-race children. On the other hand, the higher
scores of the mixed-race children suggest that parents' genes may
account for some of the difference from the black children, and that
the mixed-race children may have had a better inheritance by virtue of
having one white parent. Both black and mixed-race children scored
worse than the biological children of their adoptive parents (who
scored 116), an expected finding because the adopting parents were an
elite group and likely passed on above-average genetic potential to
their children. But they also scored considerably below the average of
118 for comparison white children adopted into similar homes.

When the same children were retested ten years later, the results were
different.^15 The scores of the children with two black parents had
dropped to about the average for blacks in the state where they lived
before they were adopted (89). The scores of the mixed-race children
had dropped too (99), but remained intermediate between those of the
children with two black parents and those of the adoptive parents'
biological children, which had also declined, to 109. The scores of
the white children raised in adoptive homes had dropped the most,
falling to 106.

The disappearance of the salutary effect of the adoptive home,
however, does not mean that genes determine black-white differences.
We can assume that as the children aged and moved out into the world,
the effect of the home environment diminished, and both whites and
blacks tended to the average for their own population because of
either genetic or environmental effects. By showing how the effect of
a child's home environment disappears by adolescence, this study
suggests [End Page 59] that environmental disadvantages experienced by
blacks as children cannot explain the deficit in their cognitive
ability as adolescents and adults. But environmental disadvantages
facing black adolescents and adults could still explain those
deficits. The transience of environmental effects on cognitive ability
is a theme to which I shall return. The persistence of the advantage
of the mixed-race children over the children with two black parents is
suggestive of a role for genes. It is not, though, definitive: several
other explanations have been offered, including the late adoption of
the children with two black parents and parental selection effects
unrelated to race.^16

Indirect Evidence on the Role of Genetic Differences

Although the direct evidence on the role of environment is not
definitive, it mostly suggests that genetic differences are not
necessary to explain racial differences. Advocates of the hereditarian
position have therefore turned to indirect evidence.^17

Several authors have argued that estimates of the heritability of
cognitive ability put limits on the plausible role of environment.^18
The argument is normally made in a mathematical form, but it boils
down to this. First, it is now widely accepted that differences in
genetic endowment explain at least 60 percent of the variance in
cognitive ability among adults in the white population in the United
States.^19 If all the environmental variation among U.S. whites can
explain only 40 percent of the variance among whites, how could
environmental differences explain the huge gap between blacks and
whites? The mathematical argument implies that the average black
environment would have to be worse than at least 95 percent of white
environments, but observable characteristics of blacks and whites are
not that different. For example, black deficits in education or in
socioeconomic status place the average black below only about 60 to 70
percent of whites.^20

The heritability of cognitive ability is also crucial to a second type
of indirect evidence for a role of genetic differences in explaining
the black-white gap. Arthur Jensen has advanced what he calls
"Spearman's Hypothesis," after the late intelligence researcher
Charles Spearman, who observed that people who had large vocabularies
were good at solving mazes and logic problems and were also more
likely to have command of a wide range of facts. Spearman posited that
a single, largely genetic, mental ability that he called g (for
general mental ability) explained the correlation of people's
performance across a wide range of tests of mental ability.
Researchers now know that a single underlying ability cannot explain
all the tendency of people who do well on one type of test to do well
on another.^21 But it is possible to interpret the evidence as
indicating that there is a single ability that differs among people,
that is subject to genetic influence, and that explains much of the
correlation across tests. Other interpretations are also possible, but
this one cannot be discounted. In a series of studies Jensen and
Rushton have argued that different types of tests tap this general
ability to different degrees; that the more a test taps g, the more it
is subject to genetic influence; and that black-white differences are
largest on the tests most reflective of the underlying general
ability, g.^22

Using several restrictive assumptions about the nature of genetic and
environmental influence on genetic ability, researchers can use this
information to estimate the fraction of the black-white gap that is
due to differences in genetic endowment. The more the [End Page 60]
pattern of black-white differences across different tests resembles
the pattern of genetic influence on different tests, the more the
statistical procedure will attribute the black-white differences to
genetic differences. Using this method, David Rowe and Jensen have
independently estimated that from one-half to two-thirds of the
black-white gap is genetic in origin.^23

