Education, Lifelong Learning and Culture Committee,

Meeting date: Wednesday, May 21, 2008

Official Report 240KB pdf

---

Petition

---

Schools (Class Sizes) (PE1046)

Good morning. I welcome everyone to the 13th meeting in 2008 of the Education, Lifelong Learning and Culture Committee. I remind everyone that mobile phones and BlackBerrys should be switched off. Ken Macintosh is at the Health and Sport Committee, where he is speaking to amendments to the Public Health etc (Scotland) Bill, which the committee is considering at stage 2. He hopes to join us later.

Agenda item 1 is the committee's consideration of petition PE1046, which was presented to the Parliament by the Educational Institute of Scotland. I am pleased to welcome, from the EIS: Ronnie Smith, general secretary; David Drever, vice-president; and Helen Connor, vice-president elect. Thank you for joining us and for lodging the petition.

What would be the benefits to children if class sizes were to be reduced along the lines that you suggest?

Thank you for inviting us to speak to you today. I hope that our contributions will be helpful to you in your further deliberations.

We believe that the evidence in favour of the reduction in class sizes is manifest. The actions of the Government are indicative of that. Indeed, the actions of the previous Scottish Executive were indicative of that, given that it undertook the largest class size reduction programme since the mid-1970s. We are living in a time when the issue of class size reduction has come of age, which our politicians acknowledge.

Significant research evidence shows the value of class size reduction, although it is interesting that none of that research was undertaken in Scotland. The committee might want to consider that. Nevertheless, the evidence shows that where class size reduction has taken place, there have been improvements in attainment in both an immediate and a longer-lasting way. Attainment is only one aspect, although it is an important one. Along with attainment, we point to achievement, by which we mean the overall ability of youngsters to benefit from the education process in the widest sense.

On the current status of Scottish education, it is perhaps fair to say that we are seeing more radical changes in the methodology and the approach to teaching and learning in Scottish schools than we have seen for three decades. The issues before us include curriculum for excellence, the programme of formative assessment and the assessment is for learning programme. There have been changes in attitudes towards behaviour and social skills in our schools. There have also been changes in methodology; there is now a wide range of teaching practice in every classroom—good teaching is taking place. All those things will benefit from class size reduction. That is the golden thread that runs through all those elements; it will hold them together and provide the possibility of progress and improvement in the quality of learning in our classrooms. There will be benefits for youngsters of every age.

One of the key points that we want to emphasise is that we welcome the changes that have taken place so far in secondary 1 and secondary 2, with the reduction to classes of 20 for mathematics and English, and the reduction to classes of 25 in primary 1. We also welcome the programme of reduction to classes of 18 in P1 to P3.

However, the petition's intention is to achieve our aim of reducing to a maximum size of 20 all classes from P1 to S6. We are not asking for that tomorrow, next year, the year after or the year after that, but we would like a programme to be initiated for a staged and phased reduction in class sizes throughout primary and secondary schools. That would move away from the current rollercoaster situation in which small class sizes of 18 are planned for P1 to P3, followed by the possibility of a jump to classes of 33 in P4 to P7. Class sizes would go down to 20 in S1 and S2 maths and English, but would be up to 30 in other subjects from S1 to S3 and in all subjects through to S6. We want a programme to reduce those inconsistencies.

I will leave it at that, because I know that members have other questions to ask.

You have raised several issues, which members will pursue in their own way.

You said that extensive research backed up the EIS's belief that smaller class sizes would improve student and pupil attainment. The committee will hear later from the University of Glasgow, whose submission suggests that the research evidence to support that case is not overwhelming. It would help if you specified the research to which you refer and how it demonstrates an improvement in attainment and the long-lasting effects of such improvements from the early years to later years in a child's school experience.

As members know, the two major gold standards for research on the subject are the student teacher achievement ratio research in Tennessee, which was done in 1990, and the longitudinal work by University of London researchers—the class size and pupil to adult ratio research, which covered early primary years to middle and later years of primary education. The evidence of that research is that significant gains in literacy and numeracy were made in the early years when class sizes were reduced, in comparison with control groups. The hardest evidence comes from the Tennessee STAR research, but the London university work also contains significant evidence.

Concomitant findings are also advantageous. Both pieces of research showed that social relationships and behaviour improved. Attention spans and on-task focusing were shown to be better when class sizes were reduced. The evidence is clearest and hardest in the early years of primary school. There is some evidence that the benefits reduce later in school. The London evidence showed that that might happen, but the Tennessee research showed that drop-out rates in middle and later secondary school were reduced among youngsters who had been members of smaller classes.

I have mentioned only two items of research. The research field on the subject is enormous. The class sizes working party that the previous Scottish Executive established, on which Helen Connor and I represented the EIS, undertook a broad-ranging review of the research. One piece of research in the review took a meta-view of class size research and examined the different pieces of research that had been undertaken. The view of that overarching research was that, in general, class size reduction is advantageous.

We can go to the files and pull out research that says that class size reduction is not worth a jot. Others will say, "Class size reduction might work but it is hellishly expensive to undertake." We are aware of that, but where it has been undertaken thoroughly, and where it is set alongside other important aspects, such as attainment, the evidence for it is strong. If you speak to practising teachers, to folk who were once teachers and are now education managers, or to the majority of people who have been through the school system, they will say, largely unequivocally, that smaller class sizes are beneficial. Some folk may pooh-pooh that, but we think that it is important experiential evidence about the value of smaller class sizes.

In August, when the purveyors of private education in Scotland advertise in The Scotsman and The Herald, they describe their small class sizes as an important benefit of attending their schools. We think that they do that for strong educational reasons, and we would like that type of improvement to be made in the state system as well.

The paper that was provided to the committee by the University of Glasgow said that there was

"little apparent change in performance between class sizes of about 18 and 25".

It said that there was little evidence of lasting benefits and no evidence that class size reduction had an impact on pupil attainment. Has Glasgow university's research centre got it wrong?

The Scottish Council for Research in Education considered the broad range of research. We think that the evidence in the two most important pieces of research, the Tennessee STAR research and the London CSPAR one, which are regarded by researchers as the most robust available, show that benefits accrue from reduced class sizes. It is not an exact science—we do not claim that it is. It is hard to specify the factors that lead to quality teaching in schools and to improvements in attainment, literacy and numeracy. Those factors are hard to isolate, and once they are isolated they are hard to quantify. I would reassert that the research evidence says that improvements will result from smaller class sizes.

The evidence is clear that a reduction in class sizes in the early years—in P1 and P2—is particularly important. That is backed up by evidence that was produced for the class sizes, staffing and resources working group, and in particular the Peter Blatchford research from inner London. The difficulty arises when the reduction is not sustained. The Peter Blatchford evidence proves that when pupils move from primary 2 into a much bigger class in primary 3, and then to bigger classes thereafter, we fairly quickly lose any benefit that we have gained. The research backs up the theory that if we have smaller classes at the early stages, we need to ensure that that continues throughout the school.

I challenge the idea that the only issue that we are talking about in relation to class sizes is attainment. I come from North Lanarkshire, some parts of which are quite deprived. Attainment is hugely important, but it is not the most important thing to a lot of our young people. We live in a changing society, and there are changing expectations, particularly in the curriculum for excellence. If we want responsible citizens, and if we want people to be sitting here in 20 years' time, we need to change the way in which we educate our youngsters. That is as important as academic attainment.

I am sure that the SCRE witnesses will answer for themselves later. However, their paper says that "there is disagreement" about whether the benefits are most marked in classes of fewer than 15 or 20. On page 22 of the report of the class sizes, staffing and resources working group, we find the same phrase. It says:

"There is some disagreement amongst researchers about how much classes must be reduced in size to achieve significant improvements in pupil performance: some argue that benefits are most marked in classes of fewer than 15 pupils … while others … suggest that the major benefits from reduced class size are obtained as size is reduced below 20 pupils."

The answer is that there is a variety of opinion. There is no disagreement, however, on the view that reduced class sizes are desirable. I have certainly never heard anyone argue for increased class sizes.

