Race and intelligence

• 1. Introduction

  ›

  The debate over race and intelligence, meaning the claim that groups differ in intellect along racial lines, has run for as long as the modern idea of race has existed. One of its sharpest early statements appeared in print in 1916. The psychologist Lewis Terman wrote that Mexican-Americans, African-Americans and Native Americans showed a mental "dullness that seems to be racial, or at least inherent in the family stocks from which they come." Terman's confidence rested on a new kind of exam, and that exam would soon be used to shape laws on immigration and segregation. Over the following century, the same argument returned in waves, often backed by private money and by bestselling books. This documentary asks how the claim first gained authority and who kept it alive. It asks what the words race and intelligence mean to the scientists who study them. It asks what test scores actually show, and what researchers found when they went looking for a genetic cause. Last, it asks whether this research should be done at all.

  56912

• 2. Prejudice Meets Testing

  ›

  Arthur de Gobineau, a French racial thinker, built his ideology of white supremacy on the assumption that black people were innately inferior to white people. He had company in respectable circles. Thomas Jefferson, a slave owner and an Enlightenment thinker, believed black people were innately inferior in both physique and intellect. Claims of intellectual inferiority were used to defend slavery, colonialism, social Darwinism and racial eugenics, and to justify British wars and colonial campaigns in Asia.

  Frederick Douglass, an autodidact and abolitionist, stood as a living counterexample to those claims. So did the sociologist W. E. B. Du Bois and the poet Paul Laurence Dunbar. In Britain, Japan's military victory over Russia in the Russo-Japanese War began to reverse stereotypes of "oriental" inferiority.

  Between 1905 and 1908, Alfred Binet and Théodore Simon in France developed the first practical intelligence test, intended to help place schoolchildren. Binet warned that its results should not be read as a measure of innate intelligence or used to label anyone permanently. Terman later revised the test for English speakers and introduced IQ scoring. It was published as the Stanford-Binet Intelligence Scales.

  Robert Yerkes developed a separate set of tests that the US Army used to evaluate draftees in World War I. These produced the first test results in the United States that showed score differences between population groups. The psychologists and eugenicists Henry H. Goddard, Harry H. Laughlin and Princeton professor Carl Brigham drew sweeping conclusions from the data. They wrote that people from southern and eastern Europe were less intelligent than native-born Americans or Nordic immigrants, and that black Americans were less intelligent than white Americans.

  Madison Grant, a conservationist and theorist of scientific racism, was among the anti-immigration activists who publicized the results. He considered the so-called Nordic race superior but threatened by immigration from "inferior breeds." Brigham's book A Study of American Intelligence used the Army data to argue for stricter immigration, limited to countries considered part of the "Nordic race." In the 1920s, eugenics lobbyists argued that the scores proved innate inferiority and used that belief to defend racial segregation. Virginia's 1924 Racial Integrity Act wrote the one-drop rule of "racial purity" into state law.

  Many scientists objected. They pointed out that environment shaped the results, for example when test takers spoke English as a second language. By the mid-1930s, many American psychologists held that environmental and cultural factors played the dominant role. Brigham himself repudiated his earlier arguments, saying he had come to realize the tests did not measure innate intelligence.

  Grant's writings had meanwhile influenced German Nazi claims that the "Nordics" were a "master race." As American public sentiment turned against the Germans, the idea of racial differences in intelligence came to look problematic. The anthropologists Franz Boas, Ruth Benedict and Gene Weltfish did much to show that racial hierarchies of intelligence were unscientific. Even so, a powerful eugenics and segregation lobby, funded largely by the textile magnate Wickliffe Draper, kept citing intelligence studies in support of eugenics, segregation and anti-immigration laws.

  23415

• 3. The Funded Debate

  ›

  In 1937, Wickliffe Draper founded the Pioneer Fund, with a stated mission "to advance the scientific study of heredity and human differences." The fund became a central driver of research on race and intelligence. The Southern Poverty Law Center lists it as a hate group, and numerous scholars have criticized it for promoting scientific racism, eugenics and white supremacy.

  Desegregation gained traction in the United States in the 1950s, and the argument about black intelligence resurfaced with it. Audrey Shuey, funded by the Pioneer Fund, published a new analysis of Yerkes' tests and concluded that black people were of inferior intellect. Segregationists used her study to claim that black children were better off schooled apart from white children.

