The broad background to this review is a long history of concepts of special pupils and special education and a faith in special pedagogical approaches. The rise of inclusive schools and some important critiques of special pedagogy (e.g. Hart, 1996; Norwich and Lewis, 2001; Thomas and Loxley, 2001) have raised the profile of teaching approaches that ordinary teachers can and do use to include children with special educational needs in mainstream classrooms. Inclusive education itself is increasingly conceived as being about the quality of learning and participation that goes on in inclusive schools rather than simplistic matters of where children are placed.
Policy and practice background
The policy of including pupils with Special Education Needs (SEN) in mainstream schools and classrooms in England and Wales was importantly marked by the Warnock Report (DES, 1978) and has since gained momentum with Codes of Practice (DfE, 1994; 2003), government guidance (DfEE, 1998; 1998) and legislation (1981, 1993 and 1996 Education Acts; SENDA). There is a statutory requirement on mainstream schools to provide ‘effective learning opportunities for all pupils’ by setting suitable learning challenges, responding to pupils’ diverse learning needs and overcoming potential barriers to learning and assessment for individuals and groups of learners.
Previous systematic literature reviews related to the area of special educational needs and inclusion have focused on behavioural concerns and behaviour management in schools (Harden, 2003); the impact of paid adult support on the participation and learning of pupils in mainstream schools, including pupils with SEN (Howes et al., 2003); and school-level approaches to facilitating the participation by all students in the cultures, curricula and communities of schools (Dyson et al., 2002). Non-systematic (in technical terms) literature reviews have addressed the question of whether there is a particular pedagogy for special educational needs or each type of SEN, particularly types of learning difficulty, but not related to mainstream contexts (Norwich and Lewis, 2001), or asked about approaches that can effectively include children in mainstream schools beyond classroom pedagogy (Sebba and Sachdev, 1997). While research has sought to establish the effectiveness of particular pedagogies or the impact of school actions on pupil participation there has been no prior systematic review that can answer the question of what pedagogical approaches can effectively include children with SEN in mainstream classrooms.
The overall aim of the three-year project is to utilize the expertise of the research team in researching the evidence base in relation to inclusive pedagogy. In year one the focus is effective pedagogical approaches in use in mainstream classrooms with children with special educational needs, aged 7–14 years. The main aim of this systematic review was to investigate what pedagogical approaches can effectively include children with special educational needs in mainstream classrooms.
Our review question was:
What pedagogical approaches can effectively include children with special educational needs in mainstream classrooms?
Our in-depth review focused on two related but more specific questions:
Does a pedagogy involving a peer group interactive approach effectively include
children with SEN in mainstream classrooms?
How do mainstream classroom teachers enhance the academic attainment and social inclusion of children with special educational needs through peer group interactions?
When the review question had been agreed a search was conducted. Search terms generated were aligned with the varying word usages in different countries and the British Education Thesaurus was used for selecting synonyms. All studies returned from searches were incorporated into Endnote bibliographic software enabling good compatibility with the EPPI-Centre systems.
The studies were screened employing specific inclusion criteria to identify studies with a specific scope (a focus on students aged 7–14 who experience special educational needs, in mainstream classrooms, including pedagogical approaches and an indication of student outcomes); study type (empirical); and time and place (written in English and published after 1994) (see Appendix 2.1). A range of electronic databases and citation indexes were interrogated (see Appendix 2.2) and internet sites were searched (see Appendix 2.3). Screening was applied first to titles and abstracts (in two iterative stages) and then to full documents. Screening was conducted by two independent screeners on all titles and studies, and the EPPI link-person on a sample for quality assurance. For pragmatic reasons document retrieval ended on 31st March 2004; any studies received after that time will be need to be included in any update.
The identified studies were taken through a series of graduated filters, culminating in the shortlist of studies. These were KEYWORDED using the EPPI-Centre (2003) Keywording Strategy (version 0.9.7) with review-specific keywords (see Appendix 2.4) in addition to EPPI keywords. This generated the ‘descriptive map’ of the studies in our review which provides a picture of the kinds of research that have been conducted together with details of their aims, methodologies, interventions, theoretical orientation, outcomes and so on. This process did NOT attempt to assess the quality of the studies.