Is the reduction in class size the only thing that will improve attainment and support the child to become a more rounded individual? If we want to ensure that all our children get the very best out of their educational experience, is reducing class sizes the most important thing that we could do? Helen Connor pointed out that she comes from North Lanarkshire. I am very proud of the teachers that we have in North Lanarkshire and the work that they do in our schools, but I am also conscious that, in my constituency, which has pretty high levels of deprivation in places, there are primary schools in which the class sizes are already 18, but that that does not necessarily mean that those people are getting the most out of their educational experience. Is the reduction in class size the most important thing?

It is the most important thing, but it is not the only thing. The EIS has never said that a reduction in class sizes will solve all the problems overnight. However, if you look at the changing expectations of our education system, you will understand that you need to have a class that is small enough to allow teachers and other education workers to engage with young people.

Take the four capacities of the curriculum for excellence, for example. Engaging with young people and helping them to set their own targets, develop their own goals and become more confident individuals and successful learners is all much more difficult in a class of 30 than it is in a class of 20. The evidence shows that a reasonable percentage of classes throughout Scotland are at the 18 to 20 level. If you talked to the youngsters in those classes, you would see that the experience that they are getting is much better than it would be in a larger class. Time with a teacher is hugely important to young people.

As I said, although class size is the most important thing, it is not the only thing. Teachers must have adequate continuous professional development to enable them to consider the curriculum. We firmly believe that inclusion is a positive thing for education but, again, adequate resources and small enough classes are needed to allow that to develop properly.

Would the EIS prefer there to be composite classes comprising 18 pupils from, for example, P1 and P2, instead of a slightly larger class—say, 20 pupils—that had only P1 or P2 children in it?

No, that is not our position. There were huge benefits to last year's reduction to 25 in P1, but we acknowledge that there were knock-on effects, one of which was the creation of composite classes further up the school. We would prefer to have staged reductions throughout primary rather than big single-stream classes of 25 or 30.

As an elected representative, I have certainly received far more representations from parents and teachers in North Lanarkshire about composite classes. That is my personal experience, but I do not suggest for one minute that the picture is necessarily uniform throughout Scotland.

On that point, we need to recognise that in almost no circumstance is the composite class the preferred model of forming classes. The formation of composite classes tends to be driven by other circumstances. There is a strong case—this has perhaps been missing from the discussion so far—for ensuring that there is no incentive to resort to composite classes as a consequence of any reductions that take place. Under the current arrangements, the maximum class size for a composite class is 25, which is lower than any other normal class size. There is a strong argument that any future reductions to class sizes should include a concomitant reduction in the maximum permitted composite class size, so that there is no almost perverse incentive to move to the composite model.

I am interested in exploring further what evidence supports the argument that smaller class sizes will be beneficial, given that more pupils with additional support needs are now taught in mainstream schools. Is that a matter of how well teachers are trained to deal with children who have additional support needs?

Yes, I think that how well trained teachers were to deal with such needs was an element especially two or three years ago, when much more of the inclusion agenda started to be introduced under the Education (Additional Support for Learning) (Scotland) Act 2004. Although we have always had youngsters with additional support needs who required learning support, we have had many more in the past two or three years. As someone whose teaching is divided among a secondary school and six primary schools, I know that we have many more children with autism and Asperger's syndrome and—to be honest—behavioural needs than we had before.

The need for more training for teachers is an element, but there is also an extent to which we are not being fair on those young people. Having come from schools in which classes had perhaps eight pupils with two adults, they have suddenly been moved into a class of perhaps 25 or 30. I do not think that we are being fair to those young people in terms of meeting their needs and expectations. That, too, has a knock-on effect. I do not for a minute suggest that we should not welcome the change, as I think that it works well in the majority of cases. However, things would work even better in a smaller setting in which those children could get better support.

That is interesting. Obviously, local authorities need to juggle with a number of aspects that might improve educational delivery in their area. If authorities focus on reducing class sizes at the expense of other measures that might improve pupils' attainment, are they striking the wrong balance?

I think that the two things hang together. The point about additional support needs was picked up in a piece of work that was done by staff at the University of Southampton for the then Teacher Training Agency in England. The research looked at successful models of inclusion for pupils with what were then called special educational needs. Let me quote the key characteristic that the research identified:

"All imply that the learning environment plays a key role, since pupils learn through social interactions, and a sense of belonging to, and participation in, the learning community has an important effect in young people's learning in schools … Teachers foster the construction of knowledge through … dialogue with peers."

There is no answer to the point of view that for such interactions to happen, class sizes have to be smaller. If the price of making it possible for class sizes to be smaller in particular circumstances for particular pupils is that other pupils have to go into larger classes—in some kind of trade-off—then we ask for a ceiling or maximum size. We acknowledge that different classes need to be different sizes. We are not saying that every class in Scotland should have 20 pupils, but that that is the cap or ceiling beyond which we should not go in constructing classes for any pupils. Of course, we take account of the circumstances of pupil cohorts in forming classes.

So the initiative to reduce class sizes in P1 to P3 is a move in the right direction, towards what you are trying to achieve.

Yes.

Good morning. Your petition calls for a significant reduction in class sizes. Could you give me an example of what you think a significant reduction is? I know that you have already said that there should be a maximum of 20.

Make us an offer.

We want to balance what we think is educationally valuable with what is economically and politically possible. We are clear that we want there to be a move towards a maximum of 20 in classes throughout primary and secondary school. We recognise that, for many reasons, that is not an immediate goal, but we want it to be factored into future planning.

Looking at current class sizes, a patchwork exists across Scottish schools. With hindsight, we would not have gone that way, but we do not live in a perfect world. The present Government is building on what previous Governments have done, and it needs to do the best that it can. We want the work that has been undertaken on primary 1 to primary 3 to be extended throughout primary school. Where there is evidence of setbacks in attainment, it is where youngsters move from a small class to a much larger one. The time will come when those primary 3 youngsters who have been sitting in classes with fewer than 18 pupils will have to move into classes that might have 30 pupils. That will cause problems with teaching and learning, and disruption to the social groupings that have been set up in the earlier classes. The situation will change again when they move up to secondary school. We want to see moving through schools in stages a smoothing out in the direction of smaller class sizes beyond primary 1 to primary 3.

At the EIS's annual general meeting this year, folk will stand at the microphone and hammer the podium, giving the figures for what they want class sizes to be, which will not be valuable to the committee. You have the task of setting the figures. However, we have said that we want the class size maximum to be 20. The word "maximum" is important. It is the top line. There is plenty of room below it to take account of flexibility and individual circumstances. We want to see a move towards a maximum of 20.

You touched on Government policy. Is current Government policy, which is tied in with the concordat and single outcome agreements, sufficient to meet your aspirations?

I had a look at the transcripts of the committee's robust questioning of Fiona Hyslop in December last year. The committee asked her searching questions about how rigorous the process will be. We share the concerns that there is not enough specificity in and robust monitoring of either the single outcome agreements or the concordat to warrant having confidence in the ability or desire of local authorities to move towards the reductions in class sizes. We want to see what advances are being made. The cabinet secretary talked about year-on-year advances, and we will be looking very closely for them. We want to see real advances being made for youngsters in classrooms.

You will be quite heartened to hear that, last week, South Lanarkshire Council announced that there was extra money to fund 11 additional teachers in 11 primary schools to reduce class sizes to 18 and below. Is the mix of Government policy, single outcome agreements, local authorities' freedom and headteachers' autonomy to run their schools a good mix to bring about good, positive outcomes such as that in South Lanarkshire?

We welcome what is happening in South Lanarkshire. However, you raise an issue that is worth considering and teasing out—the degree of autonomy that headteachers should have. I know that committee members looked hard at that issue in December when they talked to the cabinet secretary. It is a question of central control and devolved authority. If matters are left to individual headteachers, the constraints that they work under—particularly staffing constraints—might lead to situations in which making class size reductions a priority is not possible for them. Similarly, things might happen at local authority level for financial or political reasons.

Of course we welcome individual authorities' advances, but we are concerned that the degree of devolution or handing down of decision making to schools that is involved might lead to a patchwork situation developing within local authorities or at the local authority level, so that distinctly different levels of provision are available to our children. One of the great strengths of the Scottish education system is that it is a Scottish system: we take a national approach to education. That is an important aspect of Scottish educational culture as a whole, and we lose it at our peril.

You have also answered my final question. Thank you.