  In the 1960s, William Shockley publicly defended the view that black children were innately unable to learn as well as white children. Arthur Jensen wrote a Harvard Educational Review article titled "How Much Can We Boost IQ and Scholastic Achievement?" It questioned the value of compensatory education for African-American children. Jensen suggested that poor school performance reflected an underlying genetic cause rather than a lack of stimulation at home.

  The Bell Curve, published in 1994 by Richard Herrnstein and Charles Murray, supported Jensen's general viewpoint and set off another public fight. Many of its references and sources were advocates of racial hygiene whose work was funded by the Pioneer Fund and published in the fund's affiliated journal, Mankind Quarterly. Ashley Montagu observed: "The University of California's Arthur Jensen, cited twenty-three times in The Bell Curves bibliography, is the book's principal authority on the intellectual inferiority of blacks."

  A statement of support titled "Mainstream Science on Intelligence" ran in The Wall Street Journal. Critics answered with the American Psychological Association's statement "Intelligence: Knowns and Unknowns." Several books followed, among them The Bell Curve Debate (1995), Inequality by Design (1996) and a second edition of Stephen Jay Gould's The Mismeasure of Man (1996).

  J. Philippe Rushton headed the Pioneer Fund until his death in 2012. Jensen and Rushton jointly published a 2005 review article arguing that the average IQ difference between blacks and whites is partly genetic. Jensen had received $1.1 million in grants from the fund. Richard Lynn, a psychologist on the fund's board, and the political scientist Tatu Vanhanen wrote IQ and the Wealth of Nations and IQ and Global Inequality. These books tried to link national IQ scores to the size of national economies, and several researchers criticized their conclusions.

  On the 27th of July 2020, the European Human Behavior and Evolution Association issued a formal statement opposing use of Lynn's national IQ dataset, citing methodological concerns. It concluded: "Any conclusions drawn from analyses which use these data are therefore unsound, and no reliable evolutionary work should be using these data."

  On the 24th of January 2026, the New York Times reported that a group of fringe researchers with a long history of racist pseudoscience had used deception to reach protected National Institutes of Health data. The group included Bryan Pesta, Jordan Lasker, John G. R. Fuerst and Emil Kirkegaard, and it was funded in part by the Pioneer Fund. They drew on the Adolescent Brain Cognitive Development Study to produce at least 16 papers claiming biological evidence for intelligence differences between races. Mainstream geneticists rejected the work as biased and unscientific, yet the papers sparked a new wave of online disinformation. Pesta, the only one of the group to hold a university post, was fired from Cleveland State University for scientific misconduct.

  Every one of these arguments depends on two words, race and intelligence, that scientists do not define in the same way.

  6910117

• 4. Words In Dispute

  ›

  In 1904, the English psychologist Charles Spearman noticed that scores on different kinds of IQ tests correlate with one another. He proposed an underlying factor, which he called "g" or general intelligence, and he thought of it as innate. Arthur Jensen later became a leading proponent of the idea. Studies showing that education and a change of environment can significantly raise IQ test results have contradicted this view.

  There is no agreement on how to define intelligence, and not everyone accepts that a single figure can capture it. Some critics note that different societies value and cultivate different skills, so intelligence may vary by culture. Other psychometricians argue that, whether or not a general factor exists, test performance depends heavily on knowledge gained through earlier exposure to similar tasks. In that case, comparing people with widely different life experiences says little about their relative innate potential.

  A 2023 consensus report from the National Academies of Sciences, Engineering, and Medicine calls race "a socially constructed designation, a misleading and harmful surrogate for population genetic differences." The American Anthropological Association reached a similar position in 1998. It said human populations are not "unambiguous, clearly demarcated, biologically distinct groups." In 2019, the American Association of Physical Anthropologists declared that race "does not provide an accurate representation of human biological variation."

  Genetic clustering does not rescue the category. Weiss and Fullerton observed that a sample containing only Icelanders, Mayans and Maoris would yield three distinct clusters. Every other population could then be described as a blend of those three. The anthropologist C. Loring Brace, the philosophers Jonathan Kaplan and Rasmus Winther, and the geneticist Joseph Graves argue that the clusters depend on which populations a researcher chooses to sample. Kaplan and Winther add that racial groups differ in allele frequencies, but many other genetic patterns cut across racial lines.

