African Americans currently score lower than European Americans on vocabulary, reading, and math tests, as well as on tests that claim to measure scholastic aptitude and intelligence. This gap appears before children enter kindergarten, and it persists into adulthood. It has narrowed since 1970, but the median American black still scores below 75 percent of American whites on most standardized tests. On some tests the typical American black scores below more than 85 percent of whites.
The black-white test score gap does not appear to be an inevitable fact of nature. It is true that the gap shrinks only a little when black and white children attend the same schools. It is also true that the gap shrinks only a little when black and white families have the same amount of schooling, the same income, and the same wealth. But despite endless speculation, no one has found genetic evidence indicating that blacks have less innate intellectual ability than whites. Thus while it is clear that eliminating the test score gap would require enormous effort by both blacks and whites and would probably take more than one generation, we believe it can be done. This conviction -- supported at greater length in the new collection of studies we have edited, The Black-White Test Score Gap, soon to be published by the Brookings Institution -- rests mainly on three facts:
- IQ and achievement scores are sensitive to environmental change. Scores on nonverbal IQ tests have risen dramatically throughout the world since the 1930s. The average white scored higher on the Stanford-Binet test in 1978 than 82 percent of whites who took the test in 1932.
- Black-white differences in academic achievement have also narrowed throughout the twentieth century. The best trend data come from the National Assessment of Educational Progress (NAEP), which has been testing seventeen-year-olds since 1971 and has repeated many of the same items year after year. From 1971 to 1996, the black-white reading gap shrank by almost half and the math gap by a third. [See "The Diminished Gap in Reading and Math Scores."] According to a study by two sociologists, Min-Hsiung Huang and Robert Hauser, the black-white vocabulary gap also shrank by half among adults born between 1909 and 1969.
- When black or mixed-race children are raised in white rather than black homes, their pre-adolescent test scores rise dramatically. Black adoptees' scores seem to fall in adolescence, but this is what we would expect if, as seems likely, their social and cultural environment comes to resemble that of other black adolescents and becomes less like that of the average white adolescent.
|The Diminished Gap in Reading and Math Scores
Black and White 17-Year-Olds
|Source: National Assessment of Educational Progress. Tests in all years are in a common metric and have been rescaled so that the 1996 population mean is zero and the 1996 standard deviation is 1.00.|
WHY TEST SCORES MATTER
In a country as racially polarized as the United States, no single change taken in isolation could possibly eliminate the entire legacy of slavery and Jim Crow or usher in an era of full racial equality. But if racial equality is America's goal, reducing the black-white test score gap would probably do more to promote this goal than any other strategy that could command broad political support. Reducing the test score gap is probably both necessary and sufficient for substantially reducing racial inequality in educational attainment and earnings. Changes in education and earnings would in turn help reduce racial differences in crime, health, and family structure, although we do not know how large these effects would be.
This judgment contradicts the conclusion of Inequality, a study published in 1972 by one of us (Christopher Jencks), which argued that reducing cognitive inequality would not do much to reduce economic inequality. The reason for the contradiction is simple: the world has changed. In 1972, the best evidence about what happened to black workers with high test scores came from a study by Phillips Cutright, who had analyzed the 1964 earnings of men in their thirties who had taken the Armed Forces Qualification Test (AFQT) between 1949 and 1953. Overall, employed black men earned 57.5 percent of what whites earned. Among men with AFQT scores above the national average, black men earned 64.5 percent of what whites earned. In such a world, eliminating racial differences in test performance did not seem likely to reduce the earnings gap very much.
