SCHOOLS have to balance the need for academic excellence with demands to promote social integration. For the past 30 years we have used two systems of distributing pupils into classes: mixed-ability classes in S1-S2 and sets across the curriculum in S3-S4.

My argument is that mixed-ability classes cannot deliver high academic standards and do not seem to deliver social integration either, while sets enable pupils to be taught at the level and pace that suits them but at the cost of diminishing cohesion in the school community. It is surely time to look again at streaming.

The defects of mixed-ability classes are well recognised. Group work and individualised learning give teachers extra work and require an unrealistic degree of self-motivation from many pupils, especially boys. Some strategies that work well in primary school are simply not practicable in secondary schools where many teachers meet their class for short periods once or twice a week.

The strategy of presenting a unit to the whole class but differentiating the follow-up is based on the false premise that each unit can be self-contained and that pupils will begin each subsequent unit at square one without any transfer of skill or knowledge from the previous one. The biggest drawback to mixed-ability classes, however, is that they signal to pupils and teachers that academic attainment is unimportant.

At least mixed-ability classes provide pupils with a stable learning community. We intend that they should also create social harmony, inside and outside the school. It is unfortunate that we put together children with different abilities and ask them to do academic work, which focuses attention on the very differences whose effects we want to smooth.

If we dared to look at what really happens, we might find that these classes generate more antagonism than harmony. We should not be surprised. In S1, children are entering puberty and are least confident and yet they have to exchange the constant supervision of their primary 7 teacher for the minimal supervision of innumerable subject teachers.

The advantages of setting are more evident. We can allow for differences among pupils in a particular subject, and for extra and smaller sets. This means less work for teachers, more whole-class teaching and a greater sense of common purpose. When we limit setting to one or two subjects, we limit the disruption to social cohesion and provide beneficial variation. When we set across the whole of the curriculum, however, as in S3-S4, we totally destroy social cohesion. We add complexity to the pupils’ day and degrade the quality of pastoral supervision. We may be contributing significantly to absence and alienation.

If both mixed-ability classes and wholesale setting are flawed, what about streaming? Suppose we stream from S1 on the basis of ability in language so that the whole class can use the same level of text, but set for maths, in which different children are usually stronger or weaker. We then provide the one clear benefit of mixed-ability classes: a stable learning community. We also provide the benefits of setting: an education that matches attainment and aptitude, and variation in class size to the advantage of the weakest. We can go further and develop different programmes, so that pupils can progress at different speeds towards different targets.

This is where we meet the first argument against streaming: that it labels children for life into As, Bs and Cs. I would answer that we must label if our education system is to work efficiently, but we must clarify the wording. We should take a long-term view and invoke the benefits of lifelong learning. By labelling, we are not putting a ceiling on the level that the child may attain. We are simply indicating which stream offers the most appropriate pace of learning and subject content.

A second argument against streaming (and setting) is that it demoralises pupils and teachers. I do not see why a pupil should be more demoralised by being in the lowest class than by being the lowest pupil in a mixed-ability class. In fact, being in the slowest class should offer the best chance of making progress at their own level and pace, provided of course that the teacher seeks to match the subject content and style of teaching.

A third argument is that streaming gives no guarantee of social mixing, one of the prized objectives of the comprehensive system. But we should accept that social mixing is not practised widely in the adult world and note that mixed-ability classes do not effect mixing across age-groups. To repeat a point made earlier, academic work does not provide the most propitious environment in which to generate social harmony. Can we find another?

Many schools are already divided vertically into houses. If we made sure all years and all streams are represented in each, we could make them the focus for homework clubs, involvement with the community and a wide range of competitions framed to reward participation at all levels. Houses would be led by an assistant headteacher, who would work through the class tutors (register teachers), who would be the first point of contact with parents and collators of reports on pupils.

Streams from S1 through to S6 would be based on proficiency in language, modified by setting for maths, complemented by houses for extracurricular activity and underpinned by an ethos that defined success in terms of improvement as much as in level of attainment. That is more likely than a combination of mixed-ability classes and 100 per cent setting to deliver both academic performance and social integration.

* Next week: the case against.

David Hill, a former principal teacher of guidance, chairs a primary school board in Edinburgh.