Researching on inclusive education in Spain: how does action research question the roles adopted by researchers in knowledge mobilization processes?

ABSTRACT Addressing the bridge between theory and practice in the field of inclusive education in Spain involves searching for strategies of knowledge mobilisation, and action research (AR) is one of them. This study explores how its practical development questions the roles that researchers adopt throughout the process. A content analysis of several focus groups raises questions about the challenges that researchers undertake, and the dilemmas they face due to the need to develop a more participatory, transformative, and emancipatory action research. As conclusions of this study, we emphasise that the roles adopted by researchers are configured by their positioning when choosing AR as a method for investigating critical transformation of reality, their different conceptions of what knowledge mobilisation is, and their commitment to the precepts of inclusion.


Introduction
Some authors state that there is a significant gap between the knowledge generated by research on inclusive education and the one used by the teachers during their educational activity (Korsgaard, Larsen, and Wiberg 2018;Marion and Houlfort 2015). Specifically, in Spain in recent years, it seems that inclusive education does not move forward in schools, it remains stagnant (Echeita 2017;Nuñez 2019), and the positive changes experienced by the education system over the last decades, have been somewhat hampered, and the continuing drive that inclusive education demands has not taken place. Therefore, we have too much theory and a deficit in the effective development of this new paradigm. This mismatch between theory and practice may be because the way of communicating the latest advances and results of inclusive education research is not the most appropriate. Or maybe, because the knowledge generated by the research moves away from the schools' interests and possibilities.
Universities, and specifically university professors, have the responsibility of answering to this educational and social challenge, giving more prominence to social transformation, in accordance with the concept of University Social Responsibility (USR). Nevertheless, a part of the research output is included in scientific publications that have no applicability, and part of the researchers' objectives are more oriented towards disseminating knowledge and obtaining resources than towards managing its transfer to society. Nowadays, it is easy to find educational research focused on a more traditional and dominant approach in which researchers position themselves as experts and as the only actors capable of promoting and leading research projects and transferring them to society. As opposed to this transfer approach, the Social Sciences and Humanities Research Council (SSHRC) of the Canadian Government, coined, in 2004, the concept of knowledge mobilisation (KM), in order to be able to bridge the gap between research, policy, and practice (Levin 2011). This concept goes beyond the idea of knowledge transfer, referring more to the process of dissemination of the research findings in order that they can be understood by non-specialists. Knowledge mobilisation is used to designate 'a series of more interactive activities that cover the full research cycle: from the production until the distribution, and then the data usage' (Trocmé et al. 2009, 37).
Several definitions of KM circulate. According to Labbé et al. (2020) KM is intended to reduce the gaps between the knowledge users and relevant empirical knowledge, and to develop actions based on that knowledge. We use KM in this paper as an umbrella term for the range of active approaches to encourage the creation, sharing, and use of research-informed knowledge alongside other forms of knowledge. We also consider, as noted by Landry et al. (2008), that in a KM process there exist different types and degrees of interaction or collaboration between the researchers and the actors or practitioners, which in any case involve personal interactions and are multidirectional.

