Teaching for inclusion – a review of research on the cooperation between regular teachers and special educators in the work with students in need of special support

ABSTRACT This article presents a review of qualitative research on interprofessional cooperation between regular teachers and special educators published from 2005 to 2019. The aim of the review was to gain knowledge about how different forms of cooperation take shape and about factors at multiple levels that facilitate or constrain cooperation as a means of achieving inclusion. In total, 25 studies were selected. The results are discussed in relation to Thomas Skrtic’s theory of bureaucracies within the school organisation in order to compare and analyse different forms of interprofessional cooperation and schools’ organisations of special educational work. Cooperative teaching, special educational consultations and mixed forms of cooperation were found to entail different benefits and challenges related to communication and the cooperating actors’ roles. Facilitating factors included personal chemistry, an equal distribution of power and responsibilities and support from the school management through provision of professional development and adequate planning time. In several studies, a flexible cooperation was argued to be hindered by curricular constraints and standardised testing. Education policy is therefore emphasised in this review as important for understanding the conditions under which school staff are responsible for inclusion.


Introduction
Inclusive education seems to be difficult both to define and to achieve. Not least, it is unclear how teachers and special needs staff should cooperate in order to create more inclusive classrooms. In a critical research review, Göransson and Nilholm (2014) identified empirical shortcomings in research about inclusive education, arguing that there is a lack of studies about how more inclusive environments are to be constructed in schools if we understand inclusion as involving all pupils. Thus it is of extreme importance to systematically review the knowledge pertaining to this issue in order to be able to create more inclusive learning environments and to understand obstacles to such a development. The purpose of this review is to further our knowledge about opportunities and barriers to the development towards a more inclusive classroom by reviewing research that examines different aspects of the cooperation between regular teachers and special educators. Several influential scholars within the field of special education, such as Ainscow (1998) and Skrtic (1991), have clearly emphasised interprofessional cooperation, joint decision-making and the involvement of all school personnel as organisational prerequisites for success in a school's transformation towards becoming truly inclusive. Moreover, there are reasons to believe that factors at multiple levels might affect the work of schools and cooperating professionals in this area (Skrtic 1991).
In parallel with the influence of the Salamanca Statement in 1994, when the idea of inclusive education had its international breakthrough, neoliberal ideas of efficiency and accountability in education have grown stronger. Globally, standardised tests such as the OECD's Programme for International Student Assessment (PISA) have been claimed to play a vital role in narrowing the goals of education in countries competing over placements on ranking lists (Kamens 2013;Niemann, Martens, and Teltemann 2017). Locally, the emergence of New Public Management has entailed an increased focus on school competition, customer service, and teachers' individual performances (Evetts 2011;Ball 2016). This development has been criticised by scholars who claim that teachers are pushed towards standardisation and 'teaching for the test' at the expense of their instructional flexibility (e.g. Au 2007;Hopmann 2007) or towards neglecting the work with children in need of special support, since the rewards in terms of improved student performance might be limited in this area (Ball 2003). Given the importance of policy to the daily work of school staff, we strive to incorporate empirical studies of teachers and special educators' cooperation in a wider analysis where multiple levels are taken into account. Idol (2006) approaches interprofessional cooperation in the field of special education as different forms of special education service delivery. In her discussion, Idol distinguishes four different kinds of support that are used in schools to provide assistance to regular teachers in their teaching of students in need of special support. These are:

Previous research
(1) Cooperative teaching, where special education-and regular teachers collaborate and teach together within the classroom. (2) Consulting teaching, where a special educator helps the regular teacher with planning, assessment, developing material, and adapting instruction, rather than working directly with students. (3) Supportive resource programs, where special education-and regular teacher collaborate in designing students' individualised instructional programs for the resource room. (4) Instructional assistants, where paraprofessional aides accompany students in need of special support in the regular classroom.
In this paper, the main focus in on cooperation between regular teachers and special educators as a means to achieve inclusive classrooms. Thus, the first two cooperation models, here referred to as co-teaching and special educational consultations will be further described below through a summary of previous reviews and selected previous research that have not been included in the sample of literature analysed in this review.

