A return to rigid streaming of pupils by ability is unlikely to satisfy the nation's educational needs in the next century, a new study of British and international research over the past 40 years suggests.
The study will contribute to the debate on streaming and setting which is raging in political circles.
"The current intense debate in the USA and Canada, where there is considerable pressure to discontinue tracking because of its undesirable social consequences, indicates that a return to a national system of selection and structured grouping is, in the long term, no more likely to succeed in the UK now than it did earlier this century," say Dr Susan Hallam and Ms Inje Toutounji of the University of London's Institute of Education.
Alternatives which emphasise meeting the needs of individual pupils will need to be developed to prepare young people "for the future, not for the past", they say.
New technology could play an important role in helping children work at their own pace, said Dr Hallam. "It would not be a good idea to go back to rigid streaming and we need systems that are flexible." Once children were placed in streams, it could be difficult for them to switch, even if schools insisted that streams were fluid.
She said bright children in top streams progressed faster in US studies, but lower-ability children got a raw deal and felt stigmatised. There were more children from lower socio-economic groups, and more ethnic-minority children and boys in lower streams. They tended to get less well-qualified teachers, who had lower expectations and gave them less challenging work. The more streams there were, the more likely were children to feel stigmatised.
Dr Hallam, a lecturer in the psychology of education, believes problems with mixed-ability grouping or setting by ability can be mitigated when they are understood by teachers. However, there are few teaching materials designed to provide each child with work suited to his or her attainments.
"Whether classes are streamed or mixed-ability there is a tendency for teachers to teach to the imaginary 'average' student," says the paper. "Research suggests that this satisfies the needs of few students."
They also say the evidence suggests that "teachers often fail to carry out the necessary testing, registration of progress and feedback."