Action research and knowledge mobilisation in inclusive education
This idea of KM, transferred to the field of inclusive education, leads us to differentiate between what it is to research on inclusion and what it is to conduct inclusive research (Echeita et al. 2014;Parrilla 2009). This distinction places us, firstly, before the relationship that exists between the researchers and the research subjects, before the reflection on the ownership of the process and its results and, finally, before the methods and procedures for its dissemination. Both the researcher's positioning and the kind of interactions he or she promotes, determine the researcher's role, which varies tremendously, moving from the expert role to the role of 'outsider' or external person who enters the world of other people with whom he or she wants to construct credible and reliable knowledge. At the other end, we can find those who advocate for a relationship between researchers and those being researched based on equality and reciprocity (Parrilla 2010).
At the same time, action research (AR), and participatory action research (PAR), is described as an approach to research practice that 'place[s] the researchers in a position of co-learner and put[s] a heavy accent on community participation and the translation of research findings into action for education and change' (Minkler 2000, 192). Moreover, it is an approach to engage individuals to apply 'their emergent knowledge in generating an action toward social change' (Campbell 2010, 65). These definitions of AR echo ideas core to the concept of knowledge mobilisation. This relationship is addressed more extensively in the work of ACT for CFS (2014).
However, unlike other fields such as health, the literature on knowledge mobilisation through AR processes in inclusive education is not very extensive. Some studies show how university professors develop action research processes in schools, enabling group interpretative processes in which the different perspectives of professionals, students and scholars are involved in order to promote critical reflection, collaborative learning, and mutual criticism during the shared analysis of their own practices (Ainscow et al. 2016;Arnaiz, Haro, and Mirete 2017;Moliner, Sales, and Traver 2017;Murillo and Duk 2018;Sales et al. 2017). Therefore action research is conceived as a research strategy that allows a shared process of knowledge mobilisation and construction, professional development, and improvement of the classroom and school practice to develop.
We consider that inclusive education is one of the greatest challenges facing education systems around the world today. Inclusion is seen as a process of addressing and responding to the diversity of needs of all learners through increasing participation in learning, cultures and communities, and reducing exclusion within and from education. For coherence, the inclusive perspective raises the need to design other forms of research, where participation and power relations are more equitable, as in AR. In this context, Oliver (1992) considers that the only way to research on inclusive education is through processes that truly enhance the quality of life of those being researched. This includes changing the social relationships that result from the research itself and reconsidering the roles played by the researchers.

The roles played by the researchers in inclusive education
Roles, in contrast to functions that involve tasks that must be completed, are constructed interactively in the work context and in relation to the rest of the participants (Bolivar 1999;Louis 1981). According to this definition, we consider the roles are patterns of behaviour and positions one adopts in a set of interactions. They are linked to the expectations of self and others, in other words, they are personal constructs rather than assigned functions. Therefore, they are configured dynamically and are delimited by negotiation and construction processes in the course of the investigation. That is why we raised to research which are the roles that researchers in inclusive education play, which enable to mobilise knowledge in inclusive education, and thus minimize the existing gap between educational theories and practice.
The review of the state of the art allows us to distinguish between different roles that range from full membership of the investigated group (participant) to being a complete stranger (outsider) (Unluer 2012). The fact of being an external agent of the education community (teachers, families and students) allows the researcher to conduct an analysis of the processes at some distance from the reality being studied, in order to facilitate and support the process, helping the group to build coalitions and to obtain the necessary resources (Stuardo 2017;Traver, Sales, and Moliner 2010). It is also possible to act as a 'critical friend' so that the comments made serve to establish a dialogue on the dilemmas of the teaching profession and the research process (Calvo, Haya, and Susinos 2012, 15), on the dilemmas of methodological support and of thoughtful facilitators (Soto, Figueroa-Céspedes, and Yáñez-Urbina 2017). This facilitates the organisation of work that is actively involved and engaged in the process of change, providing new insights and ideas that contribute to the reflection of the coordination teams on their own educational practice (Figueroa-Céspedes, Soto, and Yáñez-Urbina 2019;Fiori 2007;Kemmis and McTaggart 2005).
On this road towards inclusive schools, the authors highlight the accompanying and advisory roles, from the assignment of managing and supporting the school in its vital processes on a global scale, directing the phases of the programme, identifying reflective rationality as an option for problem solving, and designing an action plan for the transformation process of institutional practices (Correa, Bedoya, and Agudelo 2015). Also, facilitating decision-making and participation, and gaining support among participants, relieving the sense of isolation, and providing a programme of personal development events through a mentor or adviser connected to each school (Ming et al. 2010).
At the other end, and emphasising the transforming and emancipating role of inclusive education, we can find those works in which the actors are given the role of researchers or co-researchers. This is the case with those who consider the teacher as a thoughtful practitioner who researches critically about his own praxis (Bergeron and Marchand 2015;Brookfield 1995;Iliško, Ignatjeva, and Mičule 2010;Kincheloe 2003;Schön 1983;Stenhouse 1975Stenhouse , 1981. According to Bergeron (2014), one of the most important issues in the development of inclusive practices is allowing a school to become a producer of its own inclusion capacity and conceiving the development of inclusive practices as a process to redefine the dominant social culture. For this purpose, it is essential that the researcher gives the professionals involved in the process considerable autonomy and agrees to follow a path whose boundaries are not drawn in advance.
We consider that the researchers' self-analysis of their own role in the action research processes will allow them to show the tensions they find themselves when conducting more democratic and collaborative research.
Some previous research gives us clues about how collaborative action research supported by the teachers' experimentation and reflection on inclusive practices generates social learning based on the existing conditions (Angelides, Georgiou, and Kyriakou 2008). Also, the work of Juma, Lehtomäki, and Naukkarinen (2017) shows the advantages and disadvantages of using collaborative action research in teachers' professional development to promote inclusive pedagogy. In addition, the research of Tragoulia and Strogilos (2013) considers that dialectics is crucial, and the work of Messiou (2019) focuses on the fact that collaborative AR must consider the students' opinions. Another interesting work by Abma et al. (2017) raises the importance of the reflection of the researcher on the processes employed to achieve equal participation and roles in action research. The work of Wang and Mu (2013) reveals the different roles that researchers play during a collaborative action research project: During the introductory stage, Instructor and Expert; Facilitator, Supporter and Resources Supplier, during the planning stage; Observer, Listener and Learner, during the implementation stage; Pusher and Affective Carer, during the data analysis and evaluation stage; and finally, Editor and Co-author of papers towards the end of the project.
This proposal raises many questions regarding the role played by researchers who use AR in the field of inclusive education. In respect of the idea that what we, the researchers, intend is to contribute to mobilising knowledge on inclusive education, we cannot focus on a knowledge construction and management process that is linear or unidirectional. Then, which are the roles that researchers adopt in the process? How does AR influence the adoption of these roles? Which contradictions and dilemmas do researchers face when they try to make AR 'more inclusive'?