  In intelligence research, race is almost always taken from self-reports, not from genetic analysis. David Rowe defended this because genetic markers alone ignore the cultural, behavioral and sociological variables that distinguish racial groups. Hunt and Carlson countered that self-identification is a surprisingly reliable guide to genetic composition. Sternberg and Grigorenko replied that the study they cited showed only that ancient geographic ancestry is associated with self-identification, not that biological race exists.

  Whatever the definitions, the test scores themselves have been gathered on an enormous scale.

  1821249291

• 5. Scores And The Gap

  ›

  A 2001 meta-analysis covering 6,246,729 people tested for cognitive ability or aptitude found a black-white difference in average scores of 1.1 standard deviations. College and graduate admission tests pointed the same way. The Scholastic Aptitude Test sample had 2.4 million test takers, and the Graduate Record Examination sample had 2.3 million. Job applicant tests in corporate settings covered 0.5 million people, and military tests covered 0.4 million. In the United States, Asians on average score higher than White people, who tend to score higher than Hispanics, who tend to score higher than African Americans. Yet there is more variation within each racial group, for instance when divided by income, than between groups.

  After The Bell Curve appeared, the American Psychological Association formed a task force of eleven experts. Their 1996 report, "Intelligence: Knowns and Unknowns," reaffirmed that differences within groups are much wider than differences between them. It urged careful scrutiny of claims about ethnic differences, since such claims had been used to justify discrimination. It also noted that racial categories are neither consistently applied nor homogeneous.

  In the United Kingdom, some African groups outperform the overall population in school. In 2010-2011, white British pupils were 2.3% less likely than the national average to gain five A* to C grades at GCSE. Pupils of Nigerian origin were 21.8% more likely than average, and those of Ghanaian origin 5.5% more likely. Pupils of Somali origin were 23.7% less likely, and those of Congolese origin 35.3% less likely.

  A 2014 comparison of Key Stage 2 maths results in England found a pass rate of 86.5% for white British pupils (N = 395,787) and 85.6% for Black-African pupils (N = 18,497). Yoruba, Igbo, Hausa, Akan, Ga, Swahili, Edo, Ewe and Amharic speakers each beat the white British average, and Hausa, Igbo, Yoruba and Amhara speakers topped 90%. In 2017-2018-44.3% of Black-African pupils earned a strong pass in English and maths GCSE, against 42.7% of white pupils.

  The researcher James R. Flynn gave his name to the Flynn effect, the substantial rise in raw IQ scores seen in many parts of the world during the 20th century. In the United States, the rise was continuous and roughly linear until about 1998, when gains stopped and some tests showed declines. On some IQ tests, the average score for black Americans in 1995 matched the average for white Americans in 1945. As two academics put it, "the typical African American today probably has a slightly higher IQ than the grandparents of today's average white American."

  Flynn argued that such large shifts between one generation and the next pointed to an environmental explanation, since genes could hardly change that fast. His analysis showed that environment can move IQ scores by as much as 1 standard deviation, a scale of change that had previously been doubted.

  The American black-white gap itself narrowed in the last decades of the 20th century. Vincent reported in 1991 that it was shrinking among children but staying constant among adults. A 2006 study by Dickens and Flynn estimated that the gap closed by about 5 or 6 IQ points between 1972 and 2002, a reduction of about one-third. The educational achievement disparity shrank in the same period. Flynn and Dickens concluded: "The constancy of the Black-White IQ gap is a myth and therefore cannot be cited as evidence that the racial IQ gap is genetic in origin."

  A shrinking gap raises the question of what, in children's surroundings, was changing. Childhood lead exposure, tied in the research to homes in poorer areas, is one of the first suspects.

  8932355173

• 6. Environmental Causes

  ›

  Childhood exposure to lead correlates with an average IQ drop of 7 points, and iodine deficiency lowers it by about 12. The first two years of life are critical for malnutrition, and its effects are often irreversible. Nicholas Mackintosh points out that infant mortality is about twice as high for American black people as for white people, and low birth weight is twice as prevalent. White mothers are twice as likely to breastfeed. Reichman estimates, however, that no more than 3 to 4% of the gap can be explained by disparities in low birth weight.