Today's world is different. The best recent data on test scores and earnings come from the National Longitudinal Survey of Youth (NLSY), which gave the Armed Services Vocational Aptitude Battery to a national sample of young people in 1980. Among employed men who were 31 to 36 years old in 1993, blacks earned 67.5 percent of what whites earned -- a modest but significant improvement over the situation in 1964. The big change occurred among blacks with test scores near or above the white average. Among men who scored between the thirtieth and forty-ninth percentiles nationally, black earnings rose from 62 to 84 percent of the white average. Among men who scored above the fiftieth percentile, black earnings rose from 65 to 96 percent of the white average. [See "More-Equal Scores Now Bring More-Equal Earnings," below.] In this new world, raising black workers' test scores looks far more important than it did in the 1960s.
|More-Equal Scores Now Bring More-Equal Earnings
Ratio of Black to White Annual Earnings in 1964 and 1993 for Employed Men in their Early Thirties, by Percentile Score on a Military Test Taken Between the Ages of 18 and 23
|Sources: Cutright and authors' tabulations from the National Longitudinal Survey of Youth (NLSY). Cutright's version of the AFQT included vocabulary, arithmetic, and spatial relations. Our NLSY approxiation of Cutright's AFQT included word knowledge, numerical operations, and mechanical reasoning. See our introduction to The Black-White Test Score Gap for details on the samples and standard errors.|
Some skeptics have argued that scores on tests of this kind are really just proxies for family background. Family background does affect test performance. If we compare random pairs of children, their IQ scores differ by an average of 17 points. Among pairs of children who have been adopted into the same family, the difference averages about 15 points. Even if we compare pairs of biological siblings reared in the same family, their IQ scores still differ by an average of 12 or 13 points. The claim that test scores are only a proxy for family background is therefore false. Furthermore, test score differences between siblings raised in the same family have sizable effects on their educational attainment and earnings. Thus while it is true that eliminating the black-white test score gap would not eliminate the black-white earnings gap, the effect would surely be substantial.
Reducing the black-white test score gap would reduce racial disparities in educational attainment as well as earnings. The nationwide "High School and Beyond" survey tested twelfth graders in 1982 and followed them up in 1992, when they were in their late twenties. At the time of the follow-up only 13.3 percent of the blacks had earned a bachelor's degree, compared to 30 percent of the non-Hispanic whites. Many observers blame this disparity on black parents' inability to pay college bills, black students' lack of motivation, or the hostility that black students encounter on predominantly white college campuses. All these factors probably play some role. Nonetheless, when we compare blacks and whites with the same twelfth-grade test scores, blacks are more likely than whites to complete college. Once we equalize test scores, High School and Beyond blacks' 16.7-point disadvantage in college graduation rates turns into a 5.9-point advantage.
Eliminating racial differences in test performance would also allow colleges, professional schools, and employers to phase out the racial preferences that have caused so much political trouble over the past generation. If selective colleges based their admission decisions solely on applicants' predicted college grades, their undergraduate enrollment would currently be 96 or 97 percent white and Asian. To avoid this, almost all selective colleges and professional schools admit African Americans and Hispanics whom they would not admit if they were white. If selective colleges could achieve racial diversity without making race an explicit factor in their admission decisions, blacks would do better in college and whites would nurse fewer political grudges.
Advocates of racial equality might be more willing to accept our argument that narrowing the test score gap is crucial to achieving their goals if they believed that narrowing the gap was really feasible. But pessimism on this front has become almost universal. In the 1960s, racial egalitarians routinely blamed the test score gap on the combined effects of black poverty, racial segregation, and inadequate funding for black schools. That analysis implied obvious solutions: raise black children's family income, desegregate their schools, and equalize spending on schools that remain racially segregated. All these steps still look useful, but none has made as much difference as optimists expected in the early 1960s.
- The number of affluent black parents has grown substantially since the 1960s, but their children's test scores still lag far behind those of white children from equally affluent families. Income inequality between blacks and whites appears to play some role in the test score gap, but it is quite small.
- Most southern schools desegregated in the early 1970s, and southern black nine-year-olds' reading scores seem to have risen as a result. Even today, black third graders in predominantly white schools read better than initially similar blacks who have attended predominantly black schools. But large racial differences in reading skills persist even in desegregated schools, and a school's racial mix has little effect on reading scores after sixth grade or on math scores at any age.
- Despite glaring economic inequalities between a few rich suburbs and nearby central cities, the average black child and the average white child now live in school districts that spend almost exactly the same amount per pupil. Black and white schools also have the same average number of teachers per pupil, the same pay scales, and teachers with almost the same amount of formal education and teaching experience. The most important resource difference between black and white schools seems to be that both black and white teachers in black schools have lower test scores than their counterparts in white schools.