Research method
This work is part of an R&D project funded by the Valencian Community administration and has a national scope (Spain). The research has been developed by researchers whose interest is focused on inclusive education. For this reason, this work contributes to the analysis of the complex reality of AR processes aimed at helping other researchers to learn about this research methodology. The goal is to present the reflections of inclusive education researchers on their own role in the AR process, bringing to the surface the problems and contradictions that they face. It also aims to shed light on the contribution of AR as an appropriate research methodology for knowledge mobilisation in inclusive education.
Methodologically, this is a descriptive study of a qualitative nature: 'The goal of qualitative descriptive studies is a comprehensive summarization, in everyday terms, of specific events experienced by individuals or groups of individuals.' (Lambert and Lambert 2012, 255)

Participants
The informants of this study are teaching and research staff (PDI) from different Spanish universities, whose teaching and research is developed in the field of inclusive education. All participants have in common that they are researching inclusive education using participatory research methodologies, such as action research (following Elliott 1991) or participatory action research (following Villasante, Montañes, and Martí 2009;Ander-Egg 2012). The aim of this study is to investigate the research approach of education professionals focused on inclusive education in their different research projects, in order to analyse the relationship between the researcher's role in mobilising knowledge from an inclusive research perspective and the strategies that facilitate or hinder it. For that reason, the specific topic to be researched in their projects was not so relevant, but rather how the research is carried out and from what factors (what roles are assumed by researchers and participants) in order to extract data that would help us to understand the concept of KM in PAR processes based on research using an inclusive approach. Therefore, the discussion established in the focus groups aimed to bring the researchers closer to a rethinking of how we position ourselves in PAR projects and from what perspective KM is achieved in inclusive education projects.
With these premises in mind, participants were selected using a non-probabilistic and incidental method of sample selection, using those teachers who expressed their availability and interest in participating in the research. The total number of participants was 24 university teachers (9 men and 15 women), belonging to ten universities located throughout Spain. The participants were between 29 and 62 years old. 58% of the participants belong to the Network of Universities and Inclusive Education (RUEI) and all the researchers develop or have developed research-action processes on inclusive education in schools.
The technique used to obtain information was the focus group. Specifically, four focus groups were organised in different locations in Spain: Granada, Castellón, Murcia, and Seville. In the focus group formed by RUEI researchers in Granada, the projects of each researcher were all different but shared the common basis of research on inclusive education and participatory research methodologies. In this case, the participants responded to the questions by establishing dialogues and shared reflections on the participatory research processes that each of them carried out. In the case of the focus group carried out in Murcia, the researchers also belonged to different research teams, but all of them focused on the processes of supporting educational centres in order to promote the transformation towards more inclusive schools. Meanwhile, the focus group in Seville was composed of researchers who shared research projects focused on the barriers faced by persons with disabilities in Higher Education and appropriate support. Finally, the focus group in Castellón was also made up of researchers from the same team, whose research projects focused on the support of centres through PAR processes.