A Problem for the Indirect Arguments: Gains in Cognitive Ability over Time

Over the past century, dozens of countries around the world have seen
increases in measured cognitive ability over time as large as or even
larger than the black-white gap.^24 The phenomenon has been christened
the "Flynn Effect," after James Flynn, who did the most to investigate
and popularize this worldwide trend. The score gains have been
documented even between a large group of fathers and sons taking the
same test only decades apart, making it impossible that the gains are
due to changes in genes. Clearly environmental changes can cause huge
leaps in measured cognitive ability. Although it might not seem
plausible that the average black environment today is below the 5th
percentile of the white distribution of environments, it is certainly
plausible that the average black environment in the United States
today is as deprived as the average white environment of thirty to
fifty years ago--the time it took for cognitive ability to rise by an
amount equal to the black-white gap in many countries. These gains in
measured cognitive ability over time point to a problem in the
argument that high heritability estimates for cognitive ability
preclude large environmental effects.

Gains in cognitive ability over time also challenge the logic of
Jensen's genetic explanation for the pattern of black-white
differences across different types of tests. All studies show that
gains on different tests are positively correlated with measures of
test score heritability, and most studies show that gains are
positively correlated with the extent to which a test taps the
hypothesized general cognitive ability.^25 There is little doubt that
applying the same method as Rowe and Jensen used to data on gains in
cognitive ability over time would show them to be partially genetic in
origin, something we know cannot be true.

So, what is it that is wrong with the logic of these two arguments,
that the high heritability of cognitive ability limits the possible
effect of the environment and that the pattern of black-white
differences across different tests shows those differences to be
genetic in origin? And in particular, where is the problem in the
first?

It is important to detect the flaw, because if the logic of the
argument were sound, the case for environmental causes of black-white
differences would be difficult to make, and the possibility of
remedying those differences would be remote. But before I explain, I
want to cite two other pieces of evidence marshaled by advocates of
the hereditarian position that suggest the limited power of the
environment to change cognitive ability (and therefore to explain the
entire black-white gap). The first is that the heritability of
cognitive ability rises with age. It does so at the expense of the
effect of family environment, which disappears nearly completely in
most studies of late adolescents and adults.^26 The disappearance of
the effect on black children of being raised in white families, which
I have already noted, is just one case of a general finding from
several different types of studies. A second piece of evidence is the
fade-out of the effect of preschool programs [End Page 61] on
cognitive ability. Although such programs have been shown to have
profound effects on the measured ability of children, the effects fade
once the programs end, leaving little evidence of any effect by
adolescence.^27 Is it possible to reconcile the high heritability of
cognitive ability with large, but transient, environmental effects?

The Interplay of Genes and the Environment

To explain this puzzle, James Flynn and I have proposed a formal model
in which genes and environment work together, rather than
independently, in developing a person's cognitive ability.^28 The
solution involves three aspects of the process by which individual
ability is molded that are overlooked by the logic that implies small
environmental effects. We illustrate our argument with a basketball
analogy.

How can genes and environment both be powerful in shaping ability?
Consider a young man with a small genetic predisposition toward
greater height and faster reflexes. When he is young, he is likely to
be slightly better than his playmates at basketball. His reflexes will
make him generally better at sports, and his height will be a
particular advantage when it comes to passing, catching, and
rebounding. These advantages by themselves confer only a small edge,
but they may be enough to make the game more rewarding for him than
for the average person and get him to play more than his friends and
to improve his play more over time. After a while, he will be
considerably better than the average player his age, making it likely
that he will be picked first for teams and perhaps receive more
attention from gym teachers. Eventually, he joins a school team where
he gets exhaustive practice and professional coaching. His basketball
ability is now far superior to that of his old playmates. Through a
series of feedback loops, his initial minor physical advantage has
been multiplied into a huge overall advantage. In contrast, a child
who started life with a predisposition to be pudgy, slow, and small
would be very unlikely to enjoy playing basketball, get much practice,
or receive coaching. He would therefore be unlikely to improve his
skills. Assuming children with a range of experience between these two
extremes, scientists would find that a large fraction of the variance
of basketball playing ability would be explained by differences in
genetic endowment--that basketball ability was highly heritable. And
they would be right to do so. But that most certainly would not mean
that short kids without lightning reflexes could not improve their
basketball skills enormously with practice and coaching.