  The Copenhagen consensus of 2004 stated that lack of iodine and iron is implicated in impaired brain development. About one-third of the world's population is affected by iodine deficiency, and an estimated 40% of children aged four and under in developing countries have anaemia from too little iron. A 2013 World Health Organization meta-analysis found that breastfeeding was associated with IQ gains of 2.19 points after controlling for maternal IQ. Colen and Ramey compared siblings within families and found the link between breastfeeding duration and IQ scores became insignificant. They suggest that selection into infant feeding practices by race and socioeconomic status explains much of the apparent benefit. Some recent research also proposes that infectious diseases, more prevalent in non-white populations, retard brain development.

  The Abecedarian Early Intervention Project, an intensive early childhood education program, raised the average IQ of the black children in it by 4.4 points at age 21, compared with controls. Arthur Jensen accepted that it showed education can affect IQ. He maintained that no program had narrowed the black-white gap by more than a third, so education alone was unlikely to explain it.

  Joseph Fagan and Cynthia Holland tested the effect of prior exposure to the kinds of tasks IQ tests contain. After African Americans were prepared on such tasks, the researchers found no difference in test performance between them and white test takers. A separate study by Hala Elhoweris and colleagues found that teachers' decisions to refer students to gifted programs were influenced in part by ethnicity. David Marks argues that IQ differences track differences in literacy.

  A 2006 review found that socioeconomic factors account for slightly less than half of one standard deviation of the gap. Among low-income groups in the United States, genetic differences explain a smaller share of IQ variance than among high-income groups. The review also notes that adoption studies mostly sample high and upper-middle income families, which inflates estimated genetic effects. Children adopted from lower-class into middle-class homes gain 12 to 18 IQ points compared with similar children who stay in low-income homes. A 2015 study found that family income, maternal education, maternal verbal ability, learning materials in the home, parenting factors, birth order and birth weight together accounted for the black-white gap.

  Some studies conclude that IQ tests may be biased against certain groups, and scores from outside the United States and Europe are especially hard to compare. The American Psychological Association's 1996 report says that unfamiliarity with test material can produce substantial differences. It also says the tests predict future achievement equally well for black and white Americans, and in that sense are unbiased. Mackintosh reached a similar view about prediction in his 1998 book IQ and Human Intelligence.

  Stereotype threat is the fear of confirming a stereotype about one's group, and that fear can impair performance. Testing situations that highlight the measurement of intelligence tend to lower scores for groups expected to score lower. Mackintosh considers it beyond reasonable doubt that stereotype threat contributes to the black-white gap. A 2003 study found that stereotype threat and the educational attainment of children's fathers partly explained the gap. Research published in 1997 found that part of the gap reflects differences in test motivation. Some critics suspect publication bias, though numerous meta-analyses and reviews have found significant evidence of stereotype threat effects.

  With environment taking up so much of the explanation, the remaining question is whether genes are doing hidden work.

  3940389545

• 7. Genetic Evidence

  ›

  IQ differences between individuals have a large hereditary component, and that fact has often been misapplied. Twin studies report high heritability. Within US populations, genetic factors account for 30-50% of IQ variance in early childhood and 75-80% by late adolescence. Heritability, however, describes variation within one population and says little about what causes a trait or whether it can change.

  Take two gardens of cornstalks. Within each garden, the difference between tall and short stalks is fully heritable, because stalks genetically disposed to grow tall do so. But the difference between the two gardens is entirely environmental, because they were fed different nutrients. High heritability within races therefore cannot show that differences between races are genetic. Today, the scientific consensus is that genetics does not explain differences in IQ test performance between groups.

  The geneticist Alan R. Templeton argues that racial groups are neither sub-species nor distinct evolutionary lineages. There is no genetic material shared by all Africans or by all Europeans, so he calls the search for genes behind the black-white gap flawed from the start. Others reply that genetic cluster analysis could in principle show that certain variants are more frequent in African populations. Mackintosh accepts that a biological basis cannot be ruled out on a priori grounds.

