Sloppy Statistics, Bogus Science
Please note, this is a work in progress, and will be completed with sources, etc. soon... JP
and the Assault on Racial Equity
The "science" of books like The Bell Curve is inherently flawed, mostly because the concept of distinct biological "races" that can be studied, compared and found to be "superior" or "inferior" is itself a misnomer. The term race, in biology, properly refers to subspecies (i.e. subpopulations that are sufficiently different in genetic terms to be on the verge of "speciation," or splitting into entirely new species). So, those who defend the notion of scientific "racial" differences must be able to demonstrate that human subpopulations diverge in such a manner and to such a degree.
But such "proof" is impossible to come by, because, in fact, science disproves it conclusively. To begin with, 68% of all human genes are identical between all humans, and have no polymorphic variation whatsoever. Thus, to whatever degree any two human individuals (let alone groups of persons) differ from one another, these differences are, by definition, limited to the approximately one-third of genes where difference is even theoretically possible.
When this 32% of genes that could contain "differences" have been studied in-depth (Nei and Roychoudhury 1982), it is found that the net codon differences between human "racial" groups are significantly smaller than the differences between two randomly selected genomes from within a particular group. In other words, if whites and blacks are subspecies, and inherently different, this would mean that there would have to be thousands of subspecies within the "white" group, and "black" group as well, since the in-group differences are so much larger than the inter-group differences.
In fact, 96.8% of the genetic code between blacks and whites is shared, with only a maximum of 0.032 of the genes varying between any white or black person. The variation between whites and Asians is 0.019 (98.1% similarity), and the difference between blacks and Asians is 0.047 (95.3% similarity). These differences are far too small to indicate subspeciation, as such phenomenon would typically be characterized by variation many times greater than the above numbers. There are no subspecies of a given phylum with this high a degree of genetic overlap, anywhere in nature.
Unless one can show that the miniscule variation indicated by the above data is sufficient to produce the variations in "intelligence," crime rates, fertility, etc then the claims of those arguing for scientific superiority/inferiority along racial lines is automatically nonsense. Even to classify one as black, white, asian, etc. becomes nonsensical, and most of either group (self-defined or culturally/nationally-defined) will be even closer to one another than the above variation would imply, since the above variation would be for totally "pure" versions of each group, few of which actually exist, particularly in the West.
Furthermore, evidence from population genetics has indicated for years that less than 1% of all genes are linked to the transfer of pigmentation and other "racially distinct" characteristics, further limiting the effect of actual "race" on human genetic variation. (Allport, 1956, 1979)
The history of racial "science" is replete with contradictions, and rationalizations for white racial supremacy. When data has emerged that contradicts the racist views of scientists, they have simply altered their arguments to take account of the new scientific "reality" and to allow them to maintain their existing perceptions of superiority and inferiority. As Tucker notes (1994), at the turn of the century, scientists were saying that the tendency for blacks to do far better on memory-skill tests than whites was due to their being "closer to the primitive state" where memory was more important and functional than higher-order skills, logic, reason, etc. Then 10 years later, when tests began to show that the children of rich whites scored better on memory tests, these same scientists insisted that their superior performance was due to their "greater intelligence." Then many years later, these same scientists once again changed their minds when new memory tests showed that poor and black kids with "low IQ's" had excellent memories that surpassed those of wealthier folks and whites.
The same shifting arguments have occurred with regard to tests of "reaction time." When tests have demonstrated that blacks have quicker reaction time to various external stimuli, this has been taken as proof that they were intellectually and emotionally inferior, less reflective and analytical, etc. But when some tests have demonstrated a white edge on reaction time, these same scientists have claimed it was a sign of having "quicker brains." Similarly, throughout the years, respected "scientists" have claimed that blacks evolved earlier than whites and are thus inferior, and alternately, that blacks evolved later, and are thus inferior.