Co-teaching
In Scruggs, Mastropieri, and McDuffie (2007) review of 32 qualitative studies on coteaching, five models are described, drawing on the work of Lynne Cook and Marilyn Friend (e.g. Cook and Friend 1995;Friend et al. 2010): (1) One teach, one assist, where one teacher takes the leading role in the classroom, and the other teacher observes students or assists them while circulating around the room.
(2) Station teaching, where the two teachers divide the instructional content between them and provide instruction at different learning stations. (3) Parallel teaching, where the two teachers divide the class into two heterogeneous groups and teach the same content in different parts of the classroom. (4) Alternative teaching, where one teacher temporarily pulls a small group to the side for specialised instruction. (5) Team teaching, where the two teachers share instructional responsibilities equally, for example in the form of role-play, modelling or turn-taking in instructional delivery.
Among the studies reviewed by Scruggs, Mastropieri, and McDuffie (2007), the 'one teach, one assist' approach was by far the most common. In these cases, regular teachers generally took the greatest responsibility for whole-class instruction, while special education teachers had a subordinate role, giving individual support or managing classroom behavioura type of role distribution that Scruggs et al. argue is hardly a sign of true and innovative collaboration. In their review, a number of factors were highlighted as important for establishing successful co-teaching. These included voluntary participation, administrative support through the provision of adequate planning time and training, and the compatibility of the co-teachers on both professional and personal levels. Despite the unbalanced distribution of power in the different cases of cooperation, the overall result of the reviewed studies showed that regular teachers and special education teachers perceived co-teaching as beneficial for all actors involved. Students were believed to receive more attention and support, and teachers were considered to gain opportunities for mutual learning. However, as Friend et al. (2010) conclude, most research on co-teaching has focused on co-teachers' roles and relationships or programme logistics, and there is little evidence found in research that co-teaching contributes to more inclusive classrooms or increased academic performance among students.

Special educational consultations
Different models of cooperation have also been identified in research focusing on special educational consultations. Sundqvist (2018/2019) points out that research in this field often has discussed consultations in terms of being either expert-driven and clientcentred or process-driven and consultee centred. In other words, the question is whether special educators should function as experts giving advice to regular teachers on how to teach students in need of special support, or if they should be discussion partners who help regular teachers reflect upon their practice in order to move in a more inclusive direction.
In Sundqvist's dissertation (2012), she discusses three categories of consultative meetings: In counselling conversations, the special education teacher gives instrumental advice related to specific students based on special educational knowledge. In reflective conversations, the special education teacher tries to make the regular teacher engage in self-reflection and thereby shift the focus from the student to the teacher. Cooperative conversations are characterised by professional exchange. The special education teacher contributes knowledge of students with disabilities or learning difficulties, while the regular teacher shares knowledge about subject matter or the group of students. Reflection takes place also in these kinds of conversations, but both teachers are reflecting together in dialogue. Similar kinds of collaborative consultations have also been described by, for example, Cook and Friend (2010), and has often been regarded as the foundation for co-teaching (Idol 2006). As in the case of co-teaching, previous research has identified different forms of cooperation, but we have not been able to find adequate research investigating the outcome of special educational consultations with regard to inclusive education.

Theoretical points of departure
Inspired by Max Weber, Skrtic constructs ideal types for theorising about two different kinds of bureaucracies within the school organisation, one inside of the other. The machine bureaucracy involves standardisation of work processes and direct control of labour by managers, an organisational idea that spread from the industrial sector to social organisations in the beginning of the twentieth century. According to Skrtic, managing schools like machines reduces teachers' discretion and flexibility, which they need in order to personalise instruction to meet the needs of all their students. In contrast to the machine bureaucracy, Skrtic considers schools, in line with other scholars (e.g. Weick 1976), to be typical examples of loosely coupled systems, where the complex nature of the work that takes place is not easily bureaucratised. Thus Skrtic (1991) argues, schools are informally organised as professional bureaucracies on the inside, based on the professional and normative values of teachers. In the professional bureaucracy, standardisation is created and maintained through teachers' standard programmes, rooted in institutional and cultural norms.
Since neither machine bureaucracy nor professional bureaucracy is especially adaptable, Skrtic proposes adhocracy as a third way. This approach is based on innovation and problem solving and can be found in very dynamic and uncertain environments. Workers in an organisation structured as an adhocracy need to cooperate, communicate, and be open towards re-evaluating and reconsidering basic theories and values strongly connected to their own professional roles. In the case of special education, it is not possible to accomplish this by handing the power of decision-making to either teachers or special educators. Rather, Skrtic claims, adhocracy is only achievable when power is handed to groups of professionals in the form of interdisciplinary teams who cooperate based on their expertise.
The ideal types developed by Skrtic make it possible to compare and analyse schools' organisations of special educational work. In the present review, the ideal types will be used to discuss the distribution of roles and opportunities for interprofessional sharing of expertise in the empirical studies and how the forms of cooperation and their consequences relate to organisational and policy-related factors. It should also be pointed out that we share the pragmatic view of Skrtic (1991) that educational research should strive to help schools develop more inclusive and democratic practices.