For all these reasons, the number of people who think they know how to eliminate racial differences in test performance has shrunk steadily since the mid-1960s. While many people still think the traditional liberal remedies would help, few now believe they would suffice.
Demoralization among liberals has given new legitimacy to conservative explanations for the test score gap. From an empirical viewpoint, however, the traditional conservative explanations are no more appealing than their liberal counterparts. These explanations fall into three overlapping categories: the culture of poverty, the scarcity of two-parent black families, and genes.
- In the 1960s and 1970s many conservatives blamed blacks' problems on a culture of poverty that rejected school achievement, the work ethic, and the two-parent family in favor of instant gratification and episodic violence. In the 1980s conservatives (as well as some liberals) characterized the "black underclass" in similar terms. But this description fits only a tiny fraction of the black population. It certainly cannot explain why children from affluent black families have much lower test scores than their white counterparts.
- Conservatives invoke the decline of the family to explain social problems almost as frequently as liberals invoke poverty. But once we control for a mother's family background, test scores, and years of schooling, whether she is married has even less effect on her children's test scores than whether she is poor.
- Scientists have not yet identified many of the genes that affect test performance, so we have no genetic evidence regarding innate cognitive differences between blacks and whites. But we have accumulated a fair amount of indirect evidence since 1970. Most of it suggests that whether children live in a "black" or "white" environment has far more impact on their test performance than the number of Africans or Europeans in their family tree [see "The Heredity-Environment Controversy"].
CULTURE AND SCHOOLING
Taken as a whole, then, what we have characterized as the "traditional" explanations for the black-white test score gap do not take us very far. This has led some people to dismiss the gap as unimportant, arguing that the tests are culturally biased and do not measure skills that matter in the real world. Few scholars who spend time looking at quantitative data like that in figure 2 accept either of these arguments, so they have had to look for new explanations of the gap. These new explanations can mostly be grouped under two overlapping headings: schooling and culture.
Social scientists' thinking about "school effects" has changed substantially since the late 1960s. The 1966 Coleman Report and subsequent studies convinced most economists and quantitative sociologists (including Jencks) that school resources had little impact on achievement. Since 1990, however, new statistical methods, new data, and a handful of genuine experiments have suggested that additional resources may in fact have sizable effects on student achievement. The notion that resources matter cannot in itself explain the black-white achievement gap, because most school resources are now fairly equally distributed between blacks and whites. But certain crucial resources, like teachers with high test scores, are still quite unequally distributed. And other resources, like small classes and teachers with high expectations, may help blacks more than whites.
Equally important is the fact that predominantly black schools enroll far more children with severe academic and behavioral problems than white schools do. Such children consume many times more resources than the average child. To begin with, they are often assigned to very small classes (for the "educably mentally retarded," for example). In addition, schools where many children have serious academic, emotional, or disciplinary problems need more reading specialists, more psychologists, and more security guards. That leaves less money for regular teachers. Finally, children with serious problems consume a disproportionate share of their teachers' time when they are in regular classes, leaving less time for other students. The net result is that while predominantly black schools spend about as much per pupil as predominantly white schools, ordinary black children without special problems are likely to be in larger classes, get less attention, and have less academically skilled teachers than similar white children.
Nonetheless, disparities between black and white schools cannot explain why black children enter preschool with smaller vocabularies than white children. This fact must reflect differences between black and white children's experiences before they enter school. While racial disparities in income, parental education, family size, and the like explain some of the test score gap among preschool children, they do not explain most of it. That fact has forced many scholars to take cultural explanations more seriously.