The entire enterprise of books like the Bell Curve is suspect, when one considers some of the group differences it DOESN'T consider, in favor of those it does think relevant (namely those between whites, blacks and Asians). For example, as Hacker has pointed out (1994), while Murray and Herrnstein argue that IQ is a key factor in determining educational success and attainment, and that low black IQ is what explains the generally lower academic achievement among African Americans, they conveniently avoid mentioning or explaining the large gaps between various "white" ethnic/national groups, and what those gaps must (by their own logic) say about THEIR relative IQ's. For example, according to the Census Bureau:
51% of Americans with Russian descent have a college degree, compared to only:
33.4% of Scots
28.6% of English
27.5% of Swedes
23.4% of Poles
22.1% of Germans
21.9% of Italians,
and 21.3% of Irish
Does this mean that Russian-descended "whites" are much smarter than all other whites, and that Irish whites are the least intelligent? And why has there been no study to explain these differences?
The Bell Curve is, from beginning to end, plagued by an overreliance on questionable sources, with overtly racist motives for their "science," to put it kindly:
As Lane points out (1994), the book relies heavily on articles originally published in Mankind Quarterly, and individual researchers tied to Mankind Quarterly: a publication founded in 1960 explicitly to counter the influence of the civil rights revolution in the US and "egalitarian" thinking about race and inequality. MQ was founded by Robert Gayre, who remained the editor until 1978. Gayre was an open advocate of apartheid in South Africa, and white rule in Rhodesia. In 1968, he testified as an expert in a hate speech trial of British neo-Nazis, and noted in his testimony that blacks were "worthless."
Other founders of MQ include Henry Garrett, who was a pamphleteer for the White Citizens Council; and Corrado Gini, a leader of Italy's eugenics program under Mussolini.
MQ over the years has published the work of former Nazi scientists, including those who mentored Joseph Mengele, such as Ottmar Von Verschuer, who served on the MQ Advisory Board. The publication also praised Ku Klux Klansman, Ernest Sevier Cox and his book White America as a "classic book by a truly great man," and referred to his plan for sending all blacks back to Africa, as a "practical solution" for keeping the U.S. white. (Tucker 1994)
Since 1978, MQ has been run mostly by Roger Pearson, who was a founder of the neo-Nazi Northern League in 1958, and once bragged that he had helped hide Nazi doctor Joseph Mengele. In the 1960's, Pearson hooked up with notorious racist and anti-Jewish activist Willis Carto, to publish Western Destiny, where he would rail against the "Jewish culture distorters," who were seeking to "capture the minds, morals and souls of our children." The year that he took over at Mankind Quarterly, he was the conference chair at the national convention of the World Anti-Communist League, where his conference assistant (also a staffer at Pearson's publishing house) was a former stormtrooper in the American Nazi Party.
During his tenure as Editor, he has published articles by Ralph Scott and Donald Swan, two Americans with open connections to neo-Nazis. In 1966, Pearson himself argued: "if a nation with a more advanced, more specialized, or in any way superior set of genes mingles with, instead of exterminating an inferior tribe, then it commits racial suicide."
Pearson's Institute for the Study of Man (the publisher of MQ) is funded by the Pioneer Fund: an organization founded in 1937 by two men who openly admired Hitler, and advocated the repatriation of blacks to Africa. Even today, the Fund remains openly racist. The Fund's charter provides insight into the worldview their "science" seeks to support, when it notes that Pioneer "is committed to the proposition that people of different ethnic and cultural backgrounds are, on the basis of their heredity, inherently unequal and can never be expected to behave or perform equally." (Cohen and Solomon 1995)
In the Bell Curve itself, the authors cite 13 "scholars" who have had their work funded by Pioneer. Among these is Phillipe Rushton of Ontario. Rushton is cited eleven times in the Bell Curve, and Murray and Herrnstein go to great lengths to ensure their readers that Rushton "is not a quack." This despite the fact that Rushton's "scientific methodology" has included approaching shoppers at a Toronto mall (one-third black, one-third white, and one-third Asian) and asking them "how far can you ejaculate," or "how large is your penis?" He has also said, that intelligence is inversely related to penis size, because "it's more brain or more penis. You can't have everything," and has claimed that the success of the Nazi army was due to its Aryan genetic purity.