Aim and research questions
The overarching aim of this review is to gain a deeper understanding of the cooperation between regular teachers and special educators in schools' work with students in need of special support in order to analyse beneficial factors and hindrances in the development towards more inclusive classrooms. In this effort, we want to examine and compare different types of cooperation and the different benefits and challenges they bring to the classroom. Further, we are interested in how factors at multiple levels facilitate or constrain such cooperation as a way of achieving inclusion. The aim leads to the following research questions: (1) How does the cooperative work of regular teachers and special educators take shape within different forms of cooperation? (2) How do benefits and challenges in the cooperation between regular teachers and special educators relate to facilitating and constraining factors at multiple levels?

Materials and methods
In order to gain a rich understanding of how cooperative work related to the complex act of teaching can take shape, we have focused on qualitative research and especially on observational studies. When exploring the actual work done, we considered observations to be a crucial component, since merely asking teachers about their practices might lead to descriptions of the practices they consider the most appropriate (Hofer 2002). Since many observation studies combine different types of qualitative data, interviews have also been an important source of data in this review.

Literature search
The literature was collected by searching the databases ERIC, Scopus, Web of Science, Google Scholar and Swepub. Searches were made in the abstracts of articles and dissertations, and the search words used were teacher* AND (observ* OR ethnograph*) AND (inclusi* OR mainstreaming OR special education*) AND (collaborat* OR cooperat* OR consult* OR supervis* OR interact* OR role*). Only English-language sources of literature that were available online and published in 2005-2019 were included in the search results. Predetermined selection criteria were applied in order to assess the relevance of the collected studies. To begin with, the studies had to address the teaching of students in need of special support and include issues of the joint work of regular teachers and special educators. As described above, the studies were also required to use observational data, at least partially. Finally, we only included studies conducted in mainstream schools in grades 1-12. Studies focusing on special schools, preschools, or higher education were thus eliminated. The first selection was made by eliminating titles that were obviously irrelevant. The relevance of the remaining studies was then assessed by reading their abstracts. After reading more carefully we dropped even more studies because they were not considered to meet all criteria. In total, 25 titles were selected: 19 journal articles and 6 dissertations.

Sample characteristics
In order to aid the identification of patterns among the selected studies, a table was created where they were coded with regard to the form of cooperation investigated as well as the topic, geographical region, sample, methods and theories used, and major findings (Arai et al. 2007). (The table can be obtained from the main author upon request.) As many as 18 of the 25 studies originated in the USA, and no studies were conducted outside of the USA and Europe. Another striking feature was the many studies examining co-teaching. Common topics were the cooperating teachers' roles and the implementation of inclusion programmes in schools. Two of the selected studies investigated consultations between regular teachers and occupational groups outside of the pedagogical sphere. Although these two studies did not meet all the selection criteria, they were included due to the lack of studies found that examined consultations between regular teachers and special educators.

Research synthesis
This review applies a narrative method for synthesising research, where information from many different studies is interpreted and organised in themes, and major agreements and disagreements are discussed in order to draw new conclusions (Green, Johnson, and Adams 2006). Following the creation of the table, a closer reading of each study guided an extraction of textual units, which were sorted into themes based on our two research questions. Thus, all studies were themed twice. This process resulted in the construction of five themes: 1. cooperative teaching; 2. special educational consultations; 3. mixed forms of cooperation; 4. organisational prerequisites; and 5. standardisation and curricular constraints.

Results
In this section, the five themes will be presented as subthemes relating to the research questions, as follows: (1) How does the cooperative work of regular teachers and special educators take shape within different forms of cooperation? (themes 1, 2 and 3) and (2) How do benefits and challenges in the cooperation between regular teachers and special educators relate to facilitating and constraining factors at multiple levels? (themes 4 and 5).

