Understanding Human History, Michael H. Hart. 2007.
Racial Differences in Intelligence
Section 3 – IQ data: Blacks and whites in the United States
Although the studies differ, virtually all show a large difference between the average scores of American blacks and whites, with the differences clustering around one standard deviation. A few of those studies, including several of the largest ones, are listed in Table 15-1. [footnote 4] It is worth remarking, though, that those studies (like most others) probably underestimate the true value of δ (where δ is defined as the difference between the mean IQ of American whites of European descent and the mean IQ of American blacks). At least two factors contribute to this:
1) Many such studies include only students who are attending school, and omit high school dropouts. This factor causes us to overestimate average IQs. As proportionally more blacks than whites drop out of high school, the effect is to reduce the measured value of δ. [footnote 5]
2) Almost all such studies omit the prison population. Since prisoners, on average, have much lower IQs than the public at large, omitting them has the effect of overestimating the average IQ of every racial group in the United States. Since about 3% of American blacks are prisoners (but less than . of 1% of whites), omitting this factor leads to underestimating δ.
Section 4 – Is the gap narrowing?
An interesting question is whether δ has been getting smaller, as suggested by Hedges & Nowell (1998). [footnote 10] They reached that conclusion by considering the studies listed on the first 6 lines of Table 15-1. As all those studies were confined to high school seniors (and omitted dropouts) the test sample is unrepresentative of American blacks, and probably underestimates δ.
If we instead consider all the entries in Table 15-1, the best fit to the data (using the standard linear regression method) is:
δ = 16.9 – (date – 1917) / 30
which indicates that δ had decreased at an average rate of about one IQ point per 30 years. This suggests that δ was about 14½ points in 1992, and that if the trend had continued then δ would be about 14 points by 2007. However, subsequent data suggests that the black-white gap did not continue to fall after 1990, but instead has risen slightly. [footnote 11]
It should also be noted that in the pre-1946 studies, δ meant the difference between blacks and European whites (since, at the time, there were very few non-European whites in the country). Most of the later studies, though, include many non-European whites and mestizos in the white sample. Since these groups have average IQs about 10 points less than European whites (see section 7, below), in order to make the recent studies comparable to the early ones — which is necessary, if we want to accurately measure black improvement relative to European whites — we should probably add about a point to the recent white figures, and therefore to δ. We conclude that:
1) δ was probably about 17 points in 1917.
2) δ appears to have declined somewhat by 1992.
3) The exact value of δ in 1992 is uncertain. It was probably about 14 points, but may well have been a bit higher.
4) It is probably about 15 points today.
Nature or Nurture?
Section 2 – Arguments that there is a strong genetic component in δ
A) The magnitude of δ has changed very little in the eight decades since it was first observed. Eight decades ago, when δ was first measured (from the tests given by the US Army in World War I), the average difference between the scores of American blacks and whites was about 17 points.
Of course, at that time (during the “Jim Crow” era) there was a vast difference between the environments in which most American blacks were born and raised (and continued to live in as adults) and the environment of most whites. The majority of American blacks lived in great poverty. Their job opportunities were severely limited — sometimes by company policies, sometimes by union rules, sometimes by custom, and sometimes by law. Most American colleges had miniscule numbers of black students, and many would not admit any blacks. In addition, the public schools that most blacks attended were severely underfunded. Because of their poverty, most blacks had poor housing, poor diets, and inferior medical care.
In the intervening decades, the situation of American blacks has improved enormously, both absolutely and in comparison to that of white Americans. Their average income is still considerably lower than that of whites, but there has been a very marked convergence between the environments of the two groups. It is difficult to measure the extent of this convergence precisely, but a reasonable estimate might be that the difference between the environments of typical American whites and blacks is only about one-third (or at most one-half) as great now as it was then. (We might call this fraction the remnant factor, or R.)
It follows that if δ was caused largely by environmental factors then it should have diminished enormously in the course of the last eight decades. Indeed, if hypothesis [a+] is correct, then δ should now be only about one-third to one-half of the original 17 points (i.e., 5.7 to 8.5 points). However, tests taken in recent decades indicate that δ is about 15 points today. [footnote 1] The discrepancy is so large as to clearly refute hypothesis [a+]. Indeed, since δ appears to have diminished by only 2 points during that long interval, hypotheses [a] and [b] are also implausible, and hypothesis [d] appears to be the one that best fits the facts.