Interestingly enough, Rushton's data on penis size all comes from one study, conducted in 1898 by an anonymous French Army surgeon who traveled through Africa and recorded the size of African penises, and from a second study comparing the penises of Nigerian medical students to Czech army officers. In this study, it turned out the Nigerians penises were longer, and the Czech's had greater circumference. So why is length more important in effecting brainpower than girth? Who knows? Neither the original study, nor Rushton, explains this point.
Rushton's arguments on brain size are based on the genetic distance studies of Dr Allen Wilson, from the University of California. Yet Wilson, having reviewed Rushton's work, notes that those scientists using his work to argue for innate racial differences, let alone superiority or inferiority in intelligence, have "totally misrepresented" his findings. Furthermore, respected anthropologists like Christopher Springer at the British Museum have noted that Rushton's brain size and head size data is completely without merit.
Rushton's claims that blacks have higher levels of testosterone, thereby explaining higher rates of aggression and violence, and that the difference is racial, not environmental, ignore three commonly understood, entirely non-genetic factors that influence testosterone levels: diet, exercise and emotional state. (Hoberman 1996). Higher testosterone levels also could result from higher stress levels prior to giving blood, which in turn could easily result from the general uneasiness many black folks feel about a largely white medical establishment, and their interactions with health care providers.
The Bell Curve also relies heavily on the research of Richard Lynn, described by the authors as "a leading scholar of racial and ethnic differences." As one example of Lynn's scholarship, consider this quote, cited in Newsday, November 9, 1994:
"What is called for here is not genocide, the killing off of the population of incompetent cultures. But we do need to think realistically in terms of ‘phasing out' such peoples…Evolutionary progress means the extinction of the less competent. To think otherwise is mere sentimentality."
Lynn has also explained, that in his opinion, "…the poor and the ill are weak specimens whose proliferation needs to be discouraged in the interests of the improvement of the genetic quality of the group, and ultimately of group survival," and that "the Caucasoid and the Mongoloids are the only two races that have made any significant contribution to civilization," leading one to wonder where Lynn—obviously no historian—would place the Ancient Egyptians among these two racial groups.
The Bell Curve references Lynn's work in an effort to "prove" the following propositions that are central to the book's arguments:
1) African Blacks have IQ's substantially below the African American average;
2) East Asians have higher IQ's than any other group; and,
3) Immigrants of color to the US have sub-par IQ's
Taking a look at his "evidence" on African IQ, there is little doubt of its intellectual vacuity. Lynn's "proof" was based heavily on a 1988 review by three South African psychologists who looked at Black South African test performance. But the authors of this study concluded the OPPOSITE of Lynn and Murray and Herrnstein. In fact, when presented with Lynn's interpretation of their work, they responded with the following:
"It would be rash to suppose that psychometric tests constitute valid measures of intelligence among non-Westerners. The inability of most psychologists to look beyond the confines of their own cultures has led to the kind of arrogance whereby judgments are made concerning the ‘simplicity' of African mental structure and ‘retarded cognitive growth'."
The main source for the Bell Curve's claims regarding African IQ was a Lynn article from Mankind Quarterly in 1991, in which he said mean African IQ was 70. Lynn claims that he arrived at this figure by looking at the "best studies" on the subject since 1929. The study he claimed was the "best" was conducted in 1989 and involved 1,093 16-year old blacks, who scored a mean of 69 on the South African Junior Aptitude Test. From this, Lynn then extrapolated mean IQ to the whole of Black Africa. Even worse, Lynn completely misconstrued the findings of the study in question. According to the study's author, Dr Ken Owen, his test was "not at all" evidence of genetic intelligence. In fact, Owen has noted that the results were found directly related to the existence of apartheid era oppression, and the fact that the test was in English.
