Inclusive and Supportive Education Congress International Special Education Conference Inclusion: Celebrating Diversity? 1st - 4th August 2005. Glasgow, Scotland

home about the conference programme registration accommodation contact

THE RELATIONSHIP BETWEEN INCLUSION AND
PUPIL ACHIEVEMENT IN ENGLISH SCHOOLS

Peter Farrell, Alan Dyson, Filiz Polat, Frances Gallannaugh and Graeme Hutcheson
Faculty of Education, University of Manchester

Correspondence: Professor Peter Farrell, University of Manchester, Oxford Road, Manchester, M13 9PL, UK.
Peter.farrell@man.ac.uk

Background to the research

For some years now governments all over the world have been committed to the development of an inclusive education system. The powerful arguments supporting such a move are set out in government documents - for example the SEN strategy document Removing Barriers to Achievement (DfES 2004) - and in the literature (e.g. Thomas and Vaughn, 2004).   By and large this commitment at government level has been supported by a good degree of consensual support amongst teachers, LEA officers and parents for the principle of inclusion. There has been much less agreement, however, about how far this principle can be realised in practice and, insofar as it can, about what the impacts might be on the achievements of pupils with SEN and on their peers in mainstream schools. As previous work has shown, a further concern has arisen in schools who feel that their performance might be damaged - either in reality or in the way their outcomes are reported publicly - if they were to become 'too' inclusive (Dyson & Millward, 2000).  

There is a small body of research which addresses the question of how inclusion impacts on the achievements of pupils with and without SEN.   However, this research tends to be somewhat limited in its scope and often in the context of small-scale inclusion ‘experiments’. Such evidence is illuminating and by and large finds in favour of inclusion.

To date there has been no large scale study in this country that has addressed the question as to whether inclusive schools (defined here as those that admit high proportions of pupils with SEN), do worse by many or all of their pupils because the presence of pupils with SEN distorts school processes in some way. Alternatively, do such schools actually do better because they become more skilful at responding to individual differences?

The availability of the National Pupil Database (NPD) has resulted in much more comprehensive data becoming available in England that allows for more comprehensive and powerful analyses of these questions than could be done previously. This database brings together for the first time data at individual level on pupils’ attainments (in national assessments) and their education-relevant characteristics (including school attended, gender, entitlement to free school meals [FSM] and, above all, where, if at all, they are placed on the SEN Code of Practice’s graduated approach). It thus becomes possible, not simply to identify the impacts of small-scale inclusion experiments, but to look at the relationship between inclusion and attainment at school level and, indeed, across the whole system. Moreover, it becomes possible to disaggregate the impacts of inclusion from those of other variables (such as gender and FSM entitlement) which we know to correlate with attainment.

The current study

This study sought to answer the following five questions:

1. Is there any evidence that inclusion has an effect on attainments at LEA level? In other words, is there evidence that more inclusive LEAs do better or worse than less inclusive ones in terms of pupil attainments?
2. Is there any evidence that inclusion has an effect on attainments at school level? In other words, is there evidence that more inclusive schools do better or worse than less inclusive ones in terms of pupil attainments?
3. Is there any evidence of differential effects of inclusion on the attainments of different pupil groups? In other words, do some groups of pupils do better or worse in inclusive schools than others?
4. Is there any evidence that inclusion has impacts on achievements over and beyond the attainments that are captured in national assessments? Do pupils in inclusive schools, for instance, have better social skills or higher self-esteem than pupils elsewhere?
5. Where highly inclusive schools also produce high levels of attainment amongst their pupils, how do they manage this? What strategies and forms of organisation do they use? How ‘inclusive’ (in the sense of equitable and participatory) are their practices? And how do these schools differ from equally inclusive schools which do not generate such high attainments?

These questions were addressed using a three-pronged approach

• a statistical analysis of the NPD data from key stages (KS) 1 – 4 that was obtained in 2002 ( questions 1-3);
• a series of case studies of highly inclusive schools producing differing levels of pupil attainment (questions 4 and 5); and
• a review of relevant national and international literature (all questions).

