Sociologist James S. Coleman, whose massive study in the mid-1960s has been widely used to support school desegregation, now says it is a "mistaken belief" that black students learn better integrated classrooms.

During the past decade, Coleman said, research throughout the country has shown that "it is not the case that school desegregation, as it has been carried out in American schools, generally brings achievement benefits to disadvantaged (black) children."

In some situations, Coleman said, desegregation has brought slight gains in black achievement, but in many others there has been no change or a slight loss.

Even though he had argued a decade ago that "integration would bring about achievement benefits," Coleman said, "It has not worked out this way in many of the school desegregation cases since that research. . . .

"Thus, what once appeared to be fact is now known to be fiction," Cole man said.

Coleman, a professor of sociology of the University of Chicago, presented his new conclusions in a paper in April. He repeated them in an interview this weekend.

"Desegregation has turned out to be much more complicated than any of us ever realized," Coleman said. "There appear to be beneficial effects for some black kids, those who are better students, and harmful effects for blacks who are poorer students. It all seems to balance out, which is quite the reverse of the implications of my own research" in the mid-1960s.

Coleman's 1966 report, called "Equality of Educational Opportunity," was authorized by Congress in the 1964 Civil Rights Act and sponsored by the U.S. Office of Education. It still is the most extensive piece of educational research ever conducted, involving tests and surveys of about 600,000 students and 60,000 teachers in 4,000 schools around the country.

Its most widely noted conclusions were that the social class composition of a school had more impact on student achievement than either resources or teaching methods, and that lower-class black children scored somewhat higher on standardized tests in schools with a middle-class white majority than they did in schools where all the children were poor and black.

Coleman stressed that the achievement gain occurred not because of skin color but because of the middle-class background and "educational resources" that many white children bring from home.

After his report, Coleman expressed his views widely, not only in scholarly articles but also in testimony before congressional committee and in school desegregation cases in courts.

Among these was Julius Hobson's suit against the Washington school system. Coleman testified in Hobson's behalf and was cited by U.S. Judge J. Skelly Wright to support the court's finding that "Negro students' educational achievement improves when they transfer into white or integrated educational institutions."

Coleman said he now believes that this view is "incorrect . . . wishful thinking."

In the interview, Coleman said the difference between his conclusions a decade ago and the results of desegregation since then reflects two main factors - a difference in the way desegregation has been carried out, and the availability of new research.

When he collected his data in 1965, Coleman said, nearly all the black children attending integrated school in the South were well-motivated volunteers under "open enrollment" plans. In the North almost all integration had occured in neighborhood schools where blacks and whites lived nearby.

Since then, Coleman noted, many school districts have been desegregated through mandatory busing programs, ordered by courts or state agencies, that bring children together from wide areas.

"Much of it has been accompanied by the kinds of things that don't foster achievement," Coleman said. "Often there's been some degree of turmoil and lower standards, with white teachers being afraid to apply the same standards to black students and therefore not teaching them as well."

Coleman said his 1966 report was based on data collected at one time, with conclusions drawn by comparing youngsters in schools with different proportions of black and white students.

Since then, he said, researchers have been able to follow children for several years after they switched to desegregated schools. Although Coleman has not been directly involved in any of this research, he said a review of over 100 desegregation studies in cities around the country - from Boston to Berkeley - shows "no overall gains."

"Some of the most carefully studied cases, such as in Pasadena and Riverside, Calif., "Coleman said, "show either no achievement effects or else losses."

In the South some gains by blacks have been reported by the National Assessment of Educational Progress, but Coleman said these occured in both segregated and integrated classrooms. He said these gains probably are the result of "the broader impact of desegregation in the South . . . it drew a lot of attention to schools that used to be the worst in the nation," rather than a direct result of blacks and whites being taught together.

Before going to the University of Chicago in 1973, Coleman taught at Johns Hopkins University in Baltimore. He took part in civil rights demonstrations there, and was arrested in one of them.

He said he still strongly opposes legal segregation and strongly favors integrated schools. But he said mandatory busing in many cities has been "counter-productive" because it has been followed by an extensive loss of white students.

Coleman also rejects "the belief that an all-black school is inherently bad."

"That has a curiously racist flavor," Coleman said, "which I can't accept There have been, and there are, all-black schools that are excellent schools by any standard."

"What is essential," he said, "is that if a child is in an all-black school, it should be because . . . his parents want him to be there, not because it is the only school he has a reasonable chance to attend."

Coleman said he thinks the best ways of increasing school integration now would be encourage voluntary transfer between city and suburban schools or to offer vouchers allowing parents to pcik any school for their children but providing more funds for integrated schools.

"We ought to take measures so we can become a more integrated society," Coleman said, "but we ought to be clear that integrated education does not depend on maintaining romantic notions that are not true."