Forms of cooperation
Cooperative teaching Seventeen of the 25 studies selected focused on various aspects of co-teaching. Although it was argued that this form of instructional arrangement promotes inclusion compared with the provision of support to students in a specific resource room (Buli-Holmberg and Jeyaprathaban 2016; Sanahuja-Gavaldà, Olmos-Rueda, and Morón-Velasco 2016), many studies found that co-teachers did not seem to take advantage of the potential of coteaching. Instead they tended to rely on models that required fewer instructional modifications, such as parallel teaching or one teach, one assist (e.g. Ashton 2010; Brendle, Lock, and Piazza 2017;Casale-Giannola 2012;Klein 2009;Shankland 2011;Strogilos and Avramidis 2016;Volonino 2009;Wexler et al. 2018). The frequent use of the one teach, one assist model in co-teaching was often discussed in terms of an unequal distribution of responsibilities and authority. For example, instructional decisions in the co-taught classroom investigated by Ashton (2010) were generally made by the regular teacher in accordance with general education standards, and the instruction of students with IEPs took the shape of a side activity rather than being a part of individualised teaching. In Klein's (2009) case study on the classroom implementation of an inclusion programme, the regular teachers' views and practices were more or less adopted by all coteacher pairs. These power relations seemed to be established early on in the co-teaching partnership and were then difficult to change. Some researchers reported about practices in which the co-teachers' roles were harder to categorise. The three co-teaching pairs in Terranova's (2010) study tended to share the responsibilities in the classroom rather equally, although the regular teachers were often the most dominant in the cooperative planning process. In the study by King-Sears et al. (2014), the co-teachers did use different models, and their students believed both teachers to be equally responsible for instruction, although the regular teacher was found to present new content three times as often as the special educator did and also to interact twice as often with the large group of students.
Co-teacher respondents in some of the studies (Lava 2012;Leatherman 2009; van Hover, Hicks, and Sayeski 2012) highlighted the professional relationship and relational factors such as chemistry and matching teaching styles as important. Vadala (2014) found great variation in the depth of cooperation between regular teachers and special educators and argues that co-teaching can evolve over time in parallel with the forming of a professional relationship. Such progress is described by Lava (2012), who studied the development of a co-teaching partnership over time and found that, as the co-teachers' relationship deepened, they started using a wider variety of co-teaching models adapted to the student group and the educational content. Another example is van Hover, Hicks, and Sayeski's (2012) study, where the co-teachers avoided the one teach, one assist model by clarifying their roles and structuring the instructional activities together in such a way that they were equally responsible for instructional delivery, although they were in charge of different lesson segments based on their areas of expertise. In their study on inclusion programmes in schools, Smith and Leonard (2005) show how unclear roles among co-teachers can be reflected in conflicting attitudes and different opinions of who should do what. The regular teachers in the study did not perceive themselves to have the main responsibility for the work with students in need of special support and declared that the inclusion programmes increased their workload. The special educators, on the other hand, described the inclusion programmes as successful but expressed that regular teachers often expected them to perform the tasks of an assistant.
Positive and negative consequences of co-teaching were also discussed in some of the studies. Teacher respondents in Leatherman's (2009) study stressed the value of having teachers with different types of expertise available in the classroom and perceived themselves to benefit from observing how special educators modify instruction. In Buli-Holmberg and Jeyaprathaban's (2016) study, it was reported that varied and flexible cooperation between regular teachers and special educators within the classroom created better conditions for interaction between all involved actors. Strogilos and Avramidis (2016) discovered that students in need of special support in their study were more frequently 'on task' in co-taught as opposed to non-co-taught classes. However, the cotaught students in need of special support did also display a lower degree of interaction with their peers, which Strogilos and Avramidis suggest was a consequence of an overuse of one-to-one instruction from special education teachers in co-taught settings.