Another of the "definitive" studies cited by Lynn in his own article was a 1929 study, in which 293 blacks in South Africa were given the Army Beta Test and scored a mean of 65. But this test was administered by M.L. Finch, an open protagonist of the view that blacks were inherently inferior, even before he had done any studies to "prove" such a thing: he was, in other words, hardly a pure, unbiased scientist. Furthermore, the Beta Test was one of the most culturally biased tests in the world at that time: one question on the 1929 version in dispute showed people playing tennis without a net. To get full credit for the question, one would have to draw the net in the picture—something few black Africans could have possibly known to do in 1929, having never been exposed to the game. A leading proponent of the Beta Test, C.C. Brigham, actually admitted that the test had no validity whatsoever for non-Americans: a fact totally ignored by Lynn, and by the Bell Curve.
As for the East Asian IQ superiority, this "evidence" was as weak and uncompelling as that regarding African IQ's. The data on this issue in the Bell Curve relies heavily on a Lynn article from 1982 in Nature magazine, in which he claimed the Japanese have a 10-point IQ advantage over European whites. In 1983, Lynn's piece was dissected completely in the pages of Nature by two American psychologists who noted that Lynn's study sample was made up of Japanese kids from well-off urban parents, likely to have higher IQ's than the more mixed group of Europeans against which they were compared. Murray and Herrnstein only make mention of the Nature rebuttal to Lynn in a footnote, and even there, refuse to discuss its contents.
Two other studies cited by Lynn to "prove" higher Asian IQ's are equally bogus. The first used samples of American, British and Japanese students on a test of abstract reasoning. On this test the Brits and Americans did far worse; and the second study found that 9-year-olds in the UK did worse on the Ravens Standard Progressive Matrices than 9-year-olds from Japan and Hong Kong. But if you check the footnotes for this "evidence," you find that the author Lynn was citing for both of these studies was himself. And if you look up the studies, it doesn't take long to notice the flawed methodology involved in both: The first of these studies consisted of a test given to 178 Japanese children that did not reflect the demographic makeup of the nation as a whole, economically, culturally, or in terms of gender. The testers showed up at two schools, one urban and one rural, and gave the tests to whomever was present that day. Lynn then took the results of this test and compared it to a test that was thirteen years old, had been given to 64,000 American children, and had been pre-screened for representativeness; he then compared the Japanese results to a similarly pre-screened sample of 10,000 British children who had been given a similar test in the previous decade.
In the second study, Lynn claims to have found a substantial difference between Japanese and Hong Kong student IQ's on the one hand, and those of British children on the other. Yet this study looked only at 118 9-year olds from Hong Kong, 444 children from Japan and 239 British children, and involved no known controls for environmental and demographic representativeness.
The third set of studies cited in the Bell Curve dealing with Asian IQ, comes from Harold Stevenson in Minnesota, who found that once socioeconomic status and various demographic variables were controlled for there was NO difference at all between the IQ's of Japanese, Taiwanese and American kids. Despite the fact that these studies were the most comprehensive and methodologically sound of all the studies cited on the subject in the Bell Curve, (even according to the authors themselves who noted that Stevenson "carefully matched the children on socioeconomic and demographic variables"), Murray and Herrnstein essentially dismiss them as quickly as they mention them, noting only that they are evidence of the "ongoing debate" about race and IQ, as if they are on a scientific par with the work of folks like Lynn.
Finally, in the section on "immigrant IQ," meaning, to the authors, the IQ of mostly Latino immigrants to the U.S., Murray and Herrnstein claim that the mean Latino IQ is 91—about 9 points below whites, 14 points below Asians and 7 points above blacks—but provide absolutely no source for this claim whatsoever. And of course, there is no meaningful racial category known as "Latino" anyway, as the term refers to an ethnic/national/regional heritage group within which skin color and racial phenotype varies dramatically. The evidence from Lynn that they provide on "immigrant" IQ's, which they claim indicates an IQ in the 90-95 range, takes no account of the fact that 11% of all immigrants in the period they studied were South Asian and Middle Eastern, not Latino, and many more were East Asian—the very group they have claimed to be intellectually superior.