Methodology

The analysis used data from the NPD, which includes information on over 500,000 pupils in mainstream schools at each key stage, to create variables which might have an impact on pupils’ measured attainments. The variables included school inclusivity (the proportion of pupils at School Action Plus and with a statement) and LEA inclusivity (the proportion of pupils for whom the LEA is responsible who are placed in special schools).  

Multi-level modelling techniques were used to explore the effects of these and other variables on pupils’ average points scores in national assessments and examinations. These techniques made it possible to explore the effect of one variable taking into account all other variables and any hierarchies in the data.

The aim of the case studies was to explore the factors that might explain why some schools can be both highly inclusive and high-performing. A sample of 16 schools, equally divided between primary and secondary schools, was drawn for case study. All of these schools were chosen as being highly inclusive in terms of their SEN populations. Twelve of them (six primary and six secondary) were chosen as being high-performing in terms of the average measured attainments of their pupils. Four of them were chosen as being lower-performing. In each school, teachers, headteachers and other staff were interviewed and focused lesson observations and pupil interviews were undertaken. Questionnaires were administered where possible to staff and selected pupils.

Key findings

1. We found no evidence of a relationship between inclusion and attainment at LEA level. An LEA’s policy in terms of the proportion of pupils educated in mainstream schools seems to have no bearing on overall levels of attainment in schools in that LEA when other variables are taken into account.   There may be implications here for the likely impact on overall levels of attainment of the national policy of inclusion.

1. We found a very small and negative statistical relationship between the level of inclusivity in a school and the attainments of its pupils. The possibility that this is a causal relationship cannot entirely be ruled out, though this seems unlikely. There are four reasons for this:

• There is considerable variation in the performance of schools with similar levels of inclusivity, suggesting that school level factors are more important than levels of inclusivity per se.
• The small negative relationship between inclusion and school performance can be explained by the fact that, on our definition, schools with higher levels of inclusion tend to be schools serving more disadvantaged – and hence lower-attaining – populations.
• The case studies suggest that highly-inclusive schools tend to manage inclusion in broadly similar ways which seem likely to minimise any impact inclusion might have on attainment. Both higher and lower performing schools operate a similar model of provision.
• Although there are observable differences between highly-inclusive schools with different levels of performance, these are complicated by broad similarities between schools and considerable variation within schools. Moreover, they seem to relate to factors (resourcing issues, skill in responding to the achievement agenda, detailed implementation of the inclusion model) which are not directly attributable to the level of inclusivity in the school.

We have found some evidence that the small negative relationship between inclusivity and attainment affects pupils with and without SEN differentially and that the nature of this effect is different in primary and secondary schools .  The negative relationship is a little stronger for pupils without SEN than it is for pupils with SEN in the secondary phase. In key stage 1, the opposite is the case – that is, the relationship is stronger for pupils with than without SEN. In key stage 2, there is no clear relationship.   Moreover, the relationship is more marked for schools at lower levels of inclusivity than those at higher levels. Given these variations, it seems unlikely that the relationship is causal. It seems probable that other factors are at work, such as different SEN identification practices at different key stages and ceiling effects in terms of the numbers of pupils identified with SEN in schools with low overall levels of attainment.

1. We found some evidence (chiefly in the views of teachers and pupils) that inclusion can have positive effects on the wider achievements of all pupils, such as social skills and understanding. In particular, we found evidence that pupils with SEN can and do make good progress academically, personally and socially. On the other hand, we also found indications that having SEN might be a risk factor for isolation and for low self-esteem. While it did not necessarily entail such problems and while pupils without SEN could have similar experiences, pupils with SEN seemed more likely to experience difficulties in these areas.

1. Our findings are in line with the international research evidence. Most studies find few if any negative impacts of inclusion on the attainments and achievements of pupils without SEN.