Special educational consultations
Four studies described and analysed special educational consultations. Different models of consultations were reported that could be classified, using Sundqvist's (2012) categorisation, as both reflective conversations (Kjaer and Dannesboe 2019) and cooperative conversations (Pettersson and Ström 2017). Kjaer and Dannesboe (2019) describe consultations between educational psychologists and regular teachers as having a therapeutic character, where teachers were supposed to search within themselves for alternative instructional approaches, attitudes and behaviours. Kjaer and Dannesboe (2019) argue that consultations of this kind do not take advantage of the educational psychologists' expertise and transform teachers into coached employees rather than autonomous professionals. In the consultations analysed by Pettersson and Ström (2017), there was also an emphasis on the learning environment instead of individual students' deficits. However, Pettersson and Ström (2017) describe an even distribution of power between regular teachers and special educators in these consultations, which they noted were marked by professional dialogue and problem solving through mutual sharing of expertise. The special educators were supporting teachers in their everyday challenges, but the teachers were active agents in these meetings, rather than passive receivers.
The two other studies on special educational consultations both discussed challenges related to insufficient communication. Hemmingsson, Gustavsson, and Townsend (2007) examine teachers' cooperation with therapists working at habilitation centres regarding participatory arrangements for students with physical disabilities. The results showed that the cooperating actors held different views on the purpose of their cooperation. The teachers emphasised group belonging and participation in collective learning activities, whereas the therapists focused on independence and access to all sorts of activities. Hemmingsson, Gustavsson, and Townsend (2007) argue that institutional barriers and lack of communication hindered the cooperating actors' ability to understand each other's views. On the contrary, tensions emerged, as the teachers perceived the therapists as 'controllers', while the therapists considered themselves 'pushers' struggling with teachers who did not prioritise this area of their work. The interprofessional cooperation in Bray and Russell's (2018) study on regular teachers' implementation of IEPs written and monitored by special educators can also be considered an advisory form of special educational consultations, although in written form. The special educators in the study experienced strong institutional demands to connect the IEPs tightly to general curriculum standards. This resulted in texts with limited guidance for regular teachers on how to attend to individual students' needs and to mere surface adaptions in the instructional activities.

Mixed forms of cooperation
Four other studies examined mixed forms of cooperation between regular teachers and special educators. Strogilos (2012) investigated a school's implementation of an inclusion programme that consisted of professional development for teachers, the forming of a local inclusion team, and interprofessional cooperation through co-teaching and monthly multidisciplinary meetings. Vernon-Dotson (2008) studied three schools that put together teacher leadership teamsconsisting of regular teachers, special educators and administratorswhich organised meetings and workshops in order to support regular teachers in their efforts to differentiate instruction. In one of the schools, coteaching was also introduced. Olson, Leko, and Roberts (2016) examined how a school that won a prestigious inclusive education award provided access to the general curriculum for all students through the use of a variety of forms of interprofessional cooperation. Each grade in the school had two teams consisting of regular teachers and a learning strategist, which were supposed to share ideas and solve problems together. Other forms of cooperation included joint planning, team teaching, and IEP teams that worked together with teachers to develop curricular goals for students. Another form of mixed cooperation was described by Eisenman et al. (2011), who investigated a collaborative consultation model at a high school where two special educators mixed teacher consultations with temporary co-teaching and direct work with students through coaching and additional teaching after school.
Several positive outcomes of these mixed forms of cooperation were reported. Eisenman et al. (2011) claim that the model in their study enhanced flexibility for special educators and supported an equal professional relationship between them and the regular teachers. In all studies, it was argued that the cooperation model in focus increased the number of students in need of support who were physically placed in the general classroom. Moreover, all studies described increasingly positive attitudes towards inclusion among teachers, towards shared responsibilities for students, and towards school cultures that value cooperation. Strogilos (2012) suggests that the combination of support inside the general classroom with interprofessional cooperation through discussions and consultations both challenged the teachers' assumptions about special needs and strengthened their beliefs that inclusion is possible. In some cases, benefits of the models in terms of student progress were described by respondents in the studies. For example, the special educators in Eisenman et al. (2011) study suggested that the instructional arrangement was an important reason that more students in need of support were accepted to their top choices for further education.