The data they utilize for their claims about immigrants also create some unique problems for Murray and Herrnstein that are quite unflattering for the overall thesis of their book. For example, the National Longitudinal Study of Youth (NLSY) data employed for the purpose of demonstrating inferior immigrant IQ was gathered from immigrants who spoke little English at the time they were given the entirely English-language test. And the same data indicates that blacks who immigrate to the U.S.—many of them from Africa—have, on average higher scores than native-born African Americans, a fact that utterly contradicts their arguments that the more "pure" blacks abroad are the lest intelligent of all racial types.
Furthermore, Murray and Herrnstein's positions on immigrant intelligence are internally contradictory. For example, the book praises the immigrants of the early 1900's who came mostly from Southern and Eastern Europe, claiming that they demonstrated intelligence and other functional and positive characteristics. The authors also praise the early IQ work of Carl Brigham. Yet, they conveniently ignore the fact that Brigham concluded these same Euro immigrants were 25% less intelligent than native-born whites, which is considerably worse in relative terms than modern-day Latino immigrants, who, by Murray and Herrnstein's own questionable data only score about 9% lower than native born whites, and yet, are seen as a "drag" on national intelligence, and are viewed in terms far less favorable than the white immigrants of the early 20th century who Murray and Herrnstein praise. (Feagin in Perea 1997)
Among the other "scientific" data used in the Bell Curve to "prove" racial difference and superiority/inferiority is that provided by Thomas Bouchard of the Minnesota Twin Project. Following in the footsteps of Sir Cyril Burt—whose infamous "twin studies" once were used to prove white racial superiority and the inferiority of persons of color until they were shown to be based on fraudulent and entirely fabricated evidence—Bouchard's work is questionable to say the least.
Bouchard claims to have observed twins raised apart from birth, and thus, subjected to different environments. When he found similar IQ's, as well as similarities in moral values, religiosity, political orientation and leisure-time interests among these separated twins, he pronounced that such a finding indicated that traits, including intellect were largely inherited, on the order of 70% genetic causation.
And yet, Bouchard's twin studies, which are now in their second decade, have still NEVER been published in a scholarly journal or monograph, nor have they been subjected to peer review by other researchers, nor have other scientists replicated them. Although Bouchard promised to publish a full-length study by 1987, he has failed to do so. (Mehler 1997) Since then he has refused to let any other researcher examine the case histories he claims to have studied.
Of course, even if intelligence between individuals is largely inherited or genetic, this would say nothing of the degree to which intelligence was inherited between racial groups, as such groups have no unique scientific or genetic delineations and are socially-created and demarcated collectives. If racial "groups" have no scientific basis then it is impossible to believe differences between said groups could themselves be based in science. Bouchard's studies, in that sense are meaningless, even if the science were accurate.
Other twin studies that have claimed to demonstrate the heritability of intelligence have been fatally flawed, in that the twins supposedly raised apart turned out to have been raised in close proximity, and even had contact with one another in many cases. Thus, to claim these twins were environmentally distinct in terms of their influences was a matter of bogus science.
A recent study by researchers at Carnegie-Mellon University and the University of Pittsburgh School of Medicine casts further doubt on the kinds of claims made by those like Bouchard. The researchers at these two schools found that prenatal health care was more relevant than any genetic factor in determining infant IQ. Conducting a meta-analysis of over 200 earlier studies, they determined that genes account for no more than 48% of the factors that determine IQ among individuals (to say nothing at all of groups). The study undercut the strength of twin study claims by showing that twins will be more alike in many ways, even if raised apart, because they shared a womb, and thus the same pre-natal environment and conditions. The study found that fraternal twins, who have different genes, have more similar IQ's than regular siblings, further indicating that prenatal environment is the key factor in their early intellectual development. In addition, it was determined that one-fifth of the similarity between twins is solely due to having inhabited the same womb at the same time.
Bouchard's further association with racist, bogus science was demonstrated in 1989 when he co-authored a paper with Phillipe Rushton (of the more penis, less brain school of racial science), in which they argued that blacks were more likely to contract AIDS because of "inherited reproductive strategies" that encourage promiscuous sex. (Lind 1994) And yet, geneticists trained in the field (which neither Rushton nor Bouchard have been) agree that reproductive strategies known as R and K-strategies do not differ within species, but only between species. To believe the Bouchard and Rushton thesis would require a belief in speciation, or the literal breaking off of blacks from the rest of the human family to such an extent that they were essentially no longer members of homo sapiens. As noted earlier, the evidence of genetic overlap and lack of polymorphic variation completely renders this belief untenable.
