"Strike threat over class sizes" is a familiar Easter headline as the teachers' unions hold their annual conferences. This year was no exception, with the National Union of Teachers demanding legislation to set a maximum limit of 20 pupils per class and delegates describing large state-school classes as a "national scandal". Their indignation acquired an extra edge when Jim Knight, the schools minister, told another union conference that classes could work well with as many as 70 pupils, provided there are sufficient teachers' assistants around.
Unfortunately for the NUT, research provides little evidence in favour of small classes. The best that can be said is that they lead to significant gains in academic test scores for pupils in the very early years of schooling, particularly if they are disadvantaged. But among children in Year 3 and upwards, class size has no measurable effect on literacy and numeracy levels. These results emerge from large-scale American studies as well as a current project at the London University Institute of Education.
The usual explanation - that schools put the less able and less well-behaved children in smaller classes - is exploded by the most recent research, which takes account of such factors as prior attainment and home background.
It is all monstrously counter-intuitive. All over the world, politicians promise smaller classes as a token of their commitment to education. Despite their outstanding past results in subjects such as maths, east Asian countries such as Taiwan, South Korea and Japan have policies to reduce class sizes. Here, parents pay thousands of pounds to fee-charging schools, where primary-age classes have 10.7 pupils on average, against 26.2 in the state sector. Given that teachers' salaries account for the lion's share of any school's costs, parents are being overcharged, if the research is correct, by something like 100 per cent. Can everybody be mad? It is surely common sense that children in small classes, whatever their age, ability and background, will get more of the teacher's attention and therefore learn more.
In fact, research proves at least part of the common sense. The latest findings from the Institute of Education project, presented to the American Educational Research Association this month, found that the larger the class, the less the pupils concentrated on their work (or engaged in "on-task behaviour", to use the jargon). This was particularly true of low attainers in secondary schools who, in a class of 30, spent twice as much time off-task as they did in a class of 15. However, class size had no effect at all on medium and high attainers in secondary school. And for children older than six, the research remains clear: the effects of small classes on test scores are nil, zero, zilch.
How do we explain it? The "progressive" lobby in education would argue that teachers do not sufficiently adapt their teaching to take advantage of small classes. They may, for example, still spend most of their time addressing the class as a whole and fail to use the greater opportunities to give individual attention. They may even use less small-group work because the class as a whole is easier to control. The "traditionalists" would argue that, on the contrary, teachers adapt their methods too much. Given a small class, they drop whole-class teaching, which, regardless of numbers, is the most effective method of instruction.
Another possibility is that, leaving aside the first year or so of primary school, the academic benefits of small classes kick in only when the pupil numbers drop well below 20, and perhaps below 15, as they do in the fee-charging sector. Dylan Wiliam, deputy director of the Institute of Education, argues that most teachers can't do anything in a class of 20 that they couldn't do in a class of 26. The individual attention they can give to children is still limited. The difference to the Treasury, however, is enormous, because the class of 20 entails an increase in teacher costs of more than 25 per cent. There are, Wiliam argues, more cost-effective ways of using public money.
To my surprise, I find myself in sympathy with Jim Knight. He is not the first minister to suggest that, with the growth of computer-aided learning and the advent of teachers' assistants, it is absurd to talk of "class size" at all.
Margaret Hodge, then chairing the Commons education select committee, put forward a similar argument in the New Statesman ten years ago. There may be some occasions, in secondary schools at any rate, when children manage perfectly well in groups of 75; others where they should get half an hour of individual tuition.
Small classes serve as a convenient slogan for unions and politicians, because they are easily understood and accepted by the public as self-evidently a good thing. It is time we moved beyond them and thought more creatively about how we use educational resources.