The basketball analogy so far illustrates two of the considerations
that Flynn and I believe are important for understanding the
implications of behavioral genetic studies of cognitive ability.
First, genes tend to get matched to complimentary environments. When
that happens, some of the power of environment is attributed to genes.
Only effects of environment shared by all children in the same family
and effects of environment uncorrelated with genes get counted as
environmental. Second, the effect of genetic differences gets
multiplied by positive feedback loops. Small initial differences are
multiplied by processes where people's initially varying abilities are
matched to complimentary environments that cause their abilities to
diverge further.

In theory this same multiplier process could be driven by small
environmental differences. But to drive the multiplier to its maximum,
the environmental advantage would [End Page 62] have to be as constant
over time as the genetic difference, because in the absence of the
initial advantage there will be a tendency for the whole process to
unwind. For example, suppose that midway through high school the
basketball enthusiast injures a leg, which makes him less steady and
offsets his initial advantage in height and reflexes. Because of all
his practice and learning, he will still be a superior player. But his
small decrement in performance could mean discouragement, more bench
time, or not making the cut for the varsity team. This could lead to a
further deterioration of his skills and further discouragement, until
he gives up playing on the team entirely. Although each individual's
experience will differ, the theory that Flynn and I lay out would have
people with average physical potential reverting to average ability
over time, on average.

The transitory nature of most environmental effects not driven by
genetic differences helps explain why environmental differences do not
typically drive large multipliers and produce the same large effects
as genetic differences. That same transience helps explain why
environment can be potent but still cause a relatively small share of
the variance of cognitive ability in adults.^29

Social Multipliers and the Effect of Averaging

If most external environmental influences are transitory and
transitory environmental effects are unable to drive multipliers, what
explains the large gains in cognitive ability over the past century?
That question has two answers. One is the social multiplier process.
The other is that many random transient environmental effects that
lean in one direction when averaged together can substitute for a
single persistent environmental cause. This is the third point missed
by the argument that claims that high heritability implies small
environmental effects.

Another basketball analogy will help explain social multipliers.
During the 1950s television entered many U.S. homes. Professional
basketball, with its small arena, could not reach as wide an audience
as baseball, but basketball translated much better to the small
screen. Thus public interest in basketball began to grow. The
increased interest made it easier for enthusiasts to find others to
play with, thus increasing the opportunities to improve skills. As
skills improved, standards of play rose, with players learning moves
and skills from each other. As more people played and watched the
game, interest increased still further. More resources were devoted to
coaching basketball and developing basketball programs, providing yet
more opportunities for players to improve their skills. In the end,
the small impetus provided by the introduction of television had a
huge impact on basketball skills.

A similar process may well be at work for cognitive ability. An
outpouring of studies in recent years suggests that social effects
have an important influence on school performance.^30 One study of an
experimental reduction in school class size resulting in major
achievement score gains suggests that a very large fraction of the
gains came through the children's extended association with their
peers, who shared the experience of small class sizes.^31 In this case
an arguably minor intervention had large and long-lasting effects
largely owing to a social multiplier effect.

But improvements in cognitive ability could have many triggers, rather
than a single one. Many such triggers over the past half-century
averaged together could be acting to raise cognitive ability.
Increasing cognitive demands [End Page 63] from more professional,
technical, and managerial jobs; increased leisure time; changing
cognitive demands of personal interactions; or changing attitudes
toward intellectual activity could all be playing a role. And small
initial changes along any of these dimensions would be magnified by
individual and social multipliers.

Genes and Environment and the Black-White Gap

The black-white gap in measured cognitive ability may come about in a
similar way, but it could have even more triggers. Segregation and
discrimination have caused many aspects of blacks' environment to be
inferior to that of whites. Averaged together, the total impact can be
large, even if each individual effect is small. Suppose, for example,
that environment relevant to the formation of cognitive ability
consists of 100 factors, each with an equal effect. If for each of
these 100 factors the average black were worse off than 65 percent of
whites, he would be worse off than 90 percent of whites when the
effects of all the environmental factors were considered together.
(The disparity is the necessary result of accumulating a large number
of effects when two groups have slightly different means for all the
effects.)^32 Taking the total effect of environment in this way,
considering the underestimate of the total effect of environment
because some of its power is attributed to genes, and considering
individual and social multipliers, a purely environmental explanation
for black-white differences becomes plausible despite high estimates
for the heritability of cognitive ability.

