Smaller does mean better
After a hectic day coping with 35 lively primary children, few teachers would disagree with the statement that "life would be easier if class sizes were reduced". A report from the Office for Standards in Education which claims that reducing class size is neither economical nor educationally justifiable, except for reception classes, is likely to confirm teachers' views that the Chief Inspector of Schools has it in for them and is selective in his use of evidence.
Yet even those in favour of smaller classes admit that the evidence on class size is puzzling. Research in English rural schools, where class sizes ranged from 9 to 35, revealed only a small positive association between reduced class size and pupil performance, thus supporting other correlation studies in Canada and the United States. Most commentators explain such findings by pointing out that class size was not investigated as the main effect and that, consequently, important factors such as mixed ability were not controlled. In support of this argument, the Tennessee STAR Project - which showed that children did better in smaller classes - is cited.
The weakness of these experiments, however, is that no attempt was made to discover what teachers in the smaller classes did that made them more effective. Two decades of research have shown that it is in classrooms where there is more challenge, greater amounts of feedback and more sustained question-and-answer sessions between pupils and the teacher, that children do well. It seems self-evident that with fewer pupils in a class, there should be more time available for these key interactions which are so supportive of learning. In which case, why are the learning gains relatively small?
It was this question that the National Union of Teachers commissioned us to examine. Our study focused on two groups of teachers with contrasting experience of small and large classes, who were identified by colleagues in teacher training institutions, local authority advisers and their headteachers as being "experts". We chose these teachers because research suggests that experienced, expert teachers are flexible and more likely to adapt quickly to changed situations, such as being moved from larger classes to smaller ones. By this means, we hoped to maximise the advantages of working in small classes. Next, we located seven buddy teachers from the private sector who regularly taught small classes - the largest was 21, and the smallest was 10. The private classrooms operated a regime that was not dissimilar from that found in the maintained schools within the sample.
Teachers spent two days in the others' classrooms; on the first day they familiarised themselves with their surroundings, and on the second visit they were observed teaching the classes. They were also observed teaching similar lessons with their own classes. Five of the seven teachers from the private schools went back and taught larger classes in maintained schools. Since it could be argued that having to teach unfamiliar pupils was not a fair test on teachers' abilities, we observed four of the teachers from the maintained schools teaching half their own classes on a further occasion. As a result, we were able to compare teaching small classes in private and maintained schools.
The results of the study support teachers' common sense view that smaller classes provide greater opportunities to work with individual children, provide more challenge, allow more feedback and enable longer sustained interactions with individual pupils. Typical results are shown in the three figures where levels of both feedback and sustained interaction increase as class size falls below 20. In contrast, levels of critical control decrease. The trend in the use of closed and open questions was similar to that for feedback but there were as many factual questions asked in classes with 25 or more pupils as there were in those with less than 20. In the large classes, the dominant mode of instruction consisted of giving information and routine directions rather than probing children's ideas.
However, none of the trends present in smaller classes were of sufficient magnitude to produce dramatic improvements in pupil attainment. Teachers did not appear to maximise the advantages of the smaller classes, even when only half their own class of pupils was present. These findings therefore do not necessarily lead to the conclusion that class size does not matter but rather that teachers must be trained to operate more effectively in smaller classes by maximising the use of key interactions.
The results of this pilot study suggest that the arguments about the effect of small class size will not be settled by current studies. Such studies will continue to report low correlations between class size and pupil attainment unless teachers receive training before being assigned to smaller classes.
This leads to a further conclusion. The new national curriculum for teacher training should include a requirement that all students have the opportunity at some stage of working with classes of fewer than 20 pupils. Students should be expected to work with half classes while the class teachers take the other pupils. Students would concentrate on maximising the use of key interactions such as challenging questions and giving feedback.
When we interviewed our teachers after their lessons in small classes, they were clear on the advantages of being able to give children more time to think "of not just getting glimpses so that you presume a lot in a sense". One teacher said: "I felt I was actually teaching rather than as I sometimes feel, just a manager".
Where teachers were able to spend sustained periods of time with these pupils, they produced work of high quality. In general, we felt that in today's typical classroom, with an overloaded national curriculum, it is a wonder that children do as well as they do. Currently, the same research team are evaluating part of the Government's Superhighways initiative, where children not only have to acquire basic skills but also to think creatively and productively. The research reported here indicates that such thinking requires that the numbers of pupils with any one teacher should be as small as possible.
Class size, reading and pupil achievement., Leicester University