Timothy Mason's Site

Creative Commons License
This work is licensed under a Creative Commons License.

Please note that the pages in this section are taken from e-mail posts to FLTeach. They were written quickly, in reaction to other posts. They are fleeting opinion pieces.

Class Size and Educational Outcomes

Messages to FLTeach

How many children should there be in a class-room? Most people believe that the fewer the better - but it isn't necessarily so. Although it is probably true that if class-sizes were all reduced to between 7 or 8 learners, and if they were taught by teachers with special training for small classes, a lot more learning would occur, the difference to acquisition that could be made by reducing numbers from 30 to 25 may not be very large - at least, if the research is accurate.

The following messages were all sent to the list FLTeach. I have included the dates, and have bold-typed the message-headers, so that anyone who wishes may follow up the threads in the FLTeach archives. Where I have included short extracts from posts by other members of the list, I have used italic script.

The first time I saw the question come up on FLTeach was in August of 1966, when I had only been on the list for a couple of months. In this instance, an undergraduate student sparked off the discussion by asking whether anyone could provide him with arguments to persuade the university where he was studying Chinese to lower class-numbers. Other complaints about overcrowded classes followed - from teachers. I contributed a short message, recommending a book which I still think has a number of good ideas on how to tackle teaching to large groups :

Aug 1996


For some of us, large classes are a fact of life, and the way things are going, it is unlikely that they will get any smaller. Personally, I rather like large classes. Teachers who are faced with thirty, forty, fifty or a hundred students - yes, it does happen - could do worse than to consult the little book 'Large Classes' by Rob Nolasco and Lois Arthur, published by Macmillan in the Essential Language Teaching Series, isbn 0-333-43672-5.

Best wishes

Timothy Mason

The subject then cropped up again in October. This time, I responded by referring to the general research on class-size :

Oct 1996Class-size and successI do not know of any research into the relationship between class-size and success in language learning as such. Research has been done on the relationship between class-size and general attainment. The results are intriguing and ambiguous, and on the whole, teachers don't like them very much. I will provide references as soon as I have been able to dig them up from the heaps of paper that encumber my work-room.

  1. Class size does appear to make a difference in primary schools. Younger children do better in smaller classes.
  2. For secondary schools, Rutter et al (Fifteen Thousand Hours ; Secondary Schools and their effects on children, Open Books, 1979) write the following : ... neither average size of school class nor pupil-teacher ratio seems to be consistently associated with school behaviour or attainment ....

Indeed, what trend exists is for *larger* class size to be associated with *better* attainments. What does seem to happen is that there is a qualitative change once the group becomes larger than 10-12 pupils. Efficiency then bottoms out at around 24/5, which is the class-size associated with the worst results, and *then rises* again, particularly in schools situated in working-class districts - that is, working class children do better in larger classes (One hypothesis is that this enables them to escape negative judgements from the teacher - another is that it imposes lock-step style teaching, which works better with working class children. It is middle-class children who shine when taught using liberal teaching approaches)

Rutter later reported that there did appear to be a positive correlation between class-size and delinquency - that is, children in larger classes are more likely to become delinquent. Later work has also found some positive effects of smaller class-size, but not, I believe, a great deal.(P. Mortimore et al. , 1988, School Matters, Somerset, Open Books)

It could be that language teaching is different and that smaller classes are better in this one case. And it could be that the researchers have got it all wrong (it has been claimed that the research is flawed - but you can always find flaws in educational research, if you look). But I don't think that there is much comfort for the classroom teacher in their findings.

One thing - as Jencks put it, in 'Inequality', both teachers and pupils say that they feel happier and more relaxed in smaller classes. That seems a pretty valid reason for having smaller classes - but let's be clear what we mean by smaller. I would not, in the light of present knowledge, want to create classes of 24/5 pupils. 10/12 seems about right, and might even work out cheaper.