Some readers may prefer to see a more detailed mathematical analysis of this point. Back in 1917, δ was approximately 17 points. Call that value δ1917. Let us use G to designate the fraction of δ1917 that was due to genetic factors. (According to hypothesis [a+] that was zero; according to hypothesis [e] it was at least 90%; and according to the other hypotheses it had some intermediate value.) For any assumed values of G and the remnant factor R, we can calculate what δnow (the current value of δ) should be. [footnote 2]
[footnote 2] A reasonable interpolation formula is: δnow = δ1917 × (G + R × [1 – G]). The formula expresses the notion that: (a) The part of δ1917 that was due to genetic factors is unchanged, but (b) the part of δ1917 that was not due to genetic factors (i.e., [1 – G]) is reduced to a fraction R of its original value. If G = 100% (the most extreme hereditarian position), then the formula reduces to δnow = δ1917 ; whereas if G = 0% (the extreme environmentalist position) then the formula reduces to δnow = R × δ1917.
The results of these calculations are shown in Table 16-1. Alternatively, for any assumed values of δnow and of the remnant factor R we can calculate the value of G. The results of those calculations are shown in Table 16-2. [footnote 3]
[footnote 3] Straightforward algebraic manipulations of the formula in the preceding footnote yield: G = ([ δnow/δ1917 ] – R) / (1 – R), which has been used to compute Table 16-2.
From Table 16-2 we see that any value of δ today that is greater or equal to 12 IQ points implies (even with a high estimate of the remnant factor, R) that G is at least 0.412, or 41.2%. This result is strongly inconsistent with hypothesis [a]. If we assume that the most likely value of δ today is 15 points, and that a reasonable estimate of the remnant factor is 1/3, we find that G is about 82%. This line of reasoning therefore supports hypothesis [d], although it does not completely rule out hypothesis [c].
One reason that I have spent so much time on this point is its historical importance. Most of the persons who in the middle of the 20th century urged reforms that would end the unjust treatment of blacks (or at least greatly diminish it) were aware that black scores on IQ tests were far below those of whites, and they believed that the reforms they were advocating would result in the test-score gap diminishing sharply. [footnote 4] Indeed, in 1969, when Arthur Jensen wrote an article 5 entitled “How much can we boost IQ and Scholastic Achievement?” in which he asserted that δ was due in significant part to genetic factors, and that the various civil rights reforms — no matter how justified on moral grounds — would probably not bring black IQs close to the white level, his conclusions were strongly denounced by those who believed in hypotheses [a] or [b]. Over three decades have gone by, and test scores show clearly that Jensen’s predictions were right and the predictions of the extreme environmentalists were wrong.
C) None of the environmental theories that have been suggested adequately explains δ. … In the second place, when we compare blacks and whites whose parents are from the same SES (socioeconomic status, as measured by a combination of affluence, occupation, and education) the difference in mean IQs between the races does not fall to nearly zero — as predicted by extreme environmentalism — but is still about 9 points, or more. [footnote 7] At first sight, this would suggest that genetic factors are responsible for only 9/15, or 60% of the difference in IQ scores. In fact, it indicates that genetic factors are responsible for at least 60% of the difference, since they might also be responsible for much of the difference in parental SES. (The first, incorrect inference is an example of the “sociologist’s fallacy.” [footnote 8])
H) Most of sub-Saharan Africa was extremely primitive before it came into contact with the West two centuries ago. In the 19th century, when European explorers first entered the “secluded zone” of sub-Saharan Africa, they were struck by how extremely primitive the tribes in the region were. [footnote 13] This was not because Europeans were blinded by ethnic chauvinism. When European traders had reached China, they had brought back glowing accounts of Chinese civilization: The Chinese might be heathens, but there was no disputing their wealth, nor the quality of their engineering skills, nor the volume of their literature. In like fashion, Europeans who had seen Japan, India, Persia, and the Arab world did not dispute the quality of their architecture and their textiles, nor the elegance of their art, nor the ability of their leaders to capably administer a large kingdom.