In the late 1960s and early 1970s, many liberals and radicals dismissed cultural explanations of the test score gap as an effort to put down blacks for not thinking and acting like upper-middle-class whites. Since then, cultural explanations have enjoyed a slow but steady revival. In 1978, the Nigerian anthropologist John Ogbu suggested that caste-like minorities throughout the world tended to do poorly in school, even when they were visually indistinguishable from the majority. Later, Ogbu made this argument more specific, suggesting that because of their caste-like status, blacks developed an "oppositional" culture that equated academic success with "acting white." By linking black culture directly to oppression, Ogbu made it much easier for liberals to talk about cultural differences. Jeff Howard and Ray Hammond added another important strand to this argument when they suggested that academic competence develops partly through competition, and that "rumors of inferiority" make blacks reluctant to compete in the academic arena. More recently, Claude Steele has argued that people of all races avoid situations in which they expect others to have negative stereotypes about them, even when they know that the stereotype does not apply. According to Steele, many black students "disidentify" with school because constructing a personal identity based on academic competence entails a commitment to dealing with such stereotypes on a daily basis. In a series of elegant experiments with Stanford students, Steele has also shown that merely asking test-takers to report their race or telling them that a test measures intellectual ability lowers black students' scores.
Cutright's 1964 finding that blacks with high test scores earned little more than those with low test scores may also help explain why blacks did so badly on these tests. In an economy where high-scoring blacks had very limited job opportunities, few blacks had any reason to suppose that they would win anyone's respect by acquiring a larger vocabulary, better mathematical skills, or more information about science, history, and literature. As we have seen ["More-Equal Scores Now Bring More-Equal Earnings"], the world has changed since 1964. But it always takes several generations for any group to adjust to a new reality, especially when the adjustment has significant costs (spending more time studying). The message that nerds will do well as adults is always hard to sell to children, but it is doubly hard to sell when it has only recently become true.
CAN WE EXPLAIN MORE OF THE GAP?
The available evidence shows that traditional explanations for the black-white test score gap do not work very well. If genes play any role, it is probably quite small. Poverty probably plays some role, but it too is modest. School desegregation probably reduced the black-white gap in the South during the 1970s, and desegregating northern elementary schools might raise blacks' reading scores today, but the gain would not be huge. Reducing class size in the early grades would probably raise scores more for blacks than whites, and reducing class size in later grades might help preserve these black gains, although this latter conclusion is based on conjecture rather than firm evidence. Screening teachers for verbal and mathematical competence (or raising the competence of those already employed) is also likely to raise black children's scores.
The United States should be conducting large-scale experiments aimed at reducing uncertainty about the effects of schools' racial mix, class size, teachers' test scores, ability grouping, and many other policies. We do such experiments to determine the effects of different medical treatments, different job training programs, and many other social interventions. But the U.S. Department of Education, which should in principle be funding experiments from which every state and school district would benefit, has shown almost no interest in this approach to advancing knowledge. The most important piece of educational research in the past generation, the Tennessee class size experiment, showed that small classes in the early grades made a big difference, especially for blacks -- yet it was funded by the Tennessee legislature, not the U.S. Department of Education. Experimental assessments of other educational policies that have a major impact on school spending -- salary levels, teacher selection systems, education for the physically and mentally disabled, and bilingual education, for example -- have been almost nonexistent.
If we did more experiments, we might eventually develop better theories. At present, theorizing about the causes of the black-white gap is largely a waste of time, because there is no way to resolve theoretical disagreements without data that both sides accept as valid. Most theories about human behavior start out as hunches, anecdotes, or ideological predispositions. Such theories improve only when they have to confront evidence that the theorist cannot control. In education, that seldom happens.
Our best guess is that successful new theories about the causes of the black-white gap will differ from traditional theories in at least three ways:
- Instead of looking mainly for resource differences between predominantly black and predominantly white schools, successful theories will concentrate on differences in the way black and white schools spend the resources available to them -- differences that probably derive in large part from the disproportionate number of black children with severe academic or psychological problems, but that probably prevent ordinary black children in the same schools from getting as much attention and support as their counterparts in white schools.
- Instead of concentrating on whether teachers treat black and white children differently, successful theories will probably pay more attention to the way black and white children respond to the same classroom experiences, such as having a teacher of a different race or having a teacher with low expectations for students who read below grade level.