Data collection instrument and process
The data collection instrument was a focus group guide that included three sections: a) protocol instructions, b) socio-demographic data, and c) open questions about the roles assumed by the researcher to mobilise knowledge about inclusive education in the transformation processes. Focus groups are used considering accessibility criteria, in order to have more participants and obtain more information in a shorter time. In addition, focus groups allow for dialogue and interaction between participants, which leads to interesting discussions among them. The informants, as this was a purposive selection, were personally invited to participate in the research and most of them accepted with interest.
Firstly, we turned to the RUEI network, during the National Meeting of Universities and Inclusive Education, held in Granada. Secondly, we invited researchers with experience in the field and, thirdly, we asked new teachers. The number of focus groups was decided gradually depending on the availability of participants, geographical distribution and in order to reach information saturation.
In all cases, before starting the focus groups and for ethical reasons, each participant signed an informed consent form. Each group was led by two researchers (a moderator and a speaker) and lasted between 70 and 90 minutes. All interventions were audiorecorded.

Data analysis process
A literal transcription of the data was made, from which a content analysis was conducted using ATLAS.ti software (The Qualitative Data Analysis & Research Software). Content analysis and data reduction were performed using a mixed method approach, namely using both deductive and inductive logic by the identification of units of analysis, which were then categorised and codified using the coding system derived from Wang & Mu's The researcher provides basic research knowledge and skills, instructions on how to conduct the PAR. In such cases, the actors undertake the actions and are passive subjects that receive theory. Most of the actors find this approach useful and expect the researchers to tell them what to do.
Workshops: Training course that includes how to develop questionnaires, conduct smallscale surveys, analyse data and report on results.

STAGE OF PROBLEM DELIMITATION AND ACTION PLANNING
Facilitator (S2.F) The researcher develops strategies to invite the actors to explain the problems or needs they face in their class or in the school, and why they consider them a problem.
They ask questions and facilitate reflection on the problems they want to address and why.
Partner/ Collaborator (S2.P/C) Researchers help the actors to explore their assumptions and beliefs, and try to make sense of what they did and why they did it the way they did so that they are able to make informed decisions about proposing solutions for the problems they had. In this approach, researcher and actor are viewed from a more egalitarian perspective.
In the Project meetings, they communicate in a more egalitarian way.
Resources Supplier (S2.RS) The researcher provides resources for educational research. The actors in the field (teachers) do not have habits of reading texts related to the topic addressed. With the aim of seeking solutions (action), researchers provide resources in order to know what is done in other contexts given this situation.
The researcher helps the actors to review their data collection instruments (such as questionnaires, interviews), and to analyse the data collected. They provide journal articles, organize school visits.

STAGE OF ACTION IMPLEMENTATION
Observer (S3.Ob) Researchers observe the teachers' actions. They observe the development of the action. Listener (S3.Li) Researchers listen to the teachers' explanations and exchange ideas as workmates/ equal partners.
Researchers understand the difficulties teachers face when they try to balance the lack of time and heavy workload with their research, and become increasingly more impressed by the teachers' penetrating insights into teaching, their deep love for education and care for the students' growth. Learner (S3.Le) Researchers are aware of the opportunity to learn and deepen in their understanding of teaching and learning. Researchers change their perception of teachers from knowledge consumers to thoughtful thinkers.
They understand and value the actors' work. They realize that teachers have their own contextualized knowledge, and that they need the opportunity and the appropriate means to make it explicit.

STAGE OF MONITORING AND EVALUATION OF RESULTS
Pusher (S4.P) In order that teachers can remain in the project (since the actors lack time because of the daily work overload), researchers assume a new role to guarantee that all the teachers can follow the steps taken in the project. Therefore, this role of the researchers works as a pusher for progress. Researchers encourage and remind of the pending research tasks so that the project does not become stagnant.
They remind of the data gathering in order to monitor and evaluate the action. They provide the necessary help and support for data analysis.
Proposal (2013). The codebook definition was specified through the intersubjective agreement of the researchers. Table 1 shows the codebook used during the data reduction and categorisation process. Three emerging categories were identified during the inductive analysis.