What is most amazing is that even using Murray and Herrnstein's own estimates in the Bell Curve, IQ only accounts for between 10-20% of the difference between individuals or "races" in terms of achievement. Furthermore, according to the authors, 60% of IQ derives from genetic heritability. This means that at most, inherited racial genetic differences can only account for between 6% and 12% of the achievement gaps between people or groups, leaving 88%-94% completely unexplained by "racial science." On certain tests like the Armed Forces Qualification Test (AFQT), Murray and Herrnstein admit that scores only can explain between 5% and 10% of the variation between people in terms of life achievements, meaning that if everyone had the same IQ, more than 9/10ths of social and economic inequality would remain, all other factors remaining constant. (Fischer et al 1996)
Next to their acceptance of the entirely unscientific notion of "race" as a category of biological meaning, the Bell Curve's second largest conceptual flaw is its acceptance of the validity of the field of psychometrics itself as a way to study the concept of intelligence. Psychometrics, which is defined literally as the study of measuring intelligence, takes as a given the unproven, unprovable, and frankly illogical assumption that all human talents are distributed among peoples in what amounts to a bell-shaped curve. In other words, very few people will have an overabundance of certain abilities; very few people will be utterly lacking certain abilities, and most will cluster somewhere around the middle.
By starting with this assumption (which can not be empirically tested because one must first define a "norm" against which to measure certain talents, and choosing the norm itself will skew the distribution), psychometrics will create a bell-curved distribution and ranking (on indices like intelligence), even if doing so means that people with very small actual differences on aggregate test performance will end up being separated by a huge gap on a standardized scale. For instance, if a standardized test is given, and there is a tendency among test takers to do extremely well, the demands of the psychometric enterprise—which requires a bell-shaped distribution—will require that the scores given to test takers be distributed in such a fashion (with few at the very top, few at the bottom and most in the middle), even if doing so requires taking one test taker who got 100 questions correct and separating her from another test taker who got 97 questions right, by a much larger percentile in standardized terms than would seem justified by their relatively identical performance.
In other words, the underlying assumptions that guide psychometricians and are built in to standardized tests guarantee a certain result and distribution, thereby eliminating randomness and generalizable validity. Murray and Herrnstein demonstrate this phenomenon themselves in the Bell Curve when discussing the results of the AFQT and NLSY studies. Originally, the results on these tests had been heavily clustered near the top of the distribution, with an overwhelming number of test takers doing exceptionally well in terms of questions answered correctly. Because these results did not fit into a bell-shaped distribution, Murray and Herrnstein simply recalculated the test results, using a different norming standard, to make the outcome fit their preconceived notions of what the results should look like. So they ended up giving extra credit for being at either the top or the bottom of the distribution, thereby stretching and flattening the overall curve of the distribution. Thus, at the top of the distribution, even very small differences in the number of answers one gave that were correct would result in large differences in ones actual score relative to another test taker, while in the middle of the distribution, even very large differences in terms of questions answered correctly, would have little effect on the distance between test takers in terms of final score. (Fischer 1996).
This flaw—which is largely built in to all standardized testing—is no doubt why many who administer these tests are forced to admit that even large gaps in scores between any given test-takers mean very little in terms of predicting actual intelligence or ability. For example, according to the Educational Testing Service (ETS), which writes and grades the SAT, score differences of as much as 120 points between any two students should be seen as essentially random and statistically meaningless in terms of demonstrating any actual difference in ability or intellect. (Fischer 1996; Fairtest 1998)
Psychometrics is, in fact inherently illegitimate as a social science. Unlike other disciplines, where the object being studied can be clearly defined, and then observed, psychometricians can't even provide a concrete definition of the thing they claim to be observing and studying—i.e., intelligence. The definition used is intrinsically tautological, or circular, as evidenced by Arthur Jensen—one of the "godfathers" of contemporary "racial scientists" –who is famous for saying "intelligence is what intelligence tests measure." (Fischer 1996) This means that whatever the results of an "intelligence test," scientists in the field of psychometrics will then call those results "proof" of the thing known as intelligence after the fact. But to the extent that a word or concept can't logically be defined by the word or concept itself, this means, in effect that there is no intrinsic meaning to the term "intelligence," other than that meaning given to it by society, or more to the point, those in society who have taken it upon themselves to tell the rest of us what is and is not evidence of intellect..