Moreover, our model also has explanations for the correlation of the
heritability of scores on different tests with the size of the
black-white gap on those tests and the anomalous correlation of the
size of gains in cognitive ability over time on different tests with
the heritability of those test scores. Those cognitive abilities for
which multiplier processes are most important will be the ones that
show the largest heritability, because of the environmental
augmentation of the genetic differences. But they will also be the
ones on which a persistent change in environment will have the biggest
influence. Thus we might expect that persistent environmental
differences between blacks and whites, as well as between generations,
could cause a positive correlation between test score heritabilities
and test differences.^33 Rushton and Jensen's indirect evidence of a
genetic role in black-white differences is, therefore, not probative.

Implications and Conclusions

The indirect evidence on the role of genes in explaining the
black-white gap does not tell us how much of the gap genes explain and
may be of no value at all in deciding whether genes do play a role.
Because the direct evidence on ancestry, adoption, and cross-fostering
is most consistent with little or no role for genes, it is unlikely
that the black-white gap has a large genetic component.

But what if it does? What would be the implications for the school
readiness of children? Much of the variance in human behavior,
including cognitive ability and achievement test scores, can be traced
to differences in individuals' genetic endowments. But as indisputable
as is the role of genes in shaping differences in outcomes within
races, so is the role of environment. Studies of young children show
that environmental differences explain more variation than do genetic
differences. And even studies showing an important role for genes in
no way rule out the possibility of improving the school performance of
disadvantaged children through [End Page 64] interventions aimed at
enhancing their school readiness. Interventions should stand or fall
on their own costs and benefits and not be prejudged on the basis of
genetic pessimism.

In fact, studies of the role of genes and environment in determining
school readiness offer some useful lessons in designing and evaluating
interventions. These studies show that normally occurring differences
in preschool resources or parenting practices in working- and
middle-class families have only limited effects on school readiness
once the correlation due to parents' and children's genes is taken out
of play.^34 Thus small interventions that make only modest changes in
the allocation of resources or the nature of parenting practices will
have limited to modest effects at best. Effects will likely be
somewhat larger if interventions target very disadvantaged families,
probably because the room for improvement is greater.^35

Achieving permanent effects on cognitive ability is harder than
achieving large effects. Most environmental effects on cognitive
ability seem to be like the effect of exercise on physical
conditioning: profound but short-lived. But even short-lived
improvements in cognitive ability can be valuable if they mediate
longer-term changes in achievement--for example, if improved cognitive
ability for some period of time allows students to learn to read more
quickly, putting them on a permanently higher achievement path. And
evidence suggests that programs aimed at improving cognitive ability
do have long-term effects on achievement even if they have no
significant long-term effects on cognitive ability. However, if
interventions make even small permanent changes in behavior that
support improved cognitive ability, they can set off multiplier
processes, with improved ability leading to better environments and
still further improvements in ability. If we knew what aspects of
preschool programs help elevate cognitive ability, and if we could get
children to continue to seek out such stimulation after they leave
preschool programs, their increased ability could lead them to
associate with more able peers, to have the confidence to take on more
demanding academic challenges, and to get the further advantage of yet
more positive stimulation from these activities. This, in turn, could
further develop their cognitive ability. Long-lived effects are more
likely to be large effects.

Effects are particularly likely to be large if an intervention
saturates a social group and allows the individual multiplier effects
to be reinforced by social multipliers or feedback effects. If
students find themselves among others with greater ability, individual
interactions and group activities are more likely to give rise to
further improvements in cognitive ability. In this same vein,
evaluations that do not take into account the social effects of the
intervention on children who did not directly take part may be missing
an important aspect of the effects of an intervention.

Although much of normal environmentally induced variance in cognitive
ability seems to be transient, if interventions could induce even
small long-lasting changes in behavior, they might produce very large
effects through the multiplier process. Taking advantage of such
processes may make it possible to overcome the black-white gap and put
black and white children on an even footing. [End Page 65]
William T. Dickens is a senior fellow in the Brookings Economic
Studies program. He acknowledges the excellent research assistance of
Rebecca Vichniac and Jennifer Doleac.

Endnotes

1. The review necessarily highlights only the most important studies;
a complete review of all the arguments on both sides of this debate
would require hundreds of pages and be beyond the scope of this
article.