Results
The results achieved in this study are presented below, considering the semantic content of the units of analysis. A descending order is followed, according to the percentages obtained in each of the subcategories (See Table 2), based on the referenced units of analysis. From there, the most outstanding fragments were recovered. Overall, the role the participants in this study mentioned most was the Partner role (26.1%). The researchers stated that they helped the actors to explore their assumptions and beliefs in order to make sense of what they did and why they did it the way they did so that they were able to make informed decisions when proposing solutions for the problems they had. Researchers offer the actors personal emotional care as friends and listeners. The researchers' understanding of the teachers mobilizes the latter in order that they struggle to overcome the difficulties and continue carrying on their research project.
They listen to the concerns. They send help messages.

DISSEMINATION STAGE
Editor and Co-author of the Research Report (S5.E/Co) Researchers encourage the actors, and they offer them help and guidance to write, review and improve the research reports they produce.
A report template is provided. Joint work when reviewing the report.
Reports are reviewed and rewritten. It is a role closely associated with knowledge mobilisation processes because both researcher and actor are viewed from a more egalitarian perspective. They stand 'side by side' to search together for solutions to school problems.
Another role highlighted was that of Facilitator (16.3%), developing strategies to invite the actors to explain the problems or needs that exist in their context, and why they consider them a necessity.
"Actually, it comes from real needs, but also from the needs perceived and lived by the people present. Because we can observe some needs, but if these needs aren't perceived or lived, that research may have no impact." (F_FG1_I4) This fact is closely linked to the topics object of study. In order to achieve knowledge mobilisation, it is essential to research the real problems of the context.

"As long as we conduct a research that is more focused on real problems, not because there is an approach that becomes trendy, and then we all research [. . .], but because we research on real problems that teachers are facing and we generate knowledge that they can use." (F_FG2_I12)
Several researchers highlighted, from a more participatory approach, the need to facilitate the negotiation of the demand, and start from the needs of the actors and the context. Thus, the researcher assumes the role of Facilitator as long as he or she plays the role of the 'critical friend', helping them to reinterpret their practices, or questioning them in order to prompt reflection that leads them to improve their practice or modify it towards something that is closer to inclusion. Therefore, the researcher, in the role of Facilitator, builds bridges between society and university, that is, mobilises two types of knowledge, brings practical and academic knowledge closer together and places them in interaction in a process of shared construction.
On the other hand, 14.1% of the informants stated that they adopted the role of Expert by providing basic research knowledge and skills to the actors in the field. 'Sometimes, we act as experts. [. . .] people need to have information or data [. . .] that aren't within their grasp. Then, we or somebody we know gives them this information.' (IE _FG4_I19). Although researchers did not mainly identify themselves with this role, some of them played the role of experts, considering that the fact of counting on an external and expert view could help the actors to reconceptualise and reconsider their needs before initiating the research. But they also explained that this created some dilemmas, since it could influence the orientation of the demand, whose starting point should always be the actors.

"I think that our research doesn't have to start from ourselves, but from the participatory demand. I don't know, maybe what I am saying is very utopian." (IE_FG1_I4)
Researchers do not feel comfortable in this role, which is closer to the knowledge transfer approach in which actors are passive subjects receiving theory, although they feel 'pushed' to develop it in some situations.
Researchers also referred to the role of Editor and Co-author of the research report (9.8%), aiming at disseminating the results generated.
"The role of the researcher must also cover, in addition to the proper research actions, the dissemination of this knowledge through certain alliances with the actors." (E/Co_FG2_I10) Nevertheless, they considered that, beyond the purely academic environments (for example, articles, conferences, seminars, etc.), the dissemination (outreach and accessibility), and the monitoring of the research impact was still a pending subject. And in this regard, the researchers were willing to offer help and guidance to write, review and disseminate together, recognizing their own weaknesses and considering that learning about shared dissemination is mutual and generates knowledge mobilisation.