Timothy Mason

Then in December, I came across a more recent piece of research :

Dec 1996 Class size and outcomes again

I have just stumbled across further research into the relationship between class size and success - Times Educational Supplement, 6/12/96. The research , done by a team from Leicester University, decided to tackle the question of why smaller classes produced such little measurable gain by looking at how teachers behaved when given smaller classes. They concentrated on 'expert' teachers, as indicated by trainers and headteachers, who taught in the public Primary school system, with relatively large classes. These teachers they then swapped with teachers from private schools who had small classes. The lessons were observed. Researchers also observed the public school teachers when they had their own classes, but with numbers halved. They found that time spent on critical control was lower, there was more feedback and a higher proportion of sustained interactions in the smaller classes.

These are all factors that research leads us to expect to result in better outcomes. However, the actual change in these directions was quite small. As the authors of the report state : 'Teachers did not appear to maximise the advantages of the smaller classes ... These findings 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.' They go on to say that studies of class size effects will continue to find little difference between large and small classes so long as teachers are not trained to take advantage of smaller classes.

For my own part, I wonder whether we may not advance the same argument for larger classes. Outcomes for such classes could improve considerably if teachers were given proper training, (I definitely feel that this means mainly on the job training) and if they were allowed to teach in ways which encouraged use of cognitive skills.

This may be a far better use of resources. As another article in the same issue of the TES points out, dropping average class size from 23.3 pupils to 22.3 would cost an extra £51 .58 per pupil. Dropping to 21.3 would cost £107.96 more per pupil. Dropping radically to 18.3 would cost £313.69 per pupil. This money - supposing it to be available - could be spent on cutting teacher contact time, and on providing in-service training. It could be spent on learning assistants to provide back-up support, or on clerical assistants to cut the amount of time teachers spend on administrative duties. I know that very few of my colleagues share my bizarre liking for large classes - am I totally out on a limb on that one? - but those who would advocate cutting class sizes by a meaningful number do need to make a very good case, and do need to demonstrate that they will radically alter their teaching approaches if their desires are acted upon.

Yours statistically

Timothy Mason

Another member of the list wondered whether large classes wouldn't put learners off pursuing optional subjects :

Dec 1996 Class size and outcomes again

I am curious to know how many FLTeachers experience dramatic attrition rates and have oversubscribed beginning forms and undersubscribed advanced forms. Is the experience in differing class sizes already there, or is what is available simply not applicable among forms?

Jonothan Centner raises the excellent question of the link between motivation and class-size. The figures I have cited in my last post are for primary schools, where children are not asked whether they want to study a particular subject or not. Jencks tells us that both students and teachers *prefer* smaller classes, so one may imagine that larger classes for electives will result in higher drop -out rates. I do not know whether there is any empirical backing for this supposition. Following this up might lead us to demand smaller classes for electives only .

I don't know how teachers of compulsory subjects would feel about this. Here in France, the teachers I really do feel for are teachers of French and of Philosophy who have large classes. They have to face an extremely heavy load of correcting papers - far more than the average FL teacher.

Finally, I would say that certain styles of teaching can make large classes work, and that it is what the teacher (and the students) make of the class that is the most important variable. I have taught large classes well, and had satisfied students in these classes. I have taught small classes badly, and have had dissatisfied students.

Have a good Christmas

Timothy Mason

There are grounds for skepticism about this kind of research. In a follow-up post, I outlined some of the reasons why we should take the results with a pinch of salt :

Dec 1996 Further reflections on class size

The most damning critique of class-size survey results is that they use either formal exams or standardized tests to measure educational outcomes. This at least has the advantage of judging the institution on its own terms, but almost certainly does not address the full range of behavioural changes that education brings about, whether positive or negative. Thus, for example , Rutter and his colleagues note that although class-size does not correlate with standard measures - public exams - of success, it does correlate with delinquency (although it needs to be noted that there are problems in the measurement of delinquent behaviour that we have not the space to go into here). The overall relationship between schooling and the likelihood of being labelled delinquent is a strong one, such that the peak age for delinquency was raised in England through the simple expedient of raising the school-leaving age.

The argument about class-size, then, can be seen to go to the very root of thinking about education. If it is merely the business of schools to enable children to adopt the behaviours which lead to success in standardized tests , then the number of children in a class is of very little importance, and administrators are quite right to ignore demands that these numbers should be cut. If we believe that schools serve other purposes, and that those purposes are to be fully attained - at least in part - through reducing class size, then we need to argue convincingly against the presently prevailing educational ideology, and we need to show how those ends will be furthered by smaller classes.