Discussion among the full review team and the advisory group of the most useful cluster of studies identified in the systematic map led to the re-focusing of the study onto two specific questions for the in-depth review. The in-depth review thus focused on a group of pedagogical approaches characterised by peer group interactions that were conducted by mainstream teachers without necessitating additional staff resource and asked about their effectiveness and how teachers used the approaches. New inclusion and exclusion criteria were applied to the studies in the map leading to a subset of studies for the in-depth review. Data extraction (using EPPI-Centre guidelines) was undertaken on these by two independent reviewers and any differences discussed and resolved.
The quality of studies and weight of evidence (WoE) were assessed using the EPPI data extraction framework to assess the reliability and quality of each study and focus judgements about the trustworthiness of study results and the weight of evidence that the study could contribute to answering the in-depth review questions. Judgements about the relative weight of evidence of each study were made using the following explicit criteria: the soundness of studies (internal methodological coherence); the appropriateness of the research design and analysis in relation to the review questions, and the relevance of the study topic focus to the review questions. Taking into account quality of execution, appropriateness of design and relevance of focus, an overall weight of evidence judgement was made using a consistent formula (see Chapter 2). As quality assurance, each study was independently reviewed and data extracted by two different members of the review team or a member of the review team and the EPPI-centre link person.
The findings of the individual in-depth studies were synthesised and conclusions and recommendations drawn. Synthesis took the form of eliciting a qualitative and quantitative overview for the effectiveness question and a structured narrative describing any overall, cross-study patterns/themes related to how teachers use peer group interactive approaches.
A total of 2,095 potentially relevant reports were identified for the current review. Over half (1,156) were excluded in the first screening of titles and abstracts and a further 238 were excluded in a second iteration of the process (see Table 3.1). A total of 450 were sent for and 14% of these (64) were not received within the timeframe of the review or were unavailable. A total of 383 full reports were screened resulting in the exclusion of a further 315 reports, leaving 68 that met the criteria for inclusion in the mapping study.
In the application of exclusion/inclusion criteria to the collection of titles and abstracts the measure of inter-rater reliability between the two members of the review team was good (Cohen’s Kappa 0.62). Nonetheless, for rigour, all rather than a sample of the titles and abstracts were double screened. There was 80% agreement for the two reviewers across the set of titles and abstracts. The kappa statistic for inter-rater reliability between each of the review members and our EPPI colleague was lower but fair (Cohen’s Kappa 0.35). This difference is most plausibly explained by the difference in expert knowledge of the subject matter between the review team members and our EPPI colleague.
Most of the 68 studies in the map were identified through the electronic searches on PsychInfo and ERIC. Most of the studies were researcher-manipulated evaluations and most were undertaken in the USA. The majority did not focus on curricular issues, but of those that did literacy dominated. Primary school contexts were twice as prevalent as secondary school contexts. The target groups were mostly mixed sex pupils with learning difficulties. Regular mainstream teachers mostly carried out the teaching interventions, with special teachers and peers also often involved. The most common pedagogical approach was adaptation of instruction, often combined with other types of adaptation: materials, classroom environment and assessment. Just under a quarter of the studies involved peer group interactive approaches.
Ten studies were included in the in-depth review and nine of these were conducted in the USA. With the exception of two exploration of relationships studies, the studies were evaluations, mostly researcher-manipulated (6). Six of the studies focused on literacy and six were conducted in primary school settings.
Five studies were included in the synthesis for review question (a) and seven for review question (b). Question (a) studies researched cooperative learning, guided inquiry and Circle of Friends approaches. Outcomes measured included engagement in classroom activities, curriculum performance and social interactions with peers/social acceptance. Effect sizes were reported in four studies, ranging from small to fairly large. Synthesis of question (b) studies led to five substantive themes emerging: the model of student-as-learner; integration of academic and social considerations; organisational and organised support; holistic views of 'basic skills' and shared philosophy.