Of course, it might be said that all those regions had had the benefit of at least indirect contact with the West. However, when European explorers reached the New World — which had been even more isolated from the rest of the world than sub-Saharan Africa had been — they were dazzled by the large, prosperous cities they found there. Nowhere in the secluded zone of sub-Saharan Africa have archaeologists found anything to compare with Machu Picchu in Peru, nor the ruins of Tikal in Yucatan, nor of Teotihuacan in Mexico, nor even with the statues found on tiny, remote Easter Island in the Pacific.
Furthermore, the countries in sub-Saharan Africa remain backward today, compared with third-world countries on other continents. Most of them have largely discarded the democratic institutions bequeathed them by the European countries that once ruled them, and on virtually every objective measure of social, cultural, or economic well-being they rank at the bottom. Countries such as Angola, Gabon, Zimbabwe, Guinea, and Uganda are separated from each other by thousands of miles; they differ in topography, climate, language, religion, and history. The only factors they have in common are race and the low average intelligence of their inhabitants.
J) The relative performance of blacks on different intelligence tests does not depend on the cultural content of the test, but rather it is closely correlated with the extent to which the test is “g-loaded.” It was pointed out in section 3-2 that not all types of intelligence tests are equally correlated with the g factor. Tests that measure short-term memory, for example, are only weakly “g-loaded,” while tests of paragraph comprehension or of series completion are strongly g-loaded. If the low test scores of blacks were primarily caused by poor schooling, then we would expect them to do relatively worse on tests that depended on acquired knowledge (such as vocabulary tests), and relatively better on tests (such as backward digit span) in which acquired knowledge plays no part. In fact, however, this is not the case. Instead, the relative size of the black-white difference in various mental tests depends primarily upon the g-loading of the individual tests: Blacks do worst on those tests that are strongly correlated with general intelligence. This is true whether the tests are verbal or non-verbal, written or oral, timed or untimed, culture-loaded or culture-reduced. [footnote 15]
Section 3 – Some opposing arguments
A) “Scientists agree that δ is completely due to environmental factors.” … A survey conducted by Snyderman and Rothman (1988) showed that although hypothesis [a+] (extreme environmentalism) was the one normally presented in the media as the correct view, in actual fact only 17% of experts believed it to be correct. The majority of expert opinion supported the hypothesis that both genetic and environmental factors play a role in δ. [footnote 16]
B) “Whites and blacks differ in only a small fraction of their genes.” This is one of the most common arguments presented, but it is invalid. [footnote 17] The error comes from assuming that small differences in the input to a system must yield small differences in the system’s output. That assumption is false: On the contrary, it is often the case that small differences in the input result in large differences in the final outcome. Here are a few examples:
1) A batter in a baseball game hits a 400-foot home run. Had the batter swung his bat just one inch lower, it would instead have resulted in a high pop fly, and an easy out.
2) Those who have written computer programs are aware that typing a single word incorrectly can cause a lengthy computer program to malfunction completely. Indeed, the omission of a single letter — or a single comma, or even a single blank space! — can have that effect.
3) It has often been pointed out that human beings and chimpanzees differ in less than 2% of their DNA; nevertheless, the difference in intelligence between the two species is enormous.
4) Within the human species, a large part of the difference between males and females is due to a single chromosome, indeed to a single gene (the SRY gene). Furthermore, there are many genetic diseases that are caused by a single gene, and some of them (like Huntington’s chorea) are lethal diseases. [footnote 18]
C) “The difference in mean IQ between the races is much less than the spread of IQs within each race.” This is perfectly true, but irrelevant, since it does not imply that the difference in IQ is not genetic in origin. After all, the difference between the mean height of human males and human females is only a few inches — which is much less that the variation in height found within each sex — but the male-female difference in height is almost entirely genetic in origin.
D) “There are large differences in wealth and income between whites and blacks, and these differences are responsible for the test-score gap.” This is by far the most common non-genetic explanation of δ, but it is badly contradicted by the data. For example:
1) Most of the racial difference in average IQ scores remains when we compare black and white children whose parents have similar incomes (see footnote 8).
2) Blacks whose parents have incomes over $70,000/year obtain (on average) lower scores on the SAT than whites or Asians whose parents have incomes under $20,000/year (see footnote 7).