Realistic arguments against large classes are of necessity radical arguments . And I would suggest that we, as teachers, certainly need to come to terms with the finding that working-class children do better in larger classes, for at least one of the implications is that increasing the amount of attention that the typical teacher devotes to such children actually lessens their chances of learning. If this is so, then we certainly need to subject our own practice to intense scrutiny and radical overhaul.

Yours moralistically

Timothy Mason

I returned to the subject in February the following year - once again, the hopes of those who were looking for arguments from research to back up their demands for lower class-sizes had been dashed :

Feb 1997 Depressing stuff about class sizes

According to a recent edition of the Times Educational Supplement, a reanalysis of the Tennessee project further restricts the benefits of small classes. According to Professor Sig Prais of the National Institute of Economic and Social Research ('Class size and learning", Oxford Review of Education, Vol 22, Number 4), reducing class size to 15 has only a negligible effect on most children's rate of progress during the first three years of school. He says that now research should concentrate on the benefits of alternative text books, more detailed teachers' manuals, other forms of class organization and teaching styles and *more lesson preparation time for teachers* (one of the enormous advantages of the French secondary system - Moi)

Pupils with SEN do benefit from smaller classes. But while there is still some evidence to suggest that small classes might be beneficial in first year of primary school - although even here there are reasons for skepticism - Prof Prais found that in the second year of elementary school, children in smaller classes make slightly less progress in maths than those in larger classes. Teaching assistants had a negative effect on progress - but that they may be useful in classes where there is an undue proportion of difficult pupils. More research needs to be done on the utility of teaching assistants.

Timothy Mason

Another list-member insisted that no improvement at all could be expected from reducing class-room numbers, unless there was a radical change in pedagogical approach. But research doesn't suggest you can go quite that far :

Feb 1997 Depressing stuff about class sizes

I will bet all I have (and do not) that reducing class to one-on-one tutoring will have the same result. The answer then may not lie in the teacher alone and his/her "full" access to student's brain...

I'd hold on to your money tight if I were you. Studies do suggest that, as Cindy said, once you whittle group numbers down to 10 or below, results improve considerably. This seems to be a fairly common finding across the board in social psychology - human beings work best in groups of around 10 people, or a little less. They feel more comfortable that way, and so perform better.

Of course, this is not automatic ; the behaviour of members of the team - and in educational situations, obviously the behaviour of the teacher - is a crucial variable. I've always felt that, if we kept numbers right down, and trained teachers to take advantage of the situation, we could probably teach as much in a couple or three years as schools do now in 12. But then, of course, the horrible question would need to be answered - what are we going to do with all those children for the other 9 or ten years?

Timothy Mason

Subsequently, I did not get involved in any threads on this topic - until it cropped up once again in 2001:

May 2001 HI, I am also a first year teacher but DO NOT HAVE 45 kids either!!! I have never heard of such a thing. I teach ESL and can't even relate to that. I feel that is crazy and I think you should try and teach to who is listening. if you know what I mean. You can also asign things that gets the kids working as groups. I can only imagine what it would be like trying to talk to 45 kids in SPANISH! That's hard, good luck.

45 is okay. I've done 70 and I know people who've had to work communicative classes of 120-40. And don't take the advice to just teach those who listen ; your job is teach everyone and to get them listening. And it can be done.

So - look at the FL-Teach FAQ on class-size, look at Marilyn's advice to new teachers - part of another FL-Teach FAQ, and the best introductory document to teaching I know - and get hold of 'Large Classes' by Nolasco & Arthur (ELTS Macmillan, 1988).

BTW, a recent study in the UK found a small but significant effect of lowering class-sizes. But the gains are still small enough to lead administrators to conclude that there may well be cheaper ways of getting better results. Given the measurable differences that can be traced to having a better teacher, I reckon they are probably right. We already spend far too much on schooling - and not enough of that on teachers' salaries.

Best wishes

Timothy Mason

Author's CV | Site Map | Contact Author |