Education statistics may indicate that the gap in attainment levels between African-Americans and whites has narrowed over the past 30 years. But the proportion of black children in the top 10 per cent of the academic-ability range has actually declined since the mid-1960s.
Tests of reading and maths ability reveal that a white child in the 1990s is between 10 and 15 times more likely to be a high-achiever. And even after allowance has been made for social class and several school-effects factors the ratio is still 5:1 in favour of whites.
These figures were presented to the American Educational Research Association Conference by Amy Thoreson of the University of Chicago who said that the traditional focus on the mean attainment scores of different racial groups provided an incomplete picture.
Her analysis had also revealed that there was a disproportionate number of African-Americans in the bottom 10 per cent of the ability range.
The low-achievers' figures were less bleak than they had been in 1965 when African-Americans had outnumbered whites by 10.5:1. But the data suggested that African-Americans were still twice as likely to be in the bottom 10 per cent, even after allowing for social class and school effects.
"I hope these findings get you as angry and upset as they get me," Thoreson told the conference. "We know that these disparities will only exacerbate the black-white differences in earnings.
"Black and white students are supposed to be offered equal educational opportunities but our findings suggest that they are not receiving the same quality of education."
Thoreson pointed out that the urban schools attended by most ethnic-minority children were still receiving less funding than predominantly white suburban schools. "African-American children are also being channelled into a 'general track' curriculum, a programme of studies that is less rigorous than that offered to whites."
Thoreson based her findings on the results of six large surveys of school-leavers' maths and reading ability that were carried out between 1965 and 1992. The surveys offered information on the performance of almost 200,000 17 and 18-year-olds.
They also provided her with information on family income, and parents' education. The school-effects factors that she included in her calculations were: number of children eligible for free school meals, the drop-out rate, average daily attendance and programme of studies (stream). She also collected data on the category of school the pupils had attended as achievement levels are generally higher in Roman Catholic schools.
"Although the rate of change for the top 50 per cent indicates some movement towards equality in general achievement, changes in the top 5 per cent, 10 per cent and 25 per cent areas of the distribution are trivial," she said.
Thoreson added that at the present rate of change, the black-white gap in general achievement would not close for 60 years. Furthermore, it would be another 30 years before black and white students with the same socioeconomic background and attending similar schools were achieving similar results.
"In his 1969 study Coleman concluded that unequal educational opportunities are all but inevitable due to the inequalities 'arising from outside and not overcome by the school system'," Thoreson said. "Other studies suggest educational attainment is determined more by what happens to black children inside schools than what they bring to school. The results from our study lend credence to both of these propositions."
The black-white gap may, however, be even wider than even Thoreson's study suggests, judging by a recent report from Education Trust, a Washington-based group that promotes equity in academic achievement.
It suggests that the gap in mean achievement scores narrowed during the 1970s and most of the 1980s, but has been opening up during the 1990s. In 1973, African-Americans scored 40 points below whites in the 12th grade (Years 1213) National Assessment of Educational Progress maths exam. By 1990 that gap had reduced to 21 points, but by 1994 it had widened to 27 points.