3) The mean IQ gap between blacks and whites who were adopted into the same families when they were very young children and reared together — thereby eliminating differences in parental income — is as large as the black-white IQ gap in the general population. (See Table 16-3.)
I) “Institutional racism is the cause of low black IQ scores.” This is perhaps the weakest of the hypothetical arguments, because it is so vague. What it seems to mean is that “black IQ scores and intellectual achievements are lower than those of whites, and we can’t point to any testable environmental explanation.”
J) “Among American blacks, there is no correlation between IQ and skin color (or percentage of African ancestry). Unlike the last five arguments, this one is testable and would be important if it was correct. Richard Nesbitt, [footnote 23] citing the data compiled by Audrey Shuey [footnote 24] claims that such studies show no significant positive correlation between lighter skin color and higher IQ. However, direct examination of Shuey’s data shows the contrary. Shuey examined studies made between 1913 and 1964 by 13 different researchers, each of whom tested black-white hybrids and attempted to determine if light skin color (or other evidence of Caucasoid ancestry) correlated with the results of IQ tests. Twelve of those studies (with a combined sample size of 6,520) showed a significant positive correlation between Caucasoid ancestry and IQ scores — a large correlation in seven of the studies — whereas only four small studies (with a combined sample size of 460) failed to do so.
K) … It should be added that more recent data indicates that the gap between the test scores of whites and blacks is no longer decreasing, and may even be increasing. For example, a study released in 2000 by the US Department of Education found that, on average, 17-year-old blacks read only as well as 13-year-old whites; and among 17-year-old whites and blacks the gap in science scores in 1999 (52 points) had widened slightly in the past few years and was almost the same as the 54-point difference that was recorded in 1969. [footnote 27] According to data released by the College Board, the gap in SAT scores between blacks and whites actually increased slightly between 1990 and 2000. The gap in the verbal portion of the examination increased from 91 points to 94 in the course of the decade, and the gap in the mathematics portion increased from 96 points to 104 points. [footnote 28]
L) “A study of the offspring of German women and American soldiers after World War II showed that those children with black fathers had the same mean IQ as those with white fathers.” That is true, [footnote 29] but there is no reason to assume that the black soldiers involved were a random sample of American blacks, or even of black soldiers in the US army. On the contrary, there is strong reason to believe that they were not a random sample.
We know that females are commonly selective as to whom they mate with, and intelligence is frequently an important factor in that selection. If, on average, the German females involved selected just as strongly for intelligence whether they were dealing with black or white servicemen, then we should not expect there to be any difference between the mean IQs of the white and mixed-race children. In any event, since the data presented in the study does not include any information about the IQ of the black fathers involved (nor about the white fathers, for that matter), no reliable conclusions can be drawn from the results.
N) “The backwardness of sub-Saharan Africa was entirely due to deficiencies in the natural environment, such as a dearth of domesticable plants and animals, and the direction of the principal geographic axes in Africa.” This argument was presented at length by Jared Diamond in his best-selling book Guns, Germs, and Steel (1997). Professor Diamond’s hypothesis is discussed in more detail in chapter 24 of this book, where it is shown that a comparison between the relative cultural and intellectual achievements of sub-Saharan Africa and Mesoamerica clearly refutes his theory.
In any event, the most that Dr. Diamond’s argument could do is to counter one of the arguments supporting the notion that genetic differences are involved in δ (the one concerning the backwardness of sub-Saharan Africa). It would not affect the other arguments that have been made against very strong or extreme environmentalism.
An Alternative Explanation
Section 3 – Critique: Comparison of data for SSA and Mesoamerica
There is certainly something to be said for Dr. Diamond’s thesis. Eurasia, and particularly the Middle East, did have a far greater supply of useful and easily domesticable plants and animals than any other region. It is also true that both Australia and the United States were badly lacking in such species. However, the facts do not support his theory when it is applied to a comparison between sub-Saharan Africa and Mesoamerica.