Attempts by defenders of psychometrics and racial science to precisely define "general intelligence," (what is often called "g") has been a fairly monumental flop. The idea that there is one "factor analytic" form of intelligence that is most important, and that this form can be measured, and is measured, by standard intelligence tests, has been called into question by numerous scholars (Kamin, Lewontin, Rose, Gould, etc), including, recently, Howard Gardner, who has made a strong case—entirely unrebutted in the Bell Curve—for "multiple intelligences," including musical, linguistic, bodily-kinesthetic, logical-mathematical, spatial, interpersonal, and intrapersonal forms. That this model of intelligences is far more accurate than the popular notion of "g," propagated by the Bell Curve, has been demonstrated by research on patients who have suffered some kind of brain damage to a particular portion of the brain, thereby suffering negative effects with regard to certain functions, but no effect whatsoever on other functions. One form of "intelligence," may be rendered less operative or functional at a high level, while others may proceed unscathed, calling into question the factor-analytic notion of "general" intelligence (Tucker 1994).
The convoluted arguments made by defenders of the notion of "g" would be laughable if they hadn't done so much damage to scientific understanding, and had they not been used to push such a reactionary, racist agenda. For example, Arthur Jensen (who will be discussed further below), once claimed that he could determine the amount of "g" (general intelligence) involved in carrying out virtually any task, including those that were not academic in nature. As such, he opined that the Army cooks who were scrambling eggs did so while utilizing "zero g," whereas those Army cooks preparing the jellyrolls were engaged in a process "saturated" with "g." How he made such a culinary-intellectual determination he failed to explain, but the foray into the intellectual properties of a jelly roll chef apparently failed to render Jensen unpalatable to the authors of the Bell Curve, who, like all racial science salespersons of late have relied heavily on his research (as will be examined below)
In the final analysis, to accept psychometrics as legitimate would require accepting the essential validity of standardized tests and various "measures" of intelligence, which, in turn, would require a fundamental belief in a number of contentious and problematic notions, such as the idea that teenagers are inherently smarter than adults. Why? Because teens can often whiz through math and grammar problems, on tests, and in school, while adults have problems with these types of tasks, since they tend to have forgotten many of the things they once knew, provided they don't use those skills on a regular basis.
But do Bell Curve supporters really want to say teens are smarter than adults? Doubtful. And to the extent one could claim they are smarter in some measurable and definable sense (i.e. IQ and test scores), this would simply beg the obvious question as to what kind of intelligence we seek to value and reward. If we think the kind of intelligence being demonstrated by the teenagers on these tests is the truly important kind, then why not turn over all the businesses, state, local and federal government, and educational institutions over to them? Obviously the reason we don't do that is because we realize that this form of intelligence—to the extent we prioritize it at all—is only one part, and in practice a very small part, of what it means to be competent to do a particular thing, and that other factors, like life experience, perspective, and context actually count for more in the world. And if we can realize that simple test scores and academic intelligence are thus a deeply flawed indicator of the broad range of abilities and competencies we wish to promote and reward, thanks to what we can readily see with regard to age, then why doesn't this understanding carry over to the discussion of intelligence and race?
Beyond the inherent problems with psychometrics, the Bell Curve's authors studiously ignore the wealth of counterevidence that refutes the main claims of their work: that genes are the key to IQ, and that IQ is the key to everything from professional success, to income, to educational status, to criminality. On the basis of these two claims, Murray and Herrnstein argue that the gaps in income, educational attainment, and other measures of well-being between blacks and whites in particular are the result of gaps in intelligence, and that these intelligence gaps are themselves racial/genetic, and not mostly caused by the unequal environments to which whites and blacks are often exposed.