We welcome what South Lanarkshire Council has done. Equally, we welcome what North Lanarkshire Council and Orkney Islands Council have done. In some areas, councils are moving towards fulfilling the Scottish Government's commitment, but there is a difficulty for us. The Scottish Government made a national commitment, but the expectation is that local authorities will deliver it. Glasgow City Council has clearly said that class sizes are not important to it. It does not see lowering class sizes as the way forward, and it has been up front in saying that it has no intention of doing anything about class sizes because it does not see them as a priority. My question in return—I do not know whether I am allowed to ask this, but I will—is, what will the Scottish Government do to monitor local authorities in whose schools there clearly will not be year-on-year reductions in class sizes? The expectation of class size reductions has been built up not only for teachers and young people but for parents. It is fine to show us the positives—there are positives—but what will we do about people who simply say that class sizes are not important to them and that they do not want to, or cannot, reduce class sizes?

A concordat was signed, and the single outcome agreements will be agreed to. We shall see how far Glasgow City Council goes after that.

Perhaps it would be more appropriate if the committee asked the minister that question when she comes before us and we deliberate on the petition with her. I am sure that members will want to ask that question.

Good morning. I want to pick up on an interesting point that Mr Drever made about maximum and minimum numbers of pupils being set by the Government and local authorities. Does the EIS accept that, working within those numbers, there is scope for headteachers and their staff to decide what is best for the pupils in their school, given that they are the professionals on the front line?

The issue of flexibility is different from setting a class size maximum. For example, the setting of most of the current class size maxima in 1974, when they were enshrined in the teacher contract, did not prevent flexibility being employed in schools or stop headteachers taking strategic and tactical decisions about what sizes classes would be in their schools. That power has been available to them and they have exercised it judiciously, or otherwise, over the years. Local authorities have a similar, although lesser, opportunity to do the same. However, EIS research shows that they tend not to do so, and instead they devolve that decision to schools.

We are saying that such flexibility should continue to be offered to headteachers. Our key question is whether a class size maximum will be set to ensure that there is an agreed level that classes will not go over, although we know that, on occasion, that cannot be helped. EIS members who find themselves in that unusual situation will not throw up their hands; they will try to find a way of resolving the situation as soon as possible, in order to get the class size down to what it should be. We are saying that a maximum should be set, but that within it there should be the ability for headteachers to exercise a degree of professional judgment and flexibility.

The logic of that position is that, in a local authority that has a large geographical area and a mix of types of schools, it might be sensible educationally—which is the most important point—to have a slightly different attitude to minimum and maximum levels in a class. Do you accept that, in some local authority areas, we might have to be a bit more flexible?

I would need to be persuaded on that. I would need to look at the existing mix. I come from Orkney, which has a mixture of town and rural areas. We have a remarkably wide range of class sizes, which are dictated by the demographics in the area. Sometimes, because of constraints of one sort or another, we struggle to get class sizes to the level that we want. I see no reason why that situation should not continue to pertain. The professional judgments that are currently made about class sizes, how to deploy staffing resources and so on will need to be made whether the class size maximum is 20, 25, 30 or whatever, and whether in urban or rural situations.

I am committed to lower class sizes in principle, but I have a difficulty with Government and local authorities making the final decision about class sizes—I think that it should be a matter for headteachers. Further, I am particularly concerned about what might happen in local authorities that have a broad mix of schools, because what might be right for one part of the area might not be educationally sound in another. I am concerned by the one-size-fits-all approach, and I am interested in the EIS's view of the educational aspect.

We do not think that one size fits all either, and that is not the approach that is taken at the moment. Again, if you go into the schools in the area that you represent, you will see that there is a wide variety of class sizes, according to circumstances. We want that to continue, because it exists at the moment for good educational reasons. However, we also want the bar to be lowered across the board. That is not a one-size-fits-all approach. We recognise the value of reducing class sizes, we will take steps to ensure that class sizes are reduced to the maximum, and we will deploy the existing processes by which decisions are made about how best to use the available resources in that regard.

Do you accept that if a school was doing particularly well and had good reports from Her Majesty's Inspectorate of Education but had class sizes of slightly above the maximum, it would not necessarily be a priority for that school to reduce its class sizes to below 20? Would the fact that the school simply did not meet a target be a problem?

That is a difficult question, because implicit in it is the suggestion that schools, teachers and pupils would be punished for doing well. There is the suggestion that, because that school was doing better than others, we would keep its class sizes larger. That argument does not hold water.

I want to clarify what your petition asks the Government to do. Your pay settlement was renewed over Christmas after dialogue with the Government, as part of the normal regime. Did you express your views on class sizes in discussions on the teacher contract? Did you raise that issue with the Government?

That was not part of the discussions on the salaries review. Your question relates to an issue that almost arose earlier this morning, which is the different means by which the various class size maxima are set out, or promulgated. As well as having a bit of a rollercoaster of different class size maxima according to stage, we now have, in effect, four different mechanisms for promulgating what the maxima ought to be.

In respect of certain stages, we have collective agreements that go back to the mid-1970s. In respect of primary 1 to primary 3, we have statutory instruments. We now have the new device of the concordat and the single outcome agreements. In respect of the previous Executive's commitment to class sizes of 25 for primary 1, we have a departmental advisory circular.

There is a confusing landscape of different ways in which the limits are set out. Some methods are more robust than others. One of our concerns is that where an authority faces an application and admissions request, the grounds on which it might refuse that request need to be robust. The departmental advisory circular will not wash. It is not clear to us that a single outcome agreement will be able to be relied upon.

As well as having greater consistency and coherence in the different levels, there is an argument for having greater consistency in the method. Our preferred method is for there to be a collective, tripartite agreement. We are committed to reducing class sizes, we would like to think that Scottish local authorities are all equally committed to it, and we understand that the Government is committed to it. Those are the three partners in the Scottish negotiating committee for teachers. That is the best way in which matters can be taken forward. It is about voluntary collective bargaining, which requires all three parties to play before it can come to pass.

I want to explore that further. There was an opportunity to re-examine the tripartite agreement through the SNCT, but you said that discussing class size reduction was not part of the process and that you simply had a salary review and update. You are one of the three partners in the SNCT. I did not pick up that you asked for class size reduction to be considered in a round of discussions. I am talking about the period after local authorities and the Government had agreed to a concordat, of which class size reduction was part. Why did that not trigger your saying that you wanted the issue to be discussed? Petitioning Parliament is okay, but you are one of the three partners in the SNCT who will bring class size reduction about.

The question of an approach to class size reduction is not particularly associated with the pay discussion. It is not that there is a once-a-year opportunity to meet and that anything and everything that needs to be discussed has to be brought to the table at that point.

I am sorry to interrupt, but the SNCT continues the teacher contract, and the class size requirements are included in the contract, which has been renewed, so would not the SNCT have been the mechanism for considering the requirements?

No. The agreement on salary is not about continuing or discontinuing the teacher contract: it continues unless there is a decision to do otherwise. The salary discussion was one component of the teacher contract discussions.

To return to your point, we can raise an issue with the SNCT at any time. Since the SNCT was formed, we have tried to hold behind-the-chair discussions in smaller groups to get agreement, at which point the issue is brought to the SNCT. It was made clear to us that neither the Government nor COSLA was interested in using the SNCT to take the question of class sizes forward. We could have made a political point, laid the issue on the table and formally failed to agree, but that would not have been very constructive and it would not have moved us forward.

We are in the public domain, so that is now on the record.

Just for clarity, the legislative framework is within the Parliament's scope, and it is interesting to note that that is part of the issue. The Education (Lower Primary Class Sizes) (Scotland) Regulations 1999 legislate for a maximum of 30 in a class. Are you asking Parliament to look again at the existing statutory framework or to ask the Government to introduce new regulations, or should we just leave that alone?

We are looking for the most robust mechanism. In a hierarchy of preference, I would like it to be done through the SNCT. If that is not going to happen, a statutory instrument is the second-best choice, because it is legally enforceable and it sends a strong, clear and unambiguous message to everyone who is involved. That is the most desirable and most secure solution against any external challenge.