1. The detail of how highly-inclusive and higher-performing schools manage provision is different in each case. However, there is a model which seems to depend on flexibility of grouping, customisation of provision to individual circumstances and careful individual monitoring, alongside population-wide strategies for raising attainment. As we suggested above, the model itself cannot explain differential performance, though it may be implemented with less sophistication and flexibility in some lower-performing schools.

The model is not predicated on a commitment to full participation by all pupils in common classrooms and shared learning experiences. To this extent it is not ‘fully inclusive’. Rather, it stems from a school-level commitment to ‘doing the best by all children’. In some cases this is expressed as an explicit commitment to inclusive principles; in others it is not. In either case, practice tends to be somewhat pragmatic, with opportunities for participation in the mainstream intermingled to varying degrees with individual, small group or segregated class provision.

The model is recognisable also in lower-performing schools. There are some indications that such schools fail to achieve the flexibility of provision which characterises their higher-performing counterparts. However, there are factors other than their management of inclusion which seem to explain their lower level of performance.

Some implications

• National inclusion and achievement policies

There is nothing in our findings to suggest that the current national commitment to maintain pupils in mainstream schools wherever possible is likely to have a significant impact on overall levels of attainment at national and LEA level, nor that it need have an impact at school level. Our findings suggest, in fact, that attainment is largely independent of levels of inclusivity in LEAs and schools. Other factors at the population level – such as socio-economic status, gender, ethnicity and mother tongue – seem to be much more significant and it is these factors which policy and practice might most profitably address if the aim is to make a significant impact on levels of attainment.

• LEA-level issues

There are clear indications from our data that levels of overall attainment in LEAs with higher levels of inclusivity are similar to those in LEAs with lower levels of inclusivity, once other variables are taken into account. In other words, there is nothing in our data to suggest that LEAs should not continue to pursue an inclusion agenda. However, our data can do no more than suggest how the process of change from less to more inclusive policies impacts on achievement and LEAs need to take care that their policies energise rather than destabilize.

• School-level issues

Although we can find no meaningful LEA or whole-population impacts of inclusivity on attainment, there is certainly considerable variation at school level. This suggests that there are school-level factors at work in determining attainment and achievement in schools with similar levels of inclusivity. It seems that many of these are largely unrelated to inclusion as such and that highly inclusive schools with different levels of performance nonetheless manage inclusion in broadly similar ways. However, it is possible that how a school manages inclusion and how inclusion interacts with particular school contexts might determine the strength of any ‘inclusion effect’. The implications of our findings for schools, therefore, would seem to be that, while they need not feel anxious about becoming more inclusive, they should certainly monitor the effects (both positive and negative) of such developments with care.

• The ecology of inclusion

In trying to understand the case study schools, we have found it useful to think in terms of an ‘ecology’ of inclusion. It is clear that teaching children with higher levels of SEN places considerable demands on schools and on teachers. Where schools have relatively high proportions of such pupils, there appears to be a delicate balance between the resources they can bring to bear on the task of teaching and the demands which the presence of these children create. It would seem that it does not take much to disturb it – a shortfall in classroom support, a weakness in teacher skills or managerial planning, a lack of funding and so on. There are particular issues around the presence of pupils whose behaviour disrupts lessons and, therefore, the learning of other children. In some schools, these children clearly constitute a major source of anxiety for teachers – and, to some extent, for other children. It is entirely possible that children like these have been in mainstream schools long before the current inclusion agenda was formulated. Nonetheless, those who work in and with schools should be aware of the ecological balances in schools and should seek to strengthen them.

• Caveats and future research

It is important to be clear about the limitations of this study. The NPD, like any large database, has its limitations and analyses of it have to make large numbers of assumptions. Likewise, the case studies reported here are limited in both quantity and depth. The weight of the evidence from the international research literature supports the conclusions reached by this study. Nonetheless, it would be a mistake to see the current study as definitive and it needs to be replicated and extended by future research.

Department for Education and Skills’ national database which includes information on the attainment of all state school pupils in the end of Key Stage tests with pupil characteristic variables.

home . about the conference . programme . registration . accommodation . contact