Achievement Gap Between Blacks, Whites Continues to Narrow
Anaheim, Calif--The narrowing achievement gap between black and white students, first reported about two years ago, is also beginning to show in college and graduate-school admissions tests, according to a new analysis of national data by the researcher whose earlier analyses first summarized the change.
The study, conducted by Lyle V. Jones, professor of psychology at the University of North Carolina at Chapel Hill, also examined data from the latest National Assessment of Educational Progress (naep). Those scores provide more evidence that although the gap between the performance of black and white students is still sizable, the difference continues to shrink. The study's findings were presented here last week at the annual convention of the American Psychological Association.
The report on students' scores on achievement tests came one day after Arthur Jensen of the University of California at Berkeley presented the latest results of his ongoing research on racial differences on intelligence-test scores.
In his paper, Mr. Jensen examined the relationship between the "g," or general-intelligence factor, and race. The notion of "g" was developed by Charles Spearman, a psychologist who worked in the early 1900's.
According to Mr. Jensen's theory, a strong "g" ability is a partial--but not complete--explanation of why some people are better at some tasks than are other people. He argues that on more complicated tasks--i.e. those that are "g-saturated"--the difference in scores between blacks and whites is greater. According to his hypoth-esis, one can predict the test scores that racial groups will achieve based on the degree to which a given test is "g-saturated."
Mr. Jensen, whose earlier work on the question of whether "intelligence" is an inherited or developed phenomenon was criticized by some as racist, remains controversial among psychologists and the general public alike. (Extra security guards were posted around the auditorium in which he spoke, and picketers bearing anti-Jensen signs clustered outside the hotel.)
In the paper presented last week, however, Mr. Jensen examined the nature of the racial difference in intelligence-test scores, not the explanation for its existence. He acknowledged that such an explanation is not likely to be uncovered without genetic experimentation, now deemed ethically unacceptable, on human subjects.
The latest evidence that black students are continuing to catch up with their white counterparts, Mr. Jones pointed out, comes some 15 years after the major school-desegregation initiatives, and 18 years after the passage of the Elementary and Secondary Education Act of 1965.
Thus, the students who are now taking college-admissions tests are among the first to have received their entire education in integrated schools and to have participated in federal programs such as Head Start.
That link, although not an explanation for the shift, is "consistent with the trends we've been seeing, so that desegregation in the school setting could contribute to the narrowing gap," Mr. Jones said in an interview.
But, he added, "so many things have changed. We can't point to the schools and say they did it." Also contributing, he suggested, are factors such as the rising levels of income among black families and the possibility that black students' motivation has improved because they now can realistically plan careers in fields once closed to them.
"While there's no good evidence," Mr. Jones said, "I am reasonably confident that those changes in society are a pretty powerful force.''
Looking at scores from both the Scholastic Aptitude Test (sat) and the Graduate Record Examination (gre), the researchers found that the average differences in performance for white and black students "consistently decline." For the sat, the study used 1976 to 1982 scores. For the gre, racial breakdowns of test scores were available only for 1979 to 1981.
"Despite possible shifts over time in the make-up of the samples of students who choose to take the sat and the gre," Mr. Jones stated in his paper, "the relative improvement of average scores for black students is apparent."
In an analysis of the factors that contribute to the differential levels of achievement for racial groups, the North Carolina researchers found that the number of mathematics courses taken was the strongest predictive variable for students' performance on achievement tests.
Information gathered for the naep indicates that, on average, black students take far fewer advanced mathematics courses than do white students. Thirty-seven percent of black students took only one math course, compared with 24 percent of white students. And 32 percent of white students took three math courses, compared with only 13 percent of black students.
Comparisons showed that students who took three math courses scored, on average, 82-percent correct, according to the analysis. Those who took only one course scored an average of 59-percent correct.
Moreover, Mr. Jones noted, there are pronounced differences in the average number of mathematics courses taken by students who attend predominantly white high schools, compared with those who attend predominantly black high schools.
The finding suggests--as has been recommended by numerous national reports--that one way of increasing students' achievement-test scores in mathematics is to require them to take more courses. Such increased requirements, Mr. Jones notes, would not only reduce the overall difference between white and black students' scores, but it would also shrink the achievement differential between predominantly black and predominantly white schools.
Narrowing Achievement Gap
Still an open question, however, is whether the narrowing achievement gap would be reflected on intelligence tests as well. As Mr. Jones and others pointed out, there is no national sample of intelligence-test scores analogous to the naep data for achievement.
The psychologists speculate, however, that if there were such a sample, it would show similar trends. "Any other outcome would be inconsistent with the data sets I've been tracking," Mr. Jones said. "I've seen striking improvement."
Mr. Jones suggested that the data Mr. Jensen analyzed--11 samples of intelligence-test scores for adults and children--are open to other interpretations.
"His hypothesis is tenable, but I think there are other hypotheses, too, from the same data, that haven't been disproven," Mr. Jones said.
An alternate explanation, he suggested, might be that of motivation. A task that does not require any prior knowledge still requires motivation before a person completes it successfully.
Hence, people who performs less well on a given task might react to its complexity by being unwilling to do it, while remaining perfectly capable of doing it if they chose.
Of the 11 tests analyzed by Mr. Jensen, one did not conform to his hypothesis.
That test, the Kaufman Assessment Battery for Children (kabc) was developed by Alan and Nadine Kaufman of the California School of Professional Psychology. Published last spring, the kabc uses a significantly different model of intelligence.
The kabc is designed to measure mental-processing skills--many of which are "g-saturated"--and achievement separately.
The mental-processing tests include measures of both sequential and simultaneous processing, both of which the Kaufmans regard as equally important components of intelligence. In an interview, Mr. Kaufman said that he did not see differing levels of motivation as an adequate explanation of the difference between whites' and blacks' scores on intelligence tests. Many intelligence tests that have high "g" also include a significant achievement component, he noted.
"One very possible explanation is that what he calls general intelligence may be a general achievement factor. That's one distinction we've made in kabc The kind of tests that are higher in g are often high in achievement. Even if that weak form of the hypothesis is true, if black-white differences are higher, that gets right back to the issue of opportunity for learning. It would still make no claim of the origin of the difference."
Also, Mr. Kaufman noted, in focusing so much on general intelligence, Mr. Jensen is not "taking into account anything developmental. If you look at his tables, he's grouping tests for young children with tests for adults. In that sense, he's ignoring developmental effects."
Mr. Kaufman cited research now being conducted by Cecil Reynolds, a psychologist at Texas A&M University. According to Mr. Reynolds's preliminary findings, there are racial differences in performance on the kabc
The results suggest that blacks are better at sequential reasoning, while whites are superior at simultaneous reasoning. But, Mr. Kaufman noted, these two abilities are considered equal according to the model by which the kabc was developed.
And studies that examine success on both of those scales have shown that both are equally related to success on achievement tests.
Mr. Kaufman and others, however, suggested that Mr. Jensen's research will be useful even if his hypothesis turns out to be wrong. Looking at patterns of difference, they said, is likely to be more productive than looking for global explanations of the racial differences.
Such explanations, they and Mr. Jensen point out, could be offered only after undertaking heredity experiments that would be "unthinkable'' in human populations.