As I understand it, you wish the Government to use regulations to legislate for a maximum of 20 pupils per class in all years of primary school, so the subordinate legislation would move away from requiring a maximum of 30 pupils per class in P1 to P3 to requiring a maximum of 20 pupils in all primary school classes. Is that accurate?

Yes. If we were negotiating, I would say that I would like there to be a maximum of 20 pupils per class throughout primary and secondary school.

Absolutely, but the 1999 regulations affect primary schools, and they are the only statutory requirement that we have. Would you prefer there to be a statutory requirement for all classes in all years of formal education?

Yes. We would like a clear, unambiguous and enforceable maximum class size across all stages of education.

As Jeremy Purvis said, we are on the public record here, and we do not want the EIS to be on the public record demanding class sizes of 20 in P1 to P3 rather than class sizes of 18. The EIS has a policy that the class size should be 20 throughout primary and secondary school. We are delighted to see the present Government working to reduce class sizes to 18 in primary 1 to 3. I say that for the record.

I return to the issue of how local authorities deliver on the policy. My constituency is in West Lothian, so I am sure that you will understand where I am coming from when I ask how we can address the concerns that have been raised by some local authorities about how, with increasing populations and therefore increasing school rolls, they can simultaneously reduce class sizes at an acceptable rate to meet the requirements of the policy.

The answer lies with the mechanism that the Government uses to deliver its policy, which means the robustness of the concordat and the quality of the single outcome agreements. We have done our own research. We have a local EIS association in each local authority area, and we have asked them how specific single outcome agreements are regarding class sizes. Our concern is that there is very little specificity and that, in most cases, single outcome agreements make no mention of class size targets. The generality in the concordat at the national level is echoed at the local level, which makes it difficult for us and for our members on the ground in local authorities, because we feel that we have a role, alongside the Government's role, in monitoring the policy's success. We are concerned that, judging by the early indications, there is little attention to detail in the single outcome agreements and there are few practical proposals for how the Government and local authority concordat will be carried forward in the coming year.

Ms Connor said that Glasgow had said that class size reduction was not its priority. However, some local authorities would like to make it a priority, but they have increased demands on their provision. I note what you said about the single outcome agreements. It is some months since the budget process, and the committee found it quite difficult to see where the resources were being invested to address the specific needs in local authorities. Were you any more successful in establishing whether there were additional resources for areas such as West Lothian?

If I am being honest, I do not think that we were any more successful. We met the cabinet secretary and talked through the concordat and the single outcome agreements. We monitor our local associations, which are involved in local negotiating, and there does seem to be a difficulty. The concordat has been signed up to, but although the Scottish Government says that the resources are available for a reduction in class sizes and for other issues, it is not clear that that is the case. Interestingly, while class sizes are mentioned generally in the concordat, they are not part of the list of expectations at the end of the single outcome agreements, which is quite disappointing.

That question perhaps needs to be probed with COSLA. Our best understanding is that an agreement was made between the Scottish Government and COSLA on behalf of all the local authorities. However, we detect a certain dissonance, because while there have been high-level pronouncements that there are sufficient resources to make class size reduction possible, some authorities are not singing from the same hymn sheet. That gap could be worth exploring.

That point is well made—the committee will attempt to take it up with COSLA. We may also raise with the minister the issue of how the outcome agreements have been arrived at.

I have one further question, which is on the previous experience—it has already been alluded to this morning—of reducing class sizes in primary 1 and in secondary 1 and secondary 2 for English and maths. Should any lessons from that be borne in mind to inform how the policy is implemented in the future?

We welcomed the reductions in class sizes that Peter Peacock set in process as an outcome of the partnership agreement of the previous Government, but the issue of flexibility was a matter of concern for us. Up until December 2005, we worked on the assumption that the specified class size maxima—20 pupils in secondary 1 and 2 and 25 in primary 1—would be implemented. We were therefore concerned and surprised when we received an indication from the then Scottish Executive that a flexibility factor would be introduced. We have discussed the merits of flexibility today, but there was no consultation or discussion on the issue then. Our impression was that the change was due to petitioning by the Headteachers Association of Scotland.

The proposal was perhaps not fully thought through, so it caused a degree of concern and alarm in schools both about the educational reasons behind the policy and about the practicalities of organising classes. More thought should have gone into the policy and more consultation with teachers should have taken place. One lesson that we can learn is that consideration and thought needs to be given to the policy's implementation, on which there should be consultation with the interested constituencies. That is what the committee is doing today by considering our petition on class sizes.

Ronnie Smith raised an interesting point about COSLA. The committee was unaware that COSLA had declined to give oral evidence on the petition. We discovered only yesterday that it planned to submit just written evidence to us. In light of the EIS's comments, the committee might need to revisit consideration of whether to compel COSLA to give oral evidence. The EIS has raised some legitimate points about the implementation of the concordat that we might want to pursue with COSLA directly.

Does Elizabeth Smith have any further questions on the implementation of the policy?

My question was answered, thank you.

I want to turn to financial considerations. The report from the Scottish council for research in education states that a policy of reducing class sizes would be expensive to roll out, given the costs for more teachers and the need for classroom renovations to cope with smaller classes. Has the EIS assessed how much such a policy would cost to implement?

Helen Connor is trying to find a copy of "Class Sizes, Staffing and Resources Working Group: Final Report", which suggests a number of different costs for additional teachers. More difficult to quantify—it could be done only by local authorities—are what changes would be necessary to the school estate to provide additional teaching spaces and classrooms and, possibly, extensions. We do not hold that information.

Do you foresee subsequent savings from rolling out the policy if it raises children's attainment in the early years?

It would be very difficult to give monetary figures. I completely understand why the committee has to consider the financial implications of any policy, but we also have to consider possible detrimental effects on society in the longer term. There may be future financial implications in terms of social work and health, for example.

I am the person from the EIS who harks back beyond attainment. This is not just about attainment; it is about developing young people who can contribute well to society—healthy young people for the future. If we do not consider the future financial implications now, things will be much more difficult. We cannot be short-termist; we have to look to the long term.

Does a blanket reduction represent value for money because of the long-term impacts? Do you have any sympathy with people who feel that other methods of raising attainment should be considered? Could other methods be cost effective, too? Where should the emphasis be?

A blanket reduction specifying a maximum level would be hugely beneficial in the long term. However, we have to ask whether academic attainment is the only issue. I am not saying that academic attainment is not important—it is hugely important for a vast number of our young people—but we also have to consider where some of our young people start from, what they can achieve and what they can contribute to society.

In a previous question, points were raised about flexibility within a reduction target. We are talking about the educational needs of our young people. I will give an example. I have a second-year class with 13 pupils. Although they are in second year, they are at level B in maths, which is primary 5 level. You could not meet the educational needs of those young people in a bigger class. We have to take a longer-term view that considers more than just academic attainment.

We discussed research earlier. As Helen Connor said, attainment is just one strand that we have to consider. If we believe the research that shows that smaller classes can lead to better attainment, less disruption, less exclusion and fewer instances of pupils leaving school at an early age, it would appear that smaller classes can have an economic benefit. However, it will be devilishly difficult to show a direct causal connection.

Some people have argued that reducing class sizes will not be good value for money and that it is not the best way of spending scarce resources. A few months ago, someone from London said that the return from reducing class sizes is poorer than that from formative assessment. That is an odd juxtaposition because we would say that the introduction of proper formative assessment would require reduced class sizes. It is not an either/or situation. Too often the debate is posited as, "You can have either this, or this." In fact, the two things must go together.

Do you accept that smaller classes may be achieved only in stages? Are you happy with the present direction of travel, and should we be travelling any quicker?

Yes—we accept that staging will be required. As I have said, our policy is a long-term policy that takes account of the financial and political issues as well as the predominant educational issues. The policy would be staged and phased.

The first question that Aileen Campbell asked was about the cost of the policy—you will have seen us trying to find the figures, which turn out not to be here. However, a chapter in the class sizes working party report gives projections of future education spend with different scenarios of class size reduction. It contains valuable information about projected costs. Because of changing demographic trends and the reducing birth rate, pupil populations are projected to fall in the period to 2030. It is suggested that the education budget will fall significantly in the same period. I am aware that that demographic scenario is contested by inward immigration and population change statistics, but there seems to be a general trend towards the reduction that I described. The point was made at the class sizes working party that that reduction would free up resources that could be spent on reducing class sizes. That is why we feel that holding teacher numbers at 53,000 would be valuable. If teacher numbers were held constant as the school population fell, that would allow class sizes to be reduced.