- Instead of emphasizing families' economic and educational resources, successful theories will probably pay more attention to the way family members and friends interact with one another and with the outside world. A good explanation of why white four-year-olds have bigger vocabularies than black four-year-olds is likely to focus on how much parents talk to their children, how they deal with their children's questions, and how they react when their children either learn or fail to learn something, not on how much money the parents have in the bank.
Psychological and cultural differences are hard to describe accurately and therefore are easy to exaggerate. Collecting accurate data on such differences would require a massive investment of effort, perhaps comparable to what psychologists invested in developing cognitive tests during the first half of the twentieth century. It would also require far closer cooperation between psychologists, ethnographers, and survey researchers than one ordinarily sees in academic life.
CAN WE NARROW THE GAP?
We have argued that reducing the black-white test score gap would do more to move America toward racial equality than any politically plausible alternative. This argument rests on two problematic premises: first, that we know how to reduce the gap, and second, that the policies required to reduce the gap could command broad political support.
When readers try to assess what experts know, they should start by drawing a sharp distinction between policies that are expensive but easy to implement and policies that are cheap but hard to implement. (Most policies that are cheap, easy to implement, and clearly effective -- like teaching children the alphabet or the multiplication tables -- are already universal.) The two policies that seem to us most likely to combine effectiveness with ease of implementation are cutting class size and screening out semiliterate teachers. Cutting class size is clearly expensive. Selecting teachers with higher test scores would not be expensive in districts that now have more applicants than openings, but it would require higher salaries in many big cities.
When educators look for less expensive ways of raising black children's achievement, they usually find themselves considering proposals that are quite difficult to implement. Raising teachers' expectations is not inherently expensive, for example, but how does a school administrator do it? Big-city school districts are besieged by advocates of curricular innovation who claim their programs will raise black children's test scores. These programs usually require complex and relatively subtle changes in classroom practice. School boards and administrators cannot implement such changes by decree the way they can reduce class size or require new teachers to pass an exam. Nor can teachers make such changes by a single act of will, the way they might adopt a new textbook. As a result, schools seldom implement these programs in exactly the way their designers expected or intended. A program may work well initially, when it is closely supervised by a dedicated innovator, but may have no detectable effect when the innovator tries to bottle it and sell it "off the shelf."
Proposals for reducing the black-white test score gap also arouse passionate white resistance if they seem likely to lower white children's achievement. Both school desegregation and the elimination of academically selective classes often arouse such fears. School desegregation probably raises black elementary school students' reading scores. Whether it lowers white scores is unclear. But once black enrollment in a neighborhood school rises past something like 20 percent, white parents become reluctant to move into the neighborhood. If black enrollment remains low enough to keep whites comfortable, many blacks feel uncomfortable. There is no simple way out of this dilemma. It takes years of patient work to create stable, racially mixed schools with high black enrollments, and such schools remain unusual.
Experimental evidence, while limited, suggests that the struggle over ability grouping at the elementary level is largely symbolic: eliminating such classes would probably not do black children much good, and it would probably not do whites much harm either. At the secondary level, eliminating demanding courses seems ridiculous. We should be trying to get more black students to take such classes, not trying to eliminate them as an option for whites, who will respond by sending their children elsewhere. Any politically workable educational strategy for reducing the black-white test score gap has to promise some benefits for whites as well as blacks. Reducing class size, requiring higher levels of academic competence among teachers, and raising teachers' expectations for students who have trouble with schoolwork all meet this test.
Although we believe that improving the nation's schools could reduce the black-white test score gap, schools alone cannot eliminate the entire gap. To begin with, competition for educational resources is fierce. The typical American voter might accept a system of educational finance that gave schools with unusually disadvantaged students 10 or 20 percent more money per pupil than the average school. But few Americans would accept a system that gave disadvantaged schools 50 or 100 percent more money than the average school. If smaller classes in disadvantaged schools raise children's reading skills, for example, more affluent parents will want smaller classes too. In a system of school finance that relies heavily on local funding, affluent parents who want smaller classes will usually be able to get them. Even ensuring equal funding for black and white schools is a constant struggle. Creating a system in which black schools get far more money than white schools is politically inconceivable.
