Closing the Black-White Test Score Gap

General Testing

How important is the test-score gap between racial groups in the U.S.? Can it be closed? These recurring questions have surfaced again in the media. One danger in this discussion is that in the effort to find ways to close the gap, those who score lower on tests will be subjected to more intensive teaching to the test. Yet such an approach is part of the problem, not part of the solution.


The test-score gap between African Americans, Latinos and Native Americans on one side, and European Americans and some Asian American groups on the other, has only slowly diminished over the past several decades. In fact, scores on the National Assessment of Educational Progress (NAEP) shows a widening gap in some cases in the 1990s. The average African American scores below 75 percent of U.S. whites on most standardized tests.


This data has led to numerous interpretations, from the racist arguments of genetic inferiority recently re-launched by The Bell Curve, to claims of cultural deprivation, to interpretations that have pointed to economic inequality and differential school quality. Similarly, solutions to the gap also varied widely, with many focusing on school remediation efforts, such as the federal Title I program. Others argued that changing schools would have little effect on many test scores, particularly IQ tests, or on changing economic outcomes. Among the latter were James Coleman in his well-known report and Christopher Jencks in his book Inequality, which suggested that if social equality was the goal, focusing on schools was not a good way to reach it.


Now a new study by Jencks and Meredith Phillips argues that closing the test score gap will make a difference for economic outcomes. They show that for blacks and whites with the same grade 12 test scores, blacks are more likely than whites to complete college. And by 1993, African American men who scored above the national average on the Armed Forces Qualification Test earned only four percent less than did whites with similar scores. Thus, while the test score gap could not explain differential earnings or achievement in the 1960s, now it can.


In a series of papers they edited for the Brookings Institute and in a detailed article in The American Prospect (Sept.-Oct. 1998), the authors also maintain that the test score gap can be closed. They point to the malleability of IQ test scores and the narrowing of achievement test results in the twentieth century. On NAEP, the reading score gap closed by almost one half from 1971 to 1996 (though progress has stalled since 1989).


But theoretically being able to close the gap and actually closing it are two different things. Jencks and Phillips maintain the results probably will come by focusing on schools and on culture. They argue that the effects of previous and continuing racism lead many young blacks to believe education will not benefit them, a point of view that will need to change to reflect the new reality that education, as reflected in test scores, matters for blacks as well as for whites.


They also note studies by researchers such as Stanford psychologist Claude Steele, who found that asking African Americans to report their race leads to lower scores (see Examiner, Fall-Winter 1995-96). Changes in what happens inside families, including such things as how parents talk and read with their children, probably will matter a lot, report the authors, as black children on average come to school with smaller vocabularies than white children (as do low-income children compared with children from more affluent families).


The authors are not sure how best to proceed to implement the programmatic changes that would lead to closing the score gap. They suspect there will be a lack of support for large federal programs in this area. However, efforts such as cognitively enriched schooling in the early grades, smaller class sizes, and better teachers will make a difference and, they think, might be politically feasible since they benefit all children.


Is it the Test Scores?

In the Nov. - Dec. American Prospect, Claude Steele; Stephen Ceci and Wendy Williams; Mindy Kornhaber; Jared Bernstein and Richard Rothstein; and Glenn Loury discuss Jencks and Phillips' analysis and conclusions, who then respond. Bernstein and Rothstein argue from labor market research that, in fact, test scores do not explain as much of the income differences as Jencks and Phillips claim. Rather, "the long-term erosion of labor-market institutions that formerly lifted wages of non-college-educated workers, and labor market discrimination" explain better and provide a better route to a solution of the problem of income inequality.


Most of the critics maintain that the point is not the test scores but the real abilities which may or may not be reflected in the test scores (a point with which Jencks and Phillips agree). Both Steele and Kornhaber emphasize that tests such as the SAT are limited predictors: the SAT correctly forecasts less than 20 percent of the variance in first year college grades and far less of the variance in graduation rates or later income.


However, Jencks and Phillips claim that Phillips' research for the Brookings papers shows that SAT differences fully explain the gap in college graduation rates and most of the gap in subsequent earnings. This is because on average, black and white college students as groups are quite alike on the non-test factors that explain most of the variance in college grades, making the difference in test scores more significant for explaining the gaps between the groups.


Kornhaber points to some dangerous educational consequences of the emphasis on test scores: "(P)olicies aimed at boosting test scores do not always overlap with policies for building minds... (N)othing Jencks and Phillips suggest directly bears on two crucial footholds: enabling students to engage the material at length and to reflect upon their efforts... (I)f the goal is high-level cognitive equity, or 'keen minds,' then equipping educators with classroom techniques, organizational structures, and social policies to keep diverse students engaged over the long haul may matter even more."


Closing the Gap

"The Canary in the Mine: The Achievement Gap Between Black and White Students," in the September Phi Delta Kappan offers powerful ways to address the problem. Similar to Jencks and Phillips, Mano Singham uses research by Claude Steele, John Ogbu and Signithia Fordham to point out that the cultural divide between schools and many African American youths hinders those youths' willingness to engage in school. Singham then uses research by Uri Triesman to show that students who study together in workshops did much better in their college courses. (This approach also has been found to work well in preparing nurses for the nursing board exams and in their course work.) The benefits appear for all students, but especially for groups who historically have not done well.


Singham concludes that the underlying problem is less how schools educate black students than "in the way we teach all students." He points to the low-level, fact-heavy, boring curriculum which turns many students off to learning, particularly those students with less immediate reason to become invested in schools. (Ogbu=s distinction between "voluntary" and "involuntary" immigrants is important here.)


This is, as Kornhaber notes, the very sort of curriculum which typically results from an emphasis on raising scores on standardized tests. That is, the effort to "improve" schooling by focusing on test scores neither enables students to develop the cognitive understandings and skills to do more advanced work nor gets them engaged in their own learning.


Singham concludes that what is needed is "active learning," which has been shown to produce significant academic gains for all students and reduce the achievement gap. Such active learning is precisely what makes the workshops helpful for many students. He concludes that the current system of largely passive schooling has worked for perhaps one quarter of the population. If the U.S. wants substantially more students to learn the academics schools are supposed to teach, instruction will have to change.


A clear implication of Singham's article is that even if Jencks and Phillips are correct that increasing test scores will close the gaps in higher education, employment and earnings, progress cannot be attained by teaching to the test. This is an argument test reformers, including FairTest, have made for many years. It would be counterproductive if Jencks and Meredith's focus on test scores once again served to reinforce the kinds of schooling which have helped to produce and maintain the very gap they condemn.


-- The Black-White Test Score Gap is available from Brookings, $18.95 (pap) + $4 postage; (202) 797-6258.