Like Mary Mulligan, I am concerned about the effects of increasing population and school rolls in my constituency. The General Register Office for Scotland's figures for my local authority area in the Borders project a 15 per cent population increase over the next 20 years, with increasing school rolls. If the policy of class size reduction depends on demographics, there will be no class size reduction in certain areas unless there is the corollary of additional teaching staff and teaching capacity in schools. Many schools in my area are bursting at the seams—their school rolls are not falling. You say that the demographics policy is the correct way forward; in fact, it is now enshrined in the concordat that demographic trends will deliver class size reduction. However, what kind of education system have we when class sizes are predicated by a Government policy that depends on the area in which children are born?

I reiterate that we have a national education system. The scenarios in the class sizes working party report are national projections. It is clear that the demographic trend in some parts of Scotland is different from the national trend. However, a national perspective must be taken because if the policy of class size reduction is dealt with at local authority level, it becomes a postcode lottery in which certain areas are discriminated against because of the trends there. We want the policy to be handled at national level to ensure that particular areas and schools that have particular demographic trends are not discriminated against—there must be a national perspective. The class sizes working party said that there will be opportunities from national demographic trends up to 2030.

That is on the basis that capacity can be freed up where the demographics are beneficial. The EIS wants existing teacher levels to be retained in areas where school rolls fall because of demographic changes, which would mean, de facto, that class sizes will be reduced because schools will not reduce their teaching capacity. However, that will not free up resources in the way that would be possible with a national approach, whereby the resource would be put into areas in which there were population increases. Basically, the policy will mean that demographic benefits will be entrenched in certain areas. However, that policy is to be the deliverer of maximum class sizes of 18 in P1 to P3. You seem to welcome the policy, but I am concerned about it. In the area that I represent and in other areas in which there will be population growth, there will be no additional capacity in schools.

Demographics alone will not suffice. The demographic argument is that there is a general downward trend in population figures, which will offset or reduce the additional costs of moving to class size reductions. However, as David Drever said, we want the policy to be dealt with at national level.

Jeremy Purvis's point comes back to the issue that I raised with COSLA. There are issues about distribution across 32 authorities and the nature of the settlement that is negotiated between the Government and COSLA. One would like to imagine that that settlement would take some account of shifting demographics in Scotland because it is not sufficient to look just at the global picture. There is the question of how we distribute resources and teachers across Scotland. We are highly familiar with that process, even when demographic shifts take place at authority level. Demographics are not the sole answer, but they offer an opportunity to contribute to offsetting the additional cost.

How do they do that? Demographic changes might drive a reduction in class sizes, but given that the same number of teachers and classes will be retained, although there will be fewer pupils in each class, no savings will be made. How will the shift in demographics produce cost reductions, given that it is driving the policy of delivering smaller class sizes?

The answer is to do with the interaction between the availability of resources because of what happens nationally and how those resources are deployed locally. As Ronnie Smith says, that is where COSLA should have a key role. We are not saying that if one local authority has an increasing pupil population—

Even at national level, the potential to reduce the cost of spending on education as a result of demographic changes is predicated on the assumption that fewer teachers would be needed because there would be fewer pupils and fewer classes, so the amount of capital and the size of the school estate would not need to be as large. However, if demographics are used as the driver for delivering a reduction in class sizes, teacher levels must be retained, if not increased, and the size of the school estate must be maintained, so no savings will be made.

But smaller class sizes will be made possible, because it will not be necessary to provide for an extra space if a space is freed up as a result of a fall in the number of pupils.

I am sorry, Mr Smith—the expenditure line will not go up, because the demographics will mean that there is no need to build more schools, but it will not go down, because you have said that it is necessary to retain the same number of teachers and the same size of school estate. It is just that there will be a smaller number of pupils in each class.

That will result in more classes meeting the class size limits to which we aspire.

Yes—but with the same number of teachers and the same size of schools.

Our objective is reduced class sizes, which is what the demographic changes will result in.

The committee wants to know how that will be funded. I am questioning the EIS's apparent agreement with the Government's assertion that demographic changes will allow resources to be freed up at national level to help deliver smaller class sizes, because that position is predicated on the retention of the same number of teachers and the maintenance of schools of the same size. All that will happen is that the number of pupils in each class will go down. Class sizes might be smaller, in accordance with your agreed policy, but investment will have to be maintained at the same level in areas in which there are demographic falls and increased investment will be required in areas in which the population is growing. Do you believe that the Government is funding that policy? Savings cannot be achieved as a result of demographic changes if teacher numbers and school sizes remain unchanged.

We are not asserting that demographic change alone will deliver the desired outcome. It might help to offset the cost, but additional investment will be required to bring about our objectives.

We do not know whether there is sufficient money in the settlement between the Scottish Government and COSLA: the committee must ask COSLA about that. Both partners seem to be saying that they think there is sufficient money in the settlement. We cannot gainsay that. I reiterate that we are not relying on demographic changes alone.

I want to pick up on a point that was made earlier, because it is relevant to the important financial issue that we are discussing.

It would help us to determine the deployment of teachers to meet the policy objective if we had a clearer understanding of where deprivation occurs. Even if we take the Scottish multiple deprivation index definition, it is clear that in rural areas up to 50 per cent of youngsters are living in deprived circumstances. That is far more than in urban areas, where the overall figure is lower, although pockets of severe deprivation are to be found. Only 30 per cent of the population live in rural areas. Obviously, in the context of this debate, we are talking about restricting class sizes and the financial impact of doing that. Are there other areas of debate that would help to clarify the way in which teachers are deployed?

Yes. The nature of deprivation is complex and the tendency is to represent it as an urban phenomenon. However, those of us who live and work in rural areas know of the deep deprivation there. In a sense, such deprivation is made worse by the fact that support services are not developed in those areas and because there is not necessarily a culture of support. We recognise that. The results of the recent Organization for Economic Co-operation and Development report indicate that such issues are important when it comes to education of our children. The report suggests that attainment and achievement are defined by where a child comes from and not which school they go to. The matters that Rob Gibson raises are important. We would like to see the Government elaborate on, or sharpen up on, its work into deprivation. When that is done, we will be able to see how to utilise resources.

I will add a rider, albeit that I am not sure whether it is relevant to the question. We want to see areas of deprivation resourced fully and, at the same time, ensure that resourcing is done throughout Scotland—it should be Scotland-wide. We do not want to see pockets of resourcing while other areas do not get the resources that they require.

Fair enough.

I think Mr Drever mentioned teacher numbers in the responses to a previous question. Obviously, if we are to deliver class size reductions, sufficient numbers of teachers are required. We are coming to the stage of the academic year when future employment is an issue. Are you content with the number of teachers we are training at the moment? Will we have enough teachers overall to deliver on the class size policy?

I wish that I knew the honest answer to that question, so I speak with some caution on the subject. The number of teachers in preparation is the product of the workforce planning exercise. Certainly, last year, we faced a considerable problem with teachers coming out of induction who were unable to find employment—there was an apparent oversupply of teachers. That said, it was, perversely, difficult to persuade people to work in some areas of the country. Mobility issues may be involved.

We are somewhat nervous about the prospects for the coming period in the light of some of the publicity that we have seen in recent months around local authority budget setting and the numbers that authorities are contemplating employing. It is an extraordinarily difficult task to balance exactly the number of teachers that we need and the number of vacancies that are to be found in the employing authorities. The general trend is upwards.

I am not sure what assumptions will be factored into the next workforce planning exercise because the response to the move towards reducing the class size maximum in P1 to P3 is going to be highly variable across the 32 local authorities in any given year, and it will be spread out over three years. It might be even more challenging than it has hitherto been to fine-tune and get right the number of teachers in preparation in balance with the different ways in which we might be moving towards class size reductions.

I recognise that this is not a simple question because there are issues about experience and how to provide balance within the schools, but does the EIS have a figure in mind for the number of teachers it thinks will be needed to deliver the reduction as it is planned at the moment?