Several studies were deemed to be medium or medium-high in terms of weight of evidence, but an issue remains about the scale of evidence available to address the research questions. Good quality studies, which incorporated empirical validations of effectiveness, were, unsurprisingly, based on small samples. There were no studies with high weight of evidence for question (a) and only one for question (b). The strength of confidence we can have in the evidence is therefore measured. Despite these limitations, the review did lead to some substantive findings and offered a basis on which to make some recommendations for practice. The likely effectiveness of peer group interactive approaches for inclusion of children with special educational needs in mainstream classrooms can be established and we have an (albeit small) evidence base on how teachers use these approaches, that is some qualitative understanding of the processes at work.
There is a small accumulation of evidence about the effectiveness of cooperative learning particularly in relation to the curriculum area of literacy. Cooperative learning encompasses a range of teaching practices and the evidence base relates to the elements of social grouping/teamwork, revising and adapting the curriculum and working with a cooperative learning school ethos. Specific evidence is available for the effectiveness of two specific cooperative learning programmes. Evidence of effectiveness also relates to programmes associated with other related types of peer group interactive approach, Guided Inquiry and Circle of Friends.
All of the studies showed evidence of some learning and, with the exception of the Circle of Friends approach, this has included learning in the academic domain. Three studies provide explicit evidence of impact on both the academic learning and community participation of pupils with special educational needs. A further study provides evidence of academic rather than social gains. The evidence also indicates improved attitudes toward curriculum areas and children's own views of their competence, acceptance and self-worth. The evidence indicates that peer group interactive approaches that are effective in academic terms are also often effective in terms of social participation and children's attitudes to their learning. Teasing out the elements of the approaches that are functionally related with each outcome is difficult and perhaps unnecessary in professional rather than research terms.
The model of student as learner and having active agency in the construction of personal knowledge underpinned the studies and the interventions. Teachers fostered the co-construction of knowledge through scaffolding by, and dialogue with, peers. The studies recognised that a sense of belonging to and participation in the learning community has an important effect of young people’s learning in schools. Teachers made use of organisational support for community participation and organised support for peer group interactive approaches using peers and adults together with careful planning. An holistic approach to skill development underpinned many of the interventions in contrast to the isolated skill development associated with traditional remedial programmes for special needs. Making use of peers may bring with it a necessity to make skill development socially meaningful. Finally, the studies indicate a role for shared philosophy and common concern with participation in the learning community, cooperation and collaboration.
Strengths and limitations
This systematic literature review had both strengths and limitations. It was strong in asking relevant questions of use to teachers where limited resource is an issue. It encompasses studies of pupils representing a wide range of SEN. And there was high quality assurance for the review: screening, data extraction and quality assessment was conducted by two independent review team members (or a review team member and EPPI link-person) at each stage. Confidence in the review findings is strengthened by the quality of the studies and the rigorous check on quality that were applied.
The literature review was limited in scope to material from 1994 and excluded pupils in the early years or post-14. The in-depth review was dominated by studies in primary contexts, meaning that as we move from the up through the school system to age 14, then our degree of confidence about the evidence for peer group interactive approaches drops considerably. Scope was also limited by the studies that did not arrive in time to be scrutinised in full. These tended to be unpublished theses (see Appendix 3.2) and therefore may be systematically different from the studies included in the map and in-depth review adding to the possibility of some distortion from publication bias. Negative or null outcomes are less likely to be published and which means the picture emerging from the systematic review may be over-optimistic.
The review is also limited in the strength of the evidence base arising from this systematic review. The lack of randomised control trials means that evidence of effectiveness is not as strong as it could be. The number of studies in the synthesis is small and the numbers in the samples for these are also small. While we know enough about the pupils with special needs who participated in the studies to begin to judge generalisability, we know less about the teachers themselves and how representative they may be. We also know that the contexts for the studies are likely to differ from the cont