Single-sex classes just led to deterioration in boys' behaviour.
The headteacher who came up with the idea of a trial experiment in a Welsh school to see if single-sex classes would boost the overall performance of boys said this week it had failed to deliver better results. And he claimed it had led to worse behaviour.
Frank Ciccotti, from Pembroke school in Pembrokeshire, said the results could prove a lesson to other schools thinking of using the same technique.
The school split up Year 8 boys and girls three years ago when the so-called gender gap was starting to reach worrying levels in Wales. Staff believed it might be the answer to improved results after an all-girl GCSE physics group had proved successful, improving A-level uptake and boosting results.
But when the Y8 boys were put together their behaviour soon deteriorated, particularly in lower-ability groups. It didn't lead to a marked dip in performance but neither did it improve things.
"There were slight variations in performance but nothing major. It was surprising I'd expected a fairly significant shift. But one of the most marked differences was in the predictably poor behaviour of boys," said Mr Ciccoti.
The head's experiment began in 2004, involving 120 pupils studying English, history, maths and science. Meanwhile, a control group the same size remained in mixed classes.
He was hoping to improve the performance of both sexes. Unusually, at the time boys in south Pembrokeshire were doing better, bucking the national trend of girls outperforming boys.
He has now abandoned the experiment.
"We had records of evaluation based on results from key stage 2 and Midyis (a test-based middle years information system) which gave a score that predicted future performance," said Mr Ciccotti. "It didn't show much sign of improvement."
He now believes the presence of an enthusiastic classroom teacher can make all the difference to achievement levels, regardless of whether the sexes are divided. Girls are now outperforming boys in south Pembrokeshire as elsewhere in Wales.
"I think it's a literacy issue with boys, particularly in KS3 English where there's a significant margin between them and the girls," said Mr Ciccotti.
The girl-boy gap is now so marked that education minister Jane Hutt launched an inquiry this summer. The Assembly government has asked the schools inspectorate Estyn to study those with the smallest differences to see how they achieve this, collecting evidence of good practice.
This year's GCSE results showed that while 66.5 per cent of girls in Wales obtained five or more A*-C grades, the figure was just 59.3 per cent for boys.