The New York Times this week featured on its front page a story by Sam Dillon on the slow progress that the No Child Left Behind Act has made on closing the achievement gap between students of different racial and ethnic groups—a major goal of the law. But the result should not be surprising to serious students of education research, which has shown for 40 years that educating rich and poor students separately is a recipe for disaster—one that disproportionately affects African American and Latino children.

The famous Coleman Report of 1966 found that the socioeconomic status of a child’s family is the most powerful predictor of student academic achievement, and the socioeconomic status of the school a child attends is the second most powerful driver. Subsequent research affirms that students from poor families given a chance to attend middle-class schools do far better than students from poor families who attend high poverty schools. Almost all the ingredients of good schooling—supportive peers, a disciplined learning environment, active parents, good teachers with high expectations—are more likely to be found in middle-class schools.

No Child Left Behind, with its call for testing and accountability, makes sense as far as it goes. But for the most part, it takes as a given that students will be educated in economically and racially segregated schools. In theory, NCLB provides poor and minority kids a chance to escape failing high poverty schools and transfer to better performing public schools, but in many urban communities, that opportunity has proven illusory.

Research from Ronald Fryer Jr. of Harvard and Steven D. Levitt of the University of Chicago has found that the achievement gap between blacks and whites shows up in kindergarten, but that the entire gap can be explained by socioeconomic status. Once black students begin school, however, the black/white test score gap grows, even after controlling for socioeconomic status.

Why does this happen? One plausible explanation is that poor African Americans are much more likely to go to high poverty schools than poor whites given housing segregation by race. According to the Civil Rights Project at Harvard University, only 15% of mostly white schools were high poverty in the 2003-2004 school year, compared with 76% of schools with mostly minority populations.

As yesterday’s New York Times story noted, the racial achievement gap “narrowed steadily from the 1970s through the late 1980s.” It may be no coincidence that these were precisely the years when schools were being desegregated by race (and usually by socioeconomic status). In the early 1990s, resegregation of American schools began to set in. And, in the coming years, schools are expected to grow more segregated by economic status.

Closing the racial achievement gap is complicated business, with no single solution. But to begin with, we should focus on the fact that No Child Left Behind—like most of modern education reform—is about trying to make separate but equal work. As long as we are stuck in that paradigm, we’re likely to have more articles like the one in the Times for many years to come.

Richard D. Kahlenberg is a Senior Fellow at The Century Foundation.