  One reviewer observed that "no genes related to difference in cognitive skills across the various racial and ethnic groups have ever been discovered." That reviewer added that tomorrow "has not come for quite some time now." A 2012 review concluded that known environmental factors can explain the entire gap, and Mackintosh found this plausible. Research on genetic loci tied to individual differences in IQ has shown promise. The editorial board of Nature separated that work from what it acknowledged as racist pseudoscience, and called the idea of genetically determined racial differences in intelligence definitively false. An analysis of polygenic scores from the 1000 Genomes Project found no evidence that intelligence was under diversifying selection in Africans and Europeans.

  Spearman's hypothesis, formalized by Arthur Jensen, says the black-white gap is largest on the tests that best measure g. Jensen devised the "method of correlated vectors" to test it. He and Rushton took the pattern as evidence that the cause of g and the cause of the gap are the same, and they meant genetic. Mackintosh notes that the tests Jensen and Rushton rate as most g-loaded, such as the Wechsler test, show the biggest gains for black test takers through the Flynn effect. That suggests those tests are the most sensitive to environment.

  In 2006, Flynn and Dickens found that the g gap declined "virtually in tandem with the IQ gap." In a 2014 suite of meta-analyses with Jan te Nijenhuis and Daniel Metzen, Flynn showed the same negative correlation between IQ gains and g-loading for deficits with known environmental causes. These included iodine deficiency, prenatal cocaine exposure, fetal alcohol syndrome and traumatic brain injury.

  The Minnesota Transracial Adoption Study of 1976 examined 122 adopted children and 143 nonadopted children raised by advantaged white families. When the children were restudied ten years later, white children scored higher than black children at both age 7 and age 17. The authors, Scarr and Weinberg, acknowledged confounding factors and said the results supported neither the hereditarian nor the environmentalist view.

  The Eyferth study followed the out-of-wedlock children of black and white soldiers stationed in Germany after World War II, raised by white German mothers. It found no significant difference in average IQ between the groups. In British long-stay residential nurseries, two of three tests showed no significant difference among black, white and mixed-race children. Moore observed that 23 black and interracial children raised by white parents scored a mean of 117, against 104 for 23 age-matched children raised by black parents. Frydman and Lynn reported a mean IQ of 119 for Korean infants adopted by Belgian families, still 10 points above Belgian children after correcting for the Flynn effect. Mackintosh considers the adoption evidence inconclusive and fully compatible with a 100% environmental explanation.

  African Americans typically have about 20% of their genome from European ancestors. If the gap were partly genetic, those with more European ancestry might score higher. Templeton summarizes admixture studies as showing no significant correlation between any cognitive ability and degree of African or European ancestry. Loehlin found a low correlation in some blood-group studies, and he judged the body of admixture research inconclusive.

  Mental chronometry measures the time between a stimulus and a person's response. Results comparing black and white groups are inconsistent. Nisbett notes that Chinese Americans had longer reaction times than European Americans despite higher IQs, and that movement time is faster on average for black people. Another critic points out that rhesus monkeys have shorter reaction times than American college students.

  A moderate correlation between brain size and IQ has been reported between individuals in the same group. Newborn black and white children have the same average brain size, which suggests environment explains later differences. Men and women also differ in brain size without differing in IQ. Archaeological evidence does not support the claim by Rushton and others that black people's cognitive ability was inferior in prehistoric times.

  Beyond the data lies a harder question, about whether this research should be pursued.

  7982768173

• 8. Research Ethics

  ›

  Nature published two editorials in 2009 on the ethics of this research, one by Steven Rose and one by Stephen J. Ceci and Wendy M. Williams. Debate over the question had already appeared in the American Psychological Association's 1996 report and in guidelines proposed by Gray and Thompson and by Hunt and Carlson. Rose maintains that the history of eugenics makes the field hard to reconcile with current ethical standards for science.

  James R. Flynn takes the opposite view. He argues that a ban on research into possibly poorly conceived ideas would have prevented much valuable work on intelligence testing, including his own discovery of the Flynn effect.

  Many voices call for action to close the gaps, whatever the cause. Flynn writes that "America will have to address all the aspects of black experience that are disadvantageous, beginning with the regeneration of inner city neighborhoods and their schools." Especially in developing nations, society has been urged to make preventing childhood cognitive impairment a high priority. The preventable causes on that list include malnutrition, meningitis, parasites, cerebral malaria, drug and alcohol exposure in the womb, newborn asphyxia, low birth weight, head injuries, lead poisoning and endocrine disorders.

  8884788374