2. Heritability is estimated by examining the similarity of people
with different degrees of genetic similarity raised in similar sorts
of environments, and there is some reason to believe that most
estimates are somewhat overstated by existing methods. Robert Plomin
and others, Behavioral Genetics, 4th ed. (New York: Worth Publishers,
2001), in chapter 5 and the appendix, provide a thorough discussion of
the methods used to estimate heritability. Mike Stoolmiller,
"Implications of the Restricted Range of Family Environments for
Estimates of Heritability and Nonshared Environment in
Behavior-Genetic Adoption Studies," Psychological Bulletin 125 (1999):
392-409, shows that adoption studies probably overstate the degree of
heritability and speculates on reasons why some other methods may as
well.

3. Robert Plomin and others, Behavioral Genetics (see note 2), examine
learning disorders on pp. 145-49, ADHD on pp. 227-29, and personality
in chapter 12. For the effects of genes on cognitive ability, see
Marcie L. Chambers and others, "Variation in Academic Achievement and
IQ in Twin Pairs," Intelligence (forthcoming); Lee Anne Thompson and
others, "Associations between Cognitive Abilities and Scholastic
Achievement: Genetic Overlap but Environmental Differences,"
Psychological Science 2 (1991): 158-65; and Sally J. Wadsworth,
"School Achievement," in Nature and Nurture during Middle Childhood,
edited by John C. DeFries, Robert Plomin, and David W. Fulker (Oxford:
Blackwell, 1994), pp. 86-101.

4. James R. Flynn, Race, IQ, and Jensen (London: Routledge, 1980).

5. Richard Nisbett, "Race, Genetics, and IQ," in The Black-White Test
Score Gap, edited by Christopher Jencks and Meredith Phillips
(Brookings, 1998), pp. 86-102.

6. Arthur Jensen, The g Factor (Westport, Conn.: Praeger, 1998), pp
350-531.

7. Jane R. Mercer, "What Is a Racially and Culturally
Nondiscriminatory Test? A Sociological and Pluralistic Perspective,"
in Perspectives on "Bias in Mental Testing," edited by Cecil R.
Reynolds and Robert T. Brown (New York: Plenum Press, 1984); Jonathan
Crane, "Race and Children's Cognitive Test Scores: Empirical Evidence
That Environment Explains the Entire Gap," mimeo, University of
Illinois at Chicago, 1994; and Jeanne Brooks-Gunn and others, "Ethnic
Differences in Children's Intelligence Test Scores: Role of Economic
Deprivation, Home Environment, and Maternal Characteristics," Child
Development 67, no. 2 (1996): 396-408.

8. This is based on the account by James R. Flynn (Race, IQ, and
Jensen, pp. 84-87; see note 4) of Klaus Eyferth, "Leistungen
verschiedener Gruppen von Besatzungskindern in Hamburg-Wechsler
Intelligenztest fur Kinder (HAWIK)," Archiv fur die gesamte
Psychologie 113 (1961): 222-41.

9. Flynn, Race, IQ, and Jensen, pp. 84-102 (see note 4).

10. Barbara Tizard, "IQ and Race," Nature 247, no. 5439 (February 1,
1974).

11. Flynn, Race, IQ, and Jensen, pp. 108-11 (see note 4).

12. Elsie G. J. Moore, "Family Socialization and the IQ Test
Performance of Traditionally and Transracially Adopted Black
Children," Developmental Psychology 22 (1986): 317-26; and Lee
Willerman and others, "Intellectual Development of Children from
Interracial Matings: Performance in Infancy and at 4 Years,"
Behavioral Genetics 4 (1974): 84-88. [End Page 66]

13. Sandra Scarr and Richard A. Weinberg, "IQ Test Performance of
Black Children Adopted by White Families," American Psychologist 31
(1976): 726-39; and Sandra Scarr and Richard A. Weinberg, "The
Minnesota Adoption Studies: Genetic Differences and Malleability,"
Child Development 54 (1983): 260-67.

14. These are IQ scores, which have a mean of 100 and a standard
deviation of 15 in the U.S. population.

15. Sandra Scarr and others, "The Minnesota Transracial Adoption
Study: A Follow-Up of IQ Test Performance at Adolescence,"
Intelligence 16 (1992): 117-35.