But on this score Murray and Herrnstein are flatly wrong. The overwhelming bulk of research has indicated that environmental conditions are instrumental in determining "intelligence," however one may define and measure it. Murray and Herrnstein claim to have taken into consideration environmental influences when figuring their IQ gap data, and still to have found substantial black/white gaps; but in fact, their environmental "controls" were hardly adequate and left many external factors unexamined. When those factors are included in the disaggregation of the Bell Curve "data," it becomes clear: the "racial gaps" in IQ can be explained entirely by factors having nothing to do with racial superiority, inferiority, or, in fact, race at all.
First, consider that much research has demonstrated that IQ and standardized test scores can be depressed dramatically by any number of factors: rebelliousness, apathy, and anxiety among them; and these in turn are directly correlated with conditions like attending poor schools, living in stressful and persistently poor economic conditions, having to deal with significant family stress, high levels of unemployment in one's community, and exposure to forms of oppression and discrimination, like racism. (Fischer et al 1996)
Murray and Herrnstein's claims to have taken into consideration all the relevant environmental factors that could influence IQ, leaving only "natural" and "genetic" explanations for the remaining gaps they discovered are outright distortions. The Bell Curve only looked at three environmental factors that could influence test score results: level of parental education, parental occupational status, and parental income. But these are hardly an exhaustive list of factors that could result in skewed test scores between individuals or groups of people. Many other potentially important factors were ignored, like family size, which would influence standard of living; the quality and kind of schools one had attended or was attending; local job opportunities and income levels in ones larger community, which would effect the cohesion of the neighborhood in which one lives; personal contacts and opportunities for out-of-school enrichment; geographic location.
By ignoring the impact of factors like those above, the Bell Curve's claims to scientific certitude become ever more dubious. Studies have shown (Downey, 1995; and Duncan, Featherman and Duncan, 1972) that having more siblings reduces effective family wealth and economic status and reduces chances for upward mobility, while having fewer siblings frees up space and resources, and allows a child to receive more individualized attention from parents, all of which can be critical in terms of intellectual development.
Furthermore, Murray and Herrnstein ignore the degree to which income similarities between blacks and whites may mask large wealth differentials, and the degree to which wealth reserves are key to educational, housing and overall economic opportunity. Oliver and Shapiro (1995) have shown that even when income between blacks and whites is roughly equivalent, whites, on average, continue to have at least ten times more assets than blacks, allowing them to live in "better neighborhoods," send their children to "better schools," etc., all of which will impact test scores and academic performance. Oliver and Shapiro also make clear that the black/white wealth gap is directly the result of years of racial discrimination and oppression: the result of which has been the "sedimentation" of African Americans on the bottom of the American economic hierarchy. Along these same lines, Massey and Denton (1992) note that even blacks with over $50,000 in annual income are more likely to have kids attending school in poor neighborhoods (and indeed they are more likely to LIVE in poor neighborhoods) than even POOR whites are, further effecting the IQ and other test scores of those children, despite living in middle class and above families.
A close look at Murray and Herrnstein's data from the AFQT actually demonstrates the centrality of environmental influence on IQ score. According to the data, ones AFQT score is highly correlated with the number of years of school one had completed before taking the test. Over two-thirds of those scoring in the top 5% had already had at least one year of college, while half of those on the bottom of the distribution had already dropped out of high school before taking the test. Also, age was shown to be highly correlated—in a negative direction—with AFQT score, once differences in years of education were considered, since older test takers had been out of school longer and thus forgotten much of what they had previously learned. Unless one assumes that folks simply get dumber as they get older, there is little explanation for this phenomenon, other that the obvious environmental issue: if certain skills aren't used over time, ones ability in that area(s) will decline, and certainly having nothing to do with genetics (Fischer 1996)
Homepage / Historical Views
^^ Back to top