Even if resources were not a constraint, the cognitive disparities between black and white preschool children are currently so large that it is hard to imagine how schools alone could completely eliminate them. When three and four-year-olds take vocabulary tests, for example, the typical black child's vocabulary score falls below the twentieth percentile of the national distribution. Relying entirely on educational reform to move such a child up to the fiftieth percentile does not strike us as realistic. If we want equal outcomes among twelfth graders, we will also have to narrow the skill gap between black and white children before they enter school.
Broadly speaking, there are two ways of raising three and four-year-olds' cognitive skills: we can change their preschool experiences and we can change their home experiences. Changing preschools is less important but easier than changing homes. Black preschoolers are concentrated in Head Start. At present, Head Start does not emphasize teaching cognitive skills. Few Head Start teachers are trained to do this, and some oppose the idea on principle. A review of research on preschool effects by Steven Barnett, a professor of education at Rutgers, strongly suggests that cognitively oriented preschool programs can improve black children's achievement scores, even though the benefits fade somewhat as children age. Getting Head Start to emphasize cognitive development should therefore be a higher priority than merely expanding its enrollment.
Parenting practices almost certainly have more impact on children's cognitive development than preschool practices. Indeed, changing the way parents deal with their children may be the single most important thing we can do to improve children's cognitive skills. But getting parents to change their habits is even harder than getting teachers to change. Like teachers, parents are usually suspicious of unsolicited suggestions. This is doubly true when the suggestions require fundamental changes in a parent's behavior. But once parents become convinced that changing their behavior will really help their children, many try to change, even when this is quite difficult. As a practical matter, whites cannot tell black parents to change their practices without provoking charges of ethnocentrism, racism, and much else. But black parents are not the only ones who need help. We should be promoting better parenting practices for all parents, using every tool at our disposal, from Head Start outreach programs and home visits by nurses to television soap operas, sitcoms, or anything else that looks promising.
A successful strategy for raising black children's test scores will also include convincing both blacks and whites that the gap is not genetic in origin. This is not a simple task. Genetic variation does explain a substantial fraction of the variation in cognitive skills among people of the same race. So does environmental variation. But once hereditarianism percolates into popular culture, it can easily become an excuse for treating academic failure as an inescapable fact of nature. Teaching children skills that do not seem to "come naturally" is hard work. If our culture allows us to avoid such work by saying that a child simply lacks the required aptitude to master the skill, both teachers and parents will sometimes jump at this as an excuse for not trying. While this often makes everyone's life more pleasant in the short run, in the long run it is a formula for failure. Emphasizing heredity is likely to have especially negative consequences for African-American children, who start off behind white children and therefore need to work even harder than white children if they are to catch up.
The agenda we have sketched would not be easy to implement. We are not optimistic about expanding federal support for efforts of this kind. Popular distrust of federal education programs is now quite pervasive and shows no sign of receding. We are more optimistic about state and local efforts to narrow the black-white test score gap. Everyone recognizes that racial conflict is one of the nation's most pressing and persistent problems. Other strategies for dealing with this problem, which emphasize the use of racial preferences to overcome the adverse effects of discrimination, the test score gap, or both, are clearly in political trouble. Public support for efforts to narrow the test score gap, while tempered by suspicion that "nothing works," still seems fairly widespread. One reason is that the beneficiaries appear so deserving. Hardly anyone blames black first graders' limited vocabulary on defects of character or lack of ambition. First graders of every race seem eager to please. That was why Lyndon Johnson placed so much emphasis on helping children in his original War on Poverty.
We recognize that few readers will find our sketchy agenda for reducing the black-white test score gap entirely persuasive. Such skepticism is completely reasonable. While we are convinced that reducing the black-white test score gap is both necessary and possible, we do not have a detailed blueprint for achieving this goal -- and neither does anyone else. The available research raises as many questions as it answers. This is partly because psychologists, sociologists, and educational researchers have devoted far less attention to the test score gap over the past quarter century than its political and social consequences warranted. Most social scientists have chosen safer topics and hoped the problem would go away. It didn't. We can do better.