1) Flora. Dr. Diamond rightly stresses the importance of cereal crops in the rise of agriculture. The only useful wild cereal that grew in Mesoamerica was teosinte, the ancestor of corn. However, teosinte is not nearly as nutritious as wild wheat, and it was far less amenable to domestication. In contrast, sub-Saharan Africa possessed five useful cereal crops: sorghum, bulrush millet, finger millet, teff, and African rice. [footnote 4] It seems, therefore, that SSA had an advantage in this regard. (SSA also had various useful non-cereal crops, including yams, cowpeas, watermelon, oil palm, and groundnuts; but these were balanced by the availability in Mesoamerica of beans, squash, tomatoes, and sweet potatoes.)
2) Fauna. Few regions contain as many species of large animals as SSA, but Dr. Diamond insists that none of the wild species there are domesticable. For example, he states that zebras, although very similar to horses in anatomy, are hopelessly irascible, and points out that recent attempts to tame them have all failed. I find that example unconvincing. Wild horses were long considered to be untamable animals, as was the aurochs (the wild ancestor of domestic cattle), but both were domesticated in time. Until we have spent at least a few centuries trying to domesticate zebras, we should not rush to the conclusion that such attempts are hopeless.
However, even if it turns out that zebras are untamable, it still would not be true that Africa did not possess any potentially useful farm animals. The wild ancestor of domestic cattle — the most useful of all farm animals — was indigenous to North Africa, [footnote 5] and domestic cattle were being used in the Sahara by 5.5 kya, probably earlier, [footnote 6] and south of the Sahara by 5 kya. [footnote 7] (The Sahara was much wetter several thousand years ago than it is now. [footnote 8]) In addition, domestic sheep and goats were introduced into Africa by 7.5 kya, [footnote 9] and their use had spread south of the Sahara by 4 kya. [see also “Race”, Baker, chapter 19-20, pages 343-400]
Mesoamerica, on the other hand, did not have a single large domesticable animal, since most of the megafauna in the Western Hemisphere had been killed off by the Paleo-Indians by 11 kya. As regards fauna, therefore, SSA had a great advantage over Mesoamerica.
3) Orientation of geographic axes. In Mesoamerica, the longest east-west span (from the eastern tip of Yucatan to Mazatlan, on the west coast of Mexico) is only 1300 miles. Contrast this with sub-Saharan Africa, where a vast stretch of savannah (the Sudan, situated between the Sahara and the tropical rainforest) stretches 3500 miles in an east-west direction, from the highlands of Ethiopia to Senegal. It is clear that transmission of technology and domesticates could — and repeatedly did — take place along the Sudan, and also across Ethiopia.
Furthermore, sub-Saharan Africa was not completely cut off from Eurasia, and some important aspects of Eurasian technology and culture did reach SSA. Techniques of pottery-making, bronze working, and ironworking [see chapter 41, section 5, page 311] reached SSA from the Middle East, as did the use of domesticated camels. (In addition, as already mentioned, domestic sheep and goats were introduced into SSA from the Middle East by 4 kya.) In contrast, prior to 1492, no Neolithic flora, fauna, or technology ever spread from the Old World to the Western Hemisphere.
We see, therefore, that the geographic factors mentioned by Jared Diamond strongly favored SSA over Mesoamerica. Using his criteria, civilization should have begun earlier in SSA than it did in Mesoamerica, and it should have progressed more there (prior to the European expansion of modern times) than it did in Mesoamerica.
In fact, though, by 1000 AD, Mesoamerica was far more advanced than SSA was, or ever had been. For example, Mesoamericans had originated writing on their own, had constructed many large stone structures, and had built large cities (rivaling any existing in Europe, and far larger than any in sub-Saharan Africa). Furthermore, the Mayan achievements in mathematics and astronomy dwarf any intellectual achievements in SSA.
We must therefore conclude that, although Guns, Germs, and Steel is an informative book, the obvious superiority of Mesoamerican technology to that of sub-Saharan Africa appears to be a fatal blow to the main arguments presented in it. In contrast, this book provides a simple explanation for that superiority.
Of course, even if Dr. Diamond had demonstrated that the backwardness of SSA could be completely explained by geographic factors, that would still not prove there is no genetic component to the observed difference in mean IQ between blacks and whites. Many other arguments were presented in chapter 16 for believing that that difference is due in part to genetic factors. At best, his argument would remove only one of the reasons for reaching that conclusion; the other arguments would remain, and are quite sufficient.