No, we do not have a specific figure in mind. Reference was made to the class sizes, staffing and resources working group in which a number of different models for changes in class sizes were posited alongside the possible numbers of additional teachers that might be required. The report contains different statistical modelling.

In your original answer, you talked about your concerns about teachers looking for work. For clarity, did you mean teachers who have completed their probationary year and are now looking for employment, or were you talking about those who are entering their probationary year?

As we understand it, those who are entering their probationary year are guaranteed places: the Scottish Government guarantees that a training place will be found by hook or by crook, so that ought not to be a problem, although there might well be issues with persuading authorities to offer sufficient places to meet the output from the teacher education institutions this summer.

Our main concern is about what happens at the end of probation when teachers enter the open jobs market. We know that it is a problem every year, although sometimes it has turned out to be not quite as bad as it is feared this side of the summer holidays. Last year was a bit difficult; the Government applied additional funding for, I think, 300 additional teachers to help ease that pressure. As I say, we are nervously considering what the pressures will be this coming August.

Is that additional funding on-going funding that will be available for this year?

I think that I have seen somewhere that Fiona Hyslop said that that funding has been built into the baseline figure for this year.

That is helpful, and I am sure we will come back to the point.

That concludes our questions for the panel. Thank you for your attendance at the committee. The committee will suspend for five minutes to allow our witnesses to change over.

Meeting suspended.

On resuming—

I welcome our second panel of witnesses, who are Valerie Wilson, honorary senior research fellow with the SCRE centre—formerly the Scottish Council for Research in Education—and Jon Lewin, information officer with the SCRE centre. I thank them for attending and for their written evidence.

We have several questions to ask. I have some general questions about the research on class sizes. As the witnesses sat through the evidence from the previous panel, they will be aware that the EIS defended rather robustly its view on the need to reduce class sizes and believes that the research from other parts of the United Kingdom and around the world suggests strongly the importance of doing so. Is that research on class sizes relevant, given that it has not examined the Scottish experience, or can we learn clear lessons from it that are valuable to Scotland?

Good morning, everyone. Before I answer the question, I wish to say that Jon Lewin and I welcome the opportunity to appear before the committee to discuss the research that SCRE has conducted. It is difficult to summarise concisely the work that we have undertaken. I will clarify one point about our written statement which, as members will have noticed, highlights our long association with the EIS. Some of you will know that SCRE was founded by the EIS in 1928. Although we welcome that association with the EIS and continue to have it, Jon and I are here in a separate professional capacity to try to give the committee an independent view. It occurred to me when I looked at the written statement that I should say that we have not been in communication with the EIS on the matter prior to today.

To return to the question, as researchers, we welcome the fact that politicians are taking research seriously and that you have invited us here to discuss it. In some ways, we are hoist by our own petard, in that we are trying to summarise an extensive body of research and bring it all down to answer a few direct questions and then say whether it is relevant to Scotland. To be honest—and to answer your question directly—my opinion is that much of the research is not directly relevant to Scotland.

Before I became a researcher, I used to be a teacher. In the past, I may even have paid my subscription to the EIS. However, during my early years as a teacher, I taught in the United States of America, although not in Tennessee.

The EIS says that the strongest evidence comes from the student teacher achievement ratio—STAR—project. That was a state-funded project in Tennessee, which involved 79 schools and several thousand children and ran for a number of years. It is clear from even the most cursory glance at the project that Tennessee is not Scotland and Scotland is not Tennessee. Therefore, we cannot separate out one factor—class size—and then argue that, because the approach worked in Tennessee, it will work in Scotland. That is a specious argument, which I cannot support. Lessons can be learned from other education systems, but the important point is that we are talking about the Scottish education system, as Ronnie Smith said, of which we are all part and which has its unique traditions and culture. Whatever is decided must be grounded in that culture. Therefore, I am reluctant to say that the experiment in Tennessee can be transferred to Scotland.

I will say a little about the experiment because, unless members have read the research, they will find it hard to visualise what was done. STAR was an experimental research project—I am sure that members, like the witnesses, have strong views on whether we should experiment with children. The Tennessee state government authorised researchers to divide children into classes. Small classes had fewer than 17 children and regular-sized classes had 17 to 22 children. Some regular-sized classes were assigned a teaching aide.

Members can imagine what happens when school class sizes are manipulated. In Tennessee, that was done openly, not secretly. All parents, children and teachers knew to which classes they had been assigned. That approach is not one that we researchers would subscribe to, so there were in-built problems with the research even at the design stage. That led to a further difficulty, which was the attrition rate. If I was a parent whose child had been assigned to a larger class, I would petition the school to move the child to a smaller class. That is what parents in Tennessee did, so there was movement between groups. Therefore, although the project was conceived in an experimental way, it became contaminated as the years rolled by.

We know from previous research that if we put a label on a child or a teacher, they will live up to the label. If the expectation is that children will perform better in smaller classes and that teachers are happier teaching smaller classes, the results from the groups who have been assigned to smaller classes tend to be more positive than the results from other groups.

That is the background to the Tennessee research project. It is considered a "gold standard" by many people—I use the inverted commas advisedly—because it was done on a large scale and was well funded, and because it ran for many years and involved independent researchers from the University of Tennessee and statisticians from the University of Chicago. However, it had in-built flaws, as Peter Blatchford, whose evidence is far stronger, says. The method that was used is suitable in medical research—medical colleagues regularly use randomised control trials—but such an approach is not suitable in education research.

Any results from the Tennessee project must be understood against the background that it is an experiment that we probably would not have engaged in and that it is difficult to transfer results from a very different education system with a high percentage of minority ethnic children to a much more homogeneous, smaller country such as ours. We must therefore be cautious when we interpret the results.

Both the previous Government and the current Government have been committed to reducing class sizes. If we are properly to assess the benefits of reducing class sizes, what do we need to do to make the research effective? What would you recommend that we do to evaluate the effect of reducing class sizes in Scotland, initially to 25 and potentially to 18?

You are almost inviting me to design a research project. I am not convinced that there is a need for further research and that we have not got part of the answer already.

I will explain what SCRE did, so that you can get a fuller picture. We did not conduct our own research on class size, because to do so is expensive and fraught with some of the difficulties that I have said are inherent in trying to design a research project that focuses on one factor—class size—when a child's learning and what happens in the classroom are complex.

I will take you back to 2001, when SCRE had a service level agreement with the Scottish Executive. Part of the agreement was that we provided five or six literature reviews a year to inform policy. Jon Lewin and I worked together on that. We conducted the first review in 2001. To give you an idea of what we were up against, Jon will tell you how many pieces of literature we had to handle.

I do not have exact figures. We would broadly call the kind of review that we conducted a systematic review. There was a definite system to the search and to the sifting of evidence. The guidelines laid down by the evidence for policy and practice information and co-ordinating centre at the University of London Institute of Education, which regularly conducts systematic reviews for the Government, suggest that a year is required to do a proper systematic review, but we had a few weeks in which to gather the evidence.

We initially found not far short of 1,000 articles, which we had to sift through. Many of them replicated the same bits of evidence. Due partly to the academic tenure system in the United States, there is an awful lot of repetition of articles, because the more an academic publishes, the safer their position is. When it came down to it, there were still a few hundred articles. Those examined every aspect of class size and related matters, which were impossible to disentangle. As well as class size, there were issues of classroom environment, school size, setting and streaming within class groupings, additional support in the classroom from teaching assistants—teachers' aides, as they call them in the US—and a host of other factors that were impossible to disentangle.

Therefore, we had a good few hundred articles to sift through. Valerie Wilson's conclusion in the report says something to the effect that the findings are at worst contradictory and—I cannot remember the exact phrase—

At best confusing.

Yes, the conclusions were at best confusing and at worst contradictory. That was certainly how I felt about the research process that we were involved in.

As to whether further research could shed any light, I suspect that a few more articles would just add to the list for whoever has to conduct the next literature search—instead of finding 300 articles, they would find 310. Like Valerie Wilson, I do not believe that anyone could conceive of a piece of research that would suddenly cut through the mist that surrounds the issue. Class size is perhaps a uniquely complicated issue. It is certainly one on which there is a unique amount of published evidence. It is widely discussed in parliaments around the world, and the evidence in all countries is probably puzzling to the various committees that are considering it—as it was to us.