16. But see Arthur Jensen, The g Factor, pp. 477-78 (see note 6), on
whether late adoption can explain the difference.

17. One body of evidence is difficult to judge. See J. Philippe
Rushton, Race, Evolution, and Behavior: A Life History Perspective,
3rd ed. (Port Huron, Mich.: Charles Darwin Research Institute, 2000).
Rushton has proposed a theoretical framework that would explain a
genetic gap in cognitive ability between blacks and whites and has
marshaled evidence for it. But because much of the evidence was known
before the theory was proposed, some view the theory as nothing more
than post hoc rationalization for hereditarian views on the
black-white gap. At most it suggests that some of the black-white gap
may be genetic, but it does not suggest how much.

18. Arthur Jensen, Educability and Group Differences (New York: Harper
and Row, 1973), pp. 135-39, 161-73, 186-90; Arthur Jensen, Educational
Differences (London: Methuen, 1973), pp. 408-12; Jensen, The g Factor,
pp. 445-58 (see note 6); and Richard Herrnstein and Charles Murray,
The Bell Curve: Intelligence and Class Structure in American Life (New
York: Simon and Schuster, 1994), pp. 298-99.

19. Plomin and others, Behavioral Genetics, p. 177 (see note 2); and
Ulric Neisser and others, "Intelligence: Knowns and Unknowns,"
American Psychologist 51, no. 2(1996): 85.

20. Author's calculations from the 1979 National Longitudinal Survey
of Youth.

21. John B. Carol, Human Cognitive Abilities: A Survey of
Factor-Analytic Studies (Cambridge University Press, 1993), is the
most comprehensive survey of what is known about the correlation of
scores on different types of mental tests.

22. See J. Philippe Rushton and Arthur Jensen, "Thirty Years of
Research on Race Differences in Cognitive Ability," Psychology, Public
Policy, and Law (forthcoming), for a review of this evidence and
citations to the original studies.

23. David Rowe, Alexander Vazsonyi, and Daniel Flannery, "Ethnic and
Racial Similarity in Developmental Process: A Study of Academic
Achievement," Psychological Review 101, no. 3 (1994): 396-413; Jensen,
The g Factor, pp. 464-67 (see note 6).

24. James R. Flynn, "Massive Gains in 14 Nations: What IQ Tests Really
Measure," Psychological Bulletin 101 (1987): 171-91; James R. Flynn,
"IQ Gains over Time," in Encyclopedia of Human Intelligence, edited by
Robert J. Sternberg (New York: Macmillan, 1994), pp. 617-23; James R.
Flynn, "IQ Gains over Time: Toward Finding the Causes," in The Rising
Curve: Long-Term Gains in IQ and Related Measures, edited by Ulric
Neisser (Washington: American Psychological Association, 1998), pp.
551-53.

25. Existing evidence suggests that IQ gains across subtests are
probably positively correlated with g loading. See Roberto Colom,
Manuel Juan-Espinosa, and Luís F. García, "The Secular Increase in
Test Scores Is a [End Page 67] 'Jensen effect,'" Personality and
Individual Differences 30 (2001): 553-58; and Manuel Juan-Espinosa and
others, "Individual Differences in Large-Spaces Orientation: g and
Beyond?" Personality and Individual Differences 29 (2000): 85-98, for
much stronger correlations between g loadings and IQ gains. Jensen,
The g Factor, pp. 320-21 (see note 6), reviews a number of studies of
the relation between subtests gains and g loadings, all of which show
weak positive correlations. J. Philippe Rushton, "Secular Gains in IQ
Not Related to the g Factor and Inbreeding Depression--unlike
Black-White Differences: A Reply to Flynn," Personality and Individual
Differences 26 (1999): 381-89, finds that a measure of g developed on
the Wechsler Intelligence Scale for Children has loadings that are
negatively correlated with subtest gains in several countries. But see
James R. Flynn, "The History of the American Mind in the 20th Century:
A Scenario to Explain IQ Gains over Time and a Case for the
Irrelevance of g," in Extending Intelligence: Enhancement and New
Constructs, edited by P. C. Kyllonon, R. D. Roberts, and L. Stankov
(Hillsdale, N.J.: Erlbaum, forthcoming.) for an argument that IQ gains
are greatest on tests of fluid g rather than crystallized g. He finds
a positive (though statistically insignificant) correlation between a
measure of fluid g he develops and IQ gains in the data used by
Rushton. Olev Must, Aasa Must, and Vilve Raudik, "The Flynn Effect for
Gains in Literacy Found in Estonia Is Not a Jensen Effect,"
Personality and Individual Differences 33 (2001); and Olev Must, Aasa
Must, and Vilve Raudik, "The Secular Rise in IQs: In Estonia the Flynn
Effect Is Not a Jensen Effect," Intelligence 31 (2003): 461-71, find
no correlation between g loadings and gains on two tests in Estonia,
but these are achievement tests with a strong crystallized bias.

