Single-sex schooling no longer offers middle-class girls any academic advantages, new American research suggests.
During the 1970s and 1980s, several US studies found that most girls benefited if they were taken out of male-dominated classrooms. But today only girls - and boys - from disadvantaged families show "significant gains" from single-sex education, according to Cornelius Riordan, who has been conducting research in this field for almost 20 years.
And even for poor children, the type of school attended is much less important than either home background or the school's curriculum, he says.
Dr Riordan, of Providence College, Rhode Island, believes that, up until the 1980s, American girls benefited from single-sex schooling because they were "historically and traditionally disadvantaged in school". However, research indicates that girls and boys were competing on equal terms by 1990.
Recent US studies of elite independent schools and the Catholic education sector have confirmed that girls now do equally well in single-sex and co-educational schools.
Riordan does not, however, believe that such findings should be used to attack single-sex education. In fact he complains that "political correctness" makes it difficult to challenge the predominance of co-educational schooling in the public sector.
Midle-class students may still make small undetectable gains from single-sex schooling, he says in a soon-to-be-published book, Doing Gender in Policy and Practice. He also points to the positive impact that single-sex education could have on America's troubled public schools.
"The challenge ... in the next century is to overcome the negative effects of youth culture," he says. "This is not a new problem and undoubtedly pre-dates the modern school. But the intensity and complexity of the problem is new and it is the most important obstacle in schools today.
"Single-gender schools are places where students go to learn, not to play, not to hassle teachers and other students, and not primarily to meet their friends and have fun."
Riordan does not say whether his findings could also be applicable to Britain. But he adds that a 1995 study - involving researchers in Belgium, New Zealand, Thailand and Japan - revealed that the impact of single-sex schools varies from one country to another.
"The effects appear to be limited to those national education systems in which single-sex schools are relatively rare," he says. "When they are rare there is a more selective student body who will bring with them a heightened degree of academic demands."
"Doing gender in policy and practice: perspectives on single-sex and co-educational schooling", edited by Amanda Datnow and Lea Hubbard, is to be published by Routledge Falmer.