May I follow up Jon Lewin's description of the search? We identified more than 800 pieces of information in 2001 and another 300 in 2006, when the working group on class size asked us to update our work. Basically, we found that studies fall into four types. It might be helpful to grasp the differences so that, when the committee makes a judgment, it is informed by an appreciation of the evidence.

First, there are reviews, which are similar to what we did: we read the evidence and tried to review and synthesise it and reach some conclusions. Secondly, there are correlational studies—there are lots of those—in which researchers try to find some relationship between two factors. In this case, the factors are class size and attainment.

The evidence seems to be that lots of studies can show that if you reduce class size, you can drive up attainment in certain measured areas—usually in literacy but sometimes also in maths. One problem with experimental studies is that they are expensive to undertake. There have been several in the USA, not only in Tennessee but in Wisconsin, California and Florida, as well as in Edmonton in Canada. They all seem to show that there is a relationship between a smaller class size and raised attainment, but the problem is that that relationship is not causal, in the sense that many other things may cause attainment to rise.

We can—and medical colleagues do—try to show relationships between various factors. In trying to show whether there is a relationship between, for example, people's height and their health, someone might be able to draw a correlation, but that does not mean that one causes the other. That is the important point. If you reduce class size, you have no guarantee that you will drive up attainment.

The studies show that there is a relationship, but no one seems to know why. The thing that we have not discussed—and which our colleagues from the EIS did not discuss—is what actually happens. What are the dynamics of teaching and learning in smaller classes that might be associated with increased attainment?

Our colleagues from the EIS mentioned meta-analyses. I think that they were referring to the meta-analysis by Glass and Smith, who reviewed all the experimental evidence, put the data into a computer and came up with the view that all the studies support the contention that, if we reduce class sizes, we will increase attainment. There is also the experimental STAR project in Tennessee.

The project that we have not discussed yet is Peter Blatchford's multilevel modelling project, which was undertaken at the Institute of Education at the University of London. By British standards, it was a massive project. It involved 15 local authorities in England and several thousand children. Peter Blatchford and his team followed the children from reception class, which is roughly equivalent to our nursery, through to year 6, when the children were aged 10 to 11. Peter Blatchford is an extremely good, careful researcher. He examined the experiment that was conducted in Tennessee, saw its deficits and problems, and tried to rectify them in his design. Some of the factors in his design have overcome some of my reservations about the strength of the evidence.

In some ways, one has to be a statistician to understand multilevel modelling. The idea behind it is that the researcher controls certain factors and tries to show that the factor in which he is interested really does bring about the effect. Peter Blatchford added classroom observation, so, unlike the Tennessee study, his study took into account the way in which teachers behave in classrooms when class sizes are reduced, which is important. Do they spend more time with the children? Do they have more smaller groups? Is there more interaction? His answers seem to be yes. On the positive side, that is what happens if class sizes are reduced.

The difficulty with his evidence is that he also found a disruptive effect. If pupils who were in a small reception class are moved to a larger class, the benefits disappear, so he cannot show a lasting benefit. That is one of the dangers of a policy that reduces class size in lower primary but not in upper primary. Schools might run into a disruptive effect and lose any benefit that they had gained.

To go back to the question whether there would be any benefit if Scotland tried to replicate any of the studies and design its own project, I do not think that we have the finances or resources to do that. Millions of dollars are poured into the American research. You will note from the report of the class size working group that it designed an elaborate research project but that that remains uncosted. The project was supposed to run from 2007, which coincided with the change in Administration, and report in 2011. I assume that the project, which would look for Scottish evidence, was never commissioned. I have been unable to find it on the Scottish Government's website.

That does not mean that it is not worth evaluating any policy changes—it is. However, designing a large-scale project is probably not the way to go.

Thank you. A number of members have questions.

Good morning, panel. I want to pick up on something that was said earlier about the research. Will you confirm that SCRE was asked to provide a review of both old research and current research in 2001 and in 2006?

Yes, that is correct.

I have one other short question. Did the previous Scottish Executive ask SCRE to carry out any Scottish research other than the review?

We had lots of projects that were commissioned by the previous Scottish Government. SCRE lives by bidding for projects. Yes, we did lots of research in Scotland but not specifically on class sizes.

Some of my questions have already been answered. Can you tell us a bit more about what the evidence suggests the impact of smaller class sizes is on educational attainment? Does such a policy do more than just raise attainment? We heard from the EIS witnesses that the policy would also increase achievement. Can you expand on that?

I agree entirely with my EIS colleagues that education is about more than attainment. Any research project on education would consider a host of indicators, including children's social and emotional development and their development of practical skills.

While focusing on attainment, many of the research projects also provide evidence about what happens to the dynamics of classes when class sizes are reduced. In smaller classes, there seems to be more on-task behaviour, in that pupils spend more time working on the task that is set by the teacher. On-task behaviour is used as a proxy for learning. Of course, some children might have learned that, if they keep their head down as if they are writing with a pen, they look like they are on task. That is a possibility. However, on-task behaviour is usually a fairly reliable proxy. If children look like they are working hard, they are on task. Teachers can tell when pupils are off task because there is lots of non-task-related chatter.

The research seems to indicate that, in smaller classes, there is more on-task behaviour and less aggression and pupils seem to be more solicitous of each other. An important point is that smaller classes allow good teachers to do what they know they should do without the fear and worry of larger discipline problems. In a large class, the teacher is constantly concerned about whether children are off task. As class numbers are reduced, the teacher's concerns and worries dissipate and children spend more time on task. The children also seem to be more socially related to each other. Even if the research does not prove that the increase in educational attainment is lasting, it is to be hoped that some of the social gains would continue.

One paragraph in the SCRE submission states:

"there is little evidence of lasting benefits."

However, paragraph 3.7 states:

"US evidence claims lower ‘drop out' rates".

Is that one of the contradictory findings of the research? Will you say a wee bit more about why the research did not find any lasting benefits from lower class sizes?

I have not personally done any research on whether smaller class sizes produce lasting benefits. Peter Blatchford's research states that the benefits are not evident by year 6. The American research uses different proxies because America has a different system. As members may know, in the American system it is possible to fail a year—which is practically unknown here—and to be required to repeat that year. As a consequence, children who are not making progress rapidly move down the school because their peers advance. They therefore drop out because they reach school-leaving age without finishing the statutory curriculum.

A follow-up to the STAR project concluded that the group who had entered smaller classes in kindergarten had a lower suspension rate, a better absentee rate and a lower drop-out rate in high school. The problem is that only about half the pupils who had entered the experiment were identified, so half had been lost by the time that they reached high school. That places a question mark over how reliable and robust that evidence is.

So nothing that you looked at suggested that smaller class sizes increased confidence.

No. I found no evidence of that.

What other factors did you find to have a positive impact on attainment?

Our brief was not to consider all factors that affect attainment, but to identify research on class sizes. In effect, we started from the other end.

I see. That is everything.

I do not know whether you have said everything that you can about the effect on pupils' behaviour, attendance and motivation. Can you say any more about what the research that you examined said about behaviour in class, attendance and motivation?

The American follow-up studies to the STAR project make claims on behaviour. As a proxy for behaviour, they use suspensions; that is a fairly high level, given that a child's continuing low-level misbehaviour might never result in their suspension. As a proxy, the rate of suspensions picks up the most extreme misbehaviour. The claim is that, by high school, those who had experienced the full four years of a small class—from kindergarten to grade 3—showed significantly lower suspension rates, better attendance and lower drop-out rates. Those are the three proxies that were used.

The difficulty with the STAR project is that nobody went in and observed the relevant classes—the research was hands off. One delight and pleasure of being a researcher who is no longer in front of a class is observing other people in front of a class. We see some wonderful teaching in Scottish schools, although we also see some pupils misbehaving, which we do not want to see. The STAR project researchers never went into a classroom to observe how the teacher and pupils reacted to smaller classes. However, Peter Blatchford's research team did that. He asks the interesting question whether lowering the class size alters the teacher's behaviour. That is an important issue that needs to be considered.

You have spoken a lot about the STAR project and the Blatchford project. Did you examine other research that showed that the advantages of smaller class sizes outweigh the disadvantages?