26. Plomin and others, Behavioral Genetics,pp. 173-77 (see note 2).

27. Irving Lazar and Richard Darlington, "Lasting Effects of Early
Education: A Report from the Consortium for Longitudinal Studies,"
Monographs of the Society for Research in Child Development 47,
nos.2-3 (1982).

28. William T. Dickens and James Flynn, "Heritability Estimates versus
Large Environmental Effects," Psychological Review 108, no. 2 (2001).

29. This is not to say that there are no permanent or long-lasting
environmental effects on cognitive ability. The effects of brain
damage can be severe and permanent. However, such permanent
environmental effects evidently explain only a small fraction of
normal variation in cognitive ability. Shared family environment plays
a large role in explaining variance in cognitive ability when children
are spending most of their time in the home, with their activities
strongly influenced by their parents. But that effect fades as they
spend more of their time away from home and in self-directed
activities.

30. Eric A. Hanushek and others, "Does Peer Ability Affect Student
Achievement?" Working Paper 8502 (Cambridge, Mass.: National Bureau of
Economic Research, 2001); Caroline Hoxby, "Peer Effects in the
Classroom: Learning from Gender and Race Variation," Working Paper
7867 (Cambridge, Mass.: National Bureau of Economic Research, 2001);
Dan M. Levy, "Family Income and Peer Effects as Determinants of
Educational Outcomes," Ph.D. diss., Northwestern University, 2000;
Donald Robertson and James Symons, "Do Peer Groups Matter? Peer Group
versus Schooling Effects on Academic Achievement," Economica 70
(2003): 31-53; Bruce Sacerdote, "Peer Effects with Random Assignment:
Results from Dartmouth Roommates," Quarterly Journal of Economics (May
2001): 681-704; David J. Zimmerman, "Peer Effects in Academic
Outcomes: Evidence from a Natural Experiment," Review of Economics and
Statistics 85 (2003): 9-23. [End Page 68]

31. Michael A. Boozer and Stephen E. Cacciola, "Inside the 'Black Box'
of Project STAR: Estimation of Peer Effects Using Experimental Data,"
Discussion Paper 832 (Economic Growth Center, Yale University, 2001).

32. In statistics this is referred to as the law of large
numbers--that the variance of a mean falls as the number of items
being averaged goes up. See Eugene Lukacs, Probability and Mathematics
Statistics: An Introduction (New York: Academic Press, 1972). It
applies whether or not the weights being put on the elements are
equal. Because the variance and standard deviation of the mean fall,
while the average difference stays the same, the difference in
standard deviations grows. The example assumes that the effects are
all uncorrelated with each other and that each has a normal
distribution in the white and the black populations. If the effects
were assumed to be correlated or the weights unequal, the results
would be less dramatic, but with observed values for correlations of
environmental factors, increasing the number of items to be averaged
could produce the same results.

33. Dickens and Flynn, "Heritability Estimates vs. Large Environmental
Effects" (see note 28).

34. Plomin and others, Behavioral Genetics, p. 201(see note 2).

35. Eric Turkheimer and others, "Socioeconomic Status Modifies
Heritability of IQ in Young Children," Psychological Science 14, no. 6
(2003). Their own study finds that shared family environment explains
60 percent of the variance of an IQ test score in
low-socioeconomic-status seven-year-olds, which is a much larger share
than other studies have found. For example, see Kathryn Asbury and
others, "Environmental Moderators of Genetic Influence on Verbal and
Nonverbal Abilities in Early Childhood" (Institute of Psychiatry, De
Crespigny Park, London, 2004).

[I am sending forth these memes, not because I agree wholeheartedly with 
all of them, but to impregnate females of both sexes. Ponder them and
spread them.]