Organisational prerequisites
Many studies stressed different organisational factors as crucial for successful interprofessional cooperation, and the importance of common planning time was one of the factors most frequently mentioned, especially in the studies examining co-teaching (e.g. Leatherman 2009;Olson, Leko, and Roberts 2016;Shankland 2011;Smith and Leonard 2005;Terranova 2010;Strogilos 2012;Vadala 2014;Vernon-Dotson 2008). In the study by Olson, Leko, and Roberts (2016), for example, teachers and special educators got time during weekly meetings to plan lessons together, to discuss student matters, and to share ideas, which the researchers argue led to a sense of shared responsibility for all students at the school.
Another organisational prerequisite highlighted by several researchers was provision of professional development (e.g. Brendle, Lock, and Piazza 2017;Hemmingsson, Gustavsson, and Townsend 2007;Klein 2009;Olson, Leko, and Roberts 2016;Shankland 2011;Smith and Leonard 2005;Strogilos and Avramidis 2016;Terranova 2010;Vadala 2014;Vernon-Dotson 2008). Brendle, Lock, and Piazza (2017) argue that insufficient professional development for the co-teachers in their study limited the benefits of being provided adequate time for cooperative planning. Hemmingsson, Gustavsson, and Townsend (2007) suggest that a lack of training was one of the main reasons for the tensions and conflicting views between teachers and therapists in their study. Smith and Leonard (2005) call attention to the role of the principal in their study of challenges in the implementation of school inclusion programmes. The principal whom they considered to have the greatest success in changing school culture in a more inclusive direction provided opportunities for professional development and involved regular teachers and special educators in decision-making and cooperative problem solving. In contrast, the principals in the study by Klein (2009) did not consider themselves knowledgeable enough to support the co-teachers, which meant that the responsibility for change ended up in the hands of the two cooperating teachers in the classroom. Thus Klein (2009) stresses that principals also need professional development on inclusion in order for them to be able to lead and support teachers.

Standardisation and curricular constraints
A couple of studies discuss how the educational content might relate to the degree of inclusion in teaching. From investigating strengths and weaknesses with regard to inclusion in academic and vocational classrooms, Casale-Giannola (2012) argues that active learning experiences, strong connections to the outside world, and a shared interest in a specific career are some of the elements that facilitate the building of an inclusive community in vocational settings. Sanahuja-Gavaldà, Olmos-Rueda, and Morón-Velasco (2016) discovered low degrees of inclusion in academic environments in their study on four schools' organisation of support to students with autism spectrum disorders. By using the index for inclusion developed by Booth and Ainscow, Sanahuja-Gavaldà, Olmos-Rueda, and Morón-Velasco (2016) found that students in need of support were included to a higher degree in physical and artistic education than in mathematics or language classes and that the primary schools in the study were more inclusive than the secondary school.
A possible explanation for these differences might be found in the many studies discussing the constraints placed on teachers from the policy level, which may be more evident in academic education. In Ashton's (2010) study, the co-teachers were limited to acting within a framework clearly affected by school policy reforms such as the No Child Left Behind Act, which put pressure on teachers to deliver good student results on standardised tests. Shankland (2011) found that although the co-teachers in her study were willing to make adjustments in order to meet the needs of all students, they were constrained by a tightly packed curriculum, which reduced their scope for testing alternative teaching methods and approaches. Also, Strogilos (2012) and Bray and Russell (2018) describe how a standardised curriculum and centrally determined guidelines and rules can make teaching difficult to change in a more inclusive direction. van Hover, Hicks, and Sayeski (2012) found that the two co-teachers in their study had a strong focus on very specific content areas, and strategies to memorise them, in order to prepare their students for standardised tests. This approach did not leave much room for other competences, such as critical thinking and creativity. However, while the regular teacher was almost totally committed to this task, the special educator was more ambivalent and underlined the importance of also giving the students tools necessary for handling different situations in life.