In terms of teachers' behaviour, the answer is yes. In the early 1990s, studies by Hargreaves and Jamison in London brought important issues to the research table. They suggested that smaller class sizes allowed teachers to do the things that good teachers know they should be doing. For example, more one-to-one teaching would take place; questioning would be more challenging; and individual pupils would have more contact with the teacher. Previous English research and the Peter Blatchford research has shown strongly that lowering class sizes offers the opportunity to alter the dynamic between the learner and the teacher.

You made a point earlier in your evidence about the STAR project and about class sizes of fewer than 18 and class sizes of between 18 and 22, and you said that, if you were a parent, you would be petitioning the school to have your children put into the smaller class. Do you agree that pupil engagement and positive experiences at school are more important than attainment, and that engagement and positive experiences can be achieved with smaller classes?

You are asking me to agree that pupil engagement is more important than attainment.

Is it an either/or? Or are they equally important?

No, it is not an either/or. I have read the EIS petition and, as a former teacher, I would welcome the opportunity to support it. The EIS has probably overstated the evidence, but the petition is strong where it says that Scottish education has changed dramatically. I was educated in a primary school class with more than 40 children. Okay, I managed to get to university in the end, but it was not education. A lot of it was training and disciplining and drilling.

We are now asking teachers to implement a new curriculum, which starts from a completely different philosophy. It would be easier to do that with smaller numbers in the class. We are also asking teachers to cope with a wide range of abilities and needs. It does not seem to make that much difference whether you put in a classroom assistant or classroom aide to help the teacher; the quality time is the time that is spent with the teacher. Anything that you can do to allow pupils to have more dedicated teacher-pupil engagement will be an improvement and will ultimately lead to improved attainment. The evidence of research is that that will probably be necessary. However, it will not be sufficient in itself, because other things will also need to happen.

Our principal focus is on the Government's proposals. The English evidence—from Blatchford, I presume—is that the impact of smaller class sizes on younger and less-able children has been confirmed.

Yes.

Are there particular teaching methods that are more suited to smaller class sizes?

When you go to observe teaching nowadays, the thing that you notice is that far less whole-class teaching happens in all classes. Teachers organise their classes into groups and, in primary schools, pupils sit in groups. The question then arises whether those groups should be single ability or multi ability. Teachers' views vary on that, but most teachers agree that, if the class size is reduced, teachers' flexibility to group the children is increased. Teachers group children, anyway, but large groups of seven or eight, compared with groups of three, four or five, minimise the possibility for a child to interact with their peers and the teacher, when they come round. Smaller classes give teachers the flexibility to have smaller groups.

The Cabinet Secretary for Education and Lifelong Learning has quoted the governor of Oregon, who said that children spend the first two years learning how to read and the rest of their life reading to learn. To me, that encapsulates the important point. It is not cost effective if children do not learn to read at the beginning of their educational career. Therefore, if we have smaller classes, smaller groups and a more intense relationship with the teacher, the chances are that more pupils will learn how to read with their peers and we will not need remedial action later.

Are you suggesting that the size of the class affects teaching practice and that therefore younger and less-able pupils would benefit from smaller classes?

Yes.

On another tack, is there any research evidence on how multi-teaching, for example with two teachers teaching one big class, impacts on pupil attainment?

No. Local authorities welcome the flexibility to be able to have two teachers and a large class, as that can save on capital costs or help if accommodation is not available in a school. However, that situation alters the dynamics entirely. The relationship between one teacher and 18 to 20 pupils is entirely different from a relationship between two teachers and 30-odd pupils. Those are two different entities. The research that we considered did not envisage providing two teachers simply to meet a target and bring down the ratio.

I want to ask about funding and budgets and cost considerations, which were part of the literature review. Am I correct that research was commissioned by the class sizes, staffing and resources working group on the impact of class sizes on standard grade results?

After I had presented my evidence to the working group in 2007, I ceased to have any involvement with it. I understand that the group commissioned two further pieces of research. The first, which was done by York Consulting, considered the decision-making process by which local authorities and schools determine class sizes. The second was a pilot study in, I think, North Ayrshire that was undertaken by Linda Croxford at the University of Edinburgh to consider whether there is a relationship between standard grade attainment and class size. From the working group's final report, my understanding is that the researcher concluded that the period was too short to come to any conclusion and proposed a far more extensive project.

The conclusions of the working group report state:

"Most evidence is in primary but the research commissioned by the CSWG on the impact of class size on Standard Grade results suggests that further work could be carried out to help assess the impact on secondary."

Therefore, it was not entirely accurate to say that no research had been commissioned.

In 2006, when I wrote the review, that was correct.

So the previous Government commissioned research and the working group recommended that more research be done. That is still under consideration by the Government.

The rules of tendering allow the Government to commission research up to £10,000 without going out to competitive tender. If a research project does not go out to competitive tender, there is no way for other researchers to know what research is being commissioned.

We can ask the Government about that.

The Scottish Parliament information centre briefing indicates that some research has highlighted that a focus on reducing class size does not provide the best value for money and that money could be better spent elsewhere on more one-to-one learning, more social work input, more educational psychologists and more support for pupils who are not progressing as fast as others. To what extent was value for money covered in the literature review? Did a substantial element of the literature consider alternatives?

No, but it was noticeable that value for money featured in a significantly bigger proportion of the research that we looked at in our update, which covered only five years in the early part of the decade, than it had done in the previous research. A lot of the research is American and, given that the American system is perhaps ahead in the arc that we tend to follow with regard to class size, that suggests that a backlash had begun, certainly among economists, and that it was gathering momentum by the early part of the decade. Having said that, we are talking about only a handful of economists. As I suggested earlier, their output depends on other factors, but some are more vocal than others. A number of economists, particularly in the US, are spearheading something of a backlash. I think that we could describe it in that way.

In answer to Jeremy Purvis's question, in 2001 I was not asked to look at any economics articles. In 2006, when I took the draft report to the working group, it asked me to add another section and to go back and look specifically at articles in economics journals. That work indicated—I am not sure how helpful this will be—that there were two opposing views.

The first view was represented by Hanushek, an economist from Stanford, whose argument seems fairly damning. Looking at the whole of the American system, he suggests that the pupil teacher ratio had improved; the percentage of teachers with masters degrees, which is a proxy for quality in continuing professional development, had increased; the average experience of the teachers had risen; and real-terms expenditure had risen. Although all that had happened across the system, attainment on national tests remained stubbornly flat—I think that those were his words. That seems pretty damning, but another economist, Krueger, who is also from Stanford, challenged his evidence by trying to work out the internal rate of return from a project. He looked at how much was spent on raising attainment and gave attainment a monetary value—for example, graduating from high school was given a monetary value—and then worked out the internal rate of return. His conclusion, which contradicts that of Hanushek, is that for every dollar spent two are returned, so there is a very good rate of return.

Was any comparative research done on situations in which one authority or area had used the resource to reduce class sizes, while another area had used an equivalent resource to provide more educational psychologists, more additional support staff and more social work staff, and had undertaken whole family intervention with some youngsters? Was any such research conducted, particularly in areas of deprivation or with children with complex needs?

No, but economists—including Hanushek, who is sceptical about the value of lowering class size—acknowledge that class size is only one factor. No system would do just one thing; other factors would be going on concurrently that would impact on a child's educational attainment and experience of school. Hanushek concludes that lowering class size has an effect in some cases, but the difficulty is that it is a blanket and differential effect, so it is questionable whether resources should be used in that way. As an economist, he suggests that targeting resources is more effective than applying them in a blanket approach.

You mentioned Blatchford's review and said that a conclusion of the class size and pupil adult ratio research project in England was that there was no evidence that pupils in smaller classes from year 4 onwards made more progress in maths, English or science. The EIS petitioners would like all primary and secondary class sizes to be capped at 20. What research has been done on the effect of reducing class sizes to the same level across all year groups and courses?

I do not know of any.

There is none.

That is why Peter Blatchford's hypothesis is that there is a disruptive effect.

Has all the research focused on a particular age group?

Yes. The focus has been on the early years.

That concludes our questions. I thank the witnesses for coming. I am sure that the committee will return to some of the issues that you raised when we take evidence on the petition in the future.

Meeting continued in private until 13:19.