Discussion
To sum up, the reviewed body of research focused on different forms of cooperation, associated to different benefits and challenges. Overall, the many studies that examined cooperative teaching stressed several benefits of this form of instructional arrangement. For example, it was argued that co-teaching can foster the development of deep professional relationships (e.g. Lava 2012) and can help keep students 'on task' (Strogilos and Avramidis 2016) and that regular teachers and special educators can learn from one another through observations in the instructional setting (Leatherman 2009). However, a large share of the co-teaching studies found that the regular teacher was often responsible for instruction in the classroom, while the special educator assumed the role of an assistant. This was thought to lead to negative consequences in the cotaught classrooms, for example in the shape of reduced interaction with peers among students in need of support (Strogilos and Avramidis 2016). The extensive use of the one teach, one assist model has also been repeatedly found in previous research and is discussed in the previous research review on co-teaching by Scruggs, Mastropieri, and McDuffie (2007). Following Skrtic, such a distribution of roles could be the sign of a strong professional bureaucracy in which adaptions for students with special educational needs are limited. In order to be an example of adhocracy as advocated by Skrtic (1991), the cooperation between regular teachers and special educators would need to take the professional knowledge of both cooperating actors into account and simultaneously challenge the cultural and institutional norms that facilitate professional bureaucracies.
The relatively few studies that examined special educational consultations offered some examples of cooperation in which power and responsibilities were equally shared (e.g. Eisenman et al. 2011;Pettersson and Ström 2017). Challenges found in studies focusing on interprofessional cooperation with a consultative character, on the other hand, were associated, for instance, with a lack of communication between the cooperating actors (Bray and Russell 2018;Hemmingsson, Gustavsson, and Townsend 2007). The mixtures of co-teaching and different forms of consultations and professional development activities that were described in some of the reviewed studies tended to be described by the researchers as fruitful and as having the potential to change attitudes, school cultures, or classroom practices in a more inclusive direction. As in the studies of co-teaching and consultations, this seemed to be more evident in schools where the cooperation was based on local autonomy, cooperative problem solving and interprofessional sharing of ideas. Such a kind of cooperation is in line with Skrtic's descriptions of adhocracy as resting heavily on trust, reliance upon the expertise of different professions, and an openness to change and re-evaluation of roles, values, and methods. However, we want to emphasise that the success stories of interprofessional cooperation that were described in the reviewed studies were related to various views of what desirable results consisted in. In some cases, success was understood in terms of the number of students who were placed in regular classrooms or given access to the general curriculum. Other studies based their narratives about successful cooperation or schools becoming more inclusive on the perceptions of respondents. Actual descriptions of observed, extensive re-evaluations and adaptions of pedagogical methods were rare.
The benefits and challenges associated to the different forms of cooperation were related to factors on multiple levels. Some studies stressed chemistry, matching teaching styles, and other individual factors as important for the success of co-teachers. Even more studies emphasised the organisational conditions for cooperation between teachers and special educators and thereby also the role of the principal as the leader and main decisionmaker in the school organisation. The most often mentioned organisational prerequisites for successful cooperation were provision of professional development and a sufficient amount of common planning time. Similar findings have been presented in the previous research review on co-teaching by Scruggs, Mastropieri, and McDuffie (2007).
Many of the reviewed studies discussed education policy as a constraining factor for the cooperating actors in their effort to achieve more inclusive classrooms. Regular teachers and special educators in the studies felt pressure to cover a large amount of curricular content and prepare their students for standardised tests, which hindered their flexibility and their opportunities to adapt instruction and use alternative methods. In some cases, the conflict of interest that emerged between inclusion and demands for high student performance was seen to affect special educators and regular teachers in different ways (e.g. van Hover, Hicks, and Sayeski 2012) or even to reinforce the role of the regular teacher in co-taught classrooms as responsible for whole-class instruction, while the teaching of students in need of support became a side activity (Ashton 2010).
The pressure from education policy experienced by teachers and special educators in different national contexts that is highlighted in this review can be understood in relation to the global spread of audit cultures and educational standardisation along with the growing significance of PISA and other international standardised tests (Sellar and Lingard 2013). Conflicting policy demands for inclusion on the one hand and competition and performativity on the other have been previously discussed in research that often emphasise the dominance of the latter (e.g. Barton and Slee 1999). Magnússon (2019) points out that even the Salamanca statement allows for several different interpretations of the meaning of inclusion, which range from narrow definitions focused on placement of pupils to wider ideals of creating communities. We mean that a highly standardised and competitive education system encourages an interpretation of inclusion focused on individuals' access to the general curriculum and classrooms rather than community, adaptions and meaningful participation.
By once again returning to Skrtic (1991), the constraints on interprofessional cooperation inflicted by education policy would be a sign of a strong machine bureaucracy that shape the work of professionals in schools. Based on our analysis, we argue that adhocracy is a possible way forward to achieve more inclusive classrooms, but that professional, and above all, machine bureaucratic structures hinder professionals' opportunities for flexible cooperation. In the light of our discussion, the question of how much responsibility for inclusion can be placed on teachers and special educators and on principals is a legitimate one. The influence of education policy should not be underestimated, and from our perspective, future studies that investigate this link in the chain of responsibility for inclusion would be relevant.