"Because we and the schools move at different paces. We can't always be in the school, nor practice co-writing . . . we have done some little things, such as, for example, the joint dissemination of the results. This is just great, we did it in a seminar, the teachers and us, and this was a huge knowledge-building experience." (E/Co_FG4_I20)
Other researchers, however, added an interesting nuance to this role. Beyond co-editor or co-writer, they spoke about the role of co-researcher. They considered how, through a more participatory research approach, both researcher and actor are viewed from a more egalitarian perspective, and it is an approach that is more conducive to the adoption of the role of co-researcher by the actors. Co-research makes it possible to develop a process of knowledge mobilisation based on the idea of shared knowledge construction. The seminar, as a working group that allows the establishment of egalitarian relationships, is considered a strategy for the shared construction of knowledge and professional development of both teachers and researchers. "Now, for example, in the school X, many of the people who are there play this role, but at the beginning, I think that when you start a research, taking that role of co-researcher [. . .] it is an arduous task." (FG4_I22) The researchers also considered the change of perspective required by appreciative inquiry (AI) moving from researching 'about' to researching 'with', showing the need for evolving towards a more inclusive research. 'And this shift in thinking means changing the paradigm so that the role of the researcher is actually an inclusive one.' (FG1_I1) They underline the importance of a paradigm shift, i.e. a change of positioning in line with the transformative and emancipatory role that inclusive education demands and is the purpose of any AI process.
In the same vein, regarding the role of the Learner (6.5%), the researchers stated that they were aware of the opportunity to learn and deepen their understanding of the practical processes.
"We learn a lot during these processes. When you rescue this knowledge, you name it or systematize it a little bit . . . You analyse which transformations are taking place there, in the school, this is an extraordinary source of knowledge. Then, it is bi-directional and you position yourself in that role of standing by . . . " (Le_FG4_I20) Researchers revealed how they had changed their perceptions of the actors in the field (teachers), seeing them as insightful thinkers, favouring a shared construction of knowledge.
"The construction of that knowledge has to been done in an egalitarian way, and I think that it starts out from another starting point that is that knowledge that is mobilised at any time and that is built by everybody. Thus, all branches of knowledge, not only the researchers' ones, are the important ones, or the ones that are entitled to build new knowledge." (Le_FG4_I18) Once again, the ideas of bi-directionality and shared knowledge construction emerge as defining processes of knowledge mobilisation brought about by AI.
Related to the above, the role of Listener (4.3%), which entails listening to the actors' explanations, was also referenced: 'Listening, paying attention, mostly from my field, the social field, I think you need to be completely available and open about the social and educational agents that we can find in the community. ' (Li_FG3_I15) This role involves an active listening process that includes fostering the exchange of ideas and articulating speeches shared between researchers and actors.
"We used a vocabulary they didn't understand, then they assumed our vocabulary, but without giving it content. We designed a strategy, which was to read our articles. I remember it was really interesting because at that moment our vision about what democratic practices are wasn't the only one, since they themselves gave content to what this meant. [. . .]

I think that knowledge mobilisation is somehow similar: in how our theoretical framework or foundation fits in with what is worked in the school." (Li_FG4_I24)
Again, this means mobilising knowledge in the sense of giving meaning to practice, conceptualising, theorising, constructing scientific knowledge from practical situations.
In terms of the relationships that are established between researchers and researchers, which is another key factor in the mobilisation of knowledge they highlighted the role of Affective Carer (3.3%), related the personal emotional care.
"I like working with vulnerable groups. It is like a magnet, that's how it feels to me, I have a magnet in this respect. Then, the approach between the parts isn't reached through a formal I+D project, but from the personal side, through commitment and responsibility." (AC_FG1_I6) They highlighted how life histories helped to empower those being researched. The importance of the transformative and emancipatory positioning of the researcher is once again evident here.
"(Regarding research with vulnerable groups) In relation to life histories, active listening, helping to empower that person who has been crushed and stomped on so many times that he/she can't stand up but with the help of that hand next to him/her." (AC_FG1_I6) In addition, another meaning of care emerges, this time about the care relationships between research group members (intragroup and towards the group): "We have to take care of ourselves. The group isn't only an efficient machine for doing research, but sometimes it is also therapy, sometimes family, sometimes friends. [. . ..] Having those other perceptions is essential in order not to lose our way and not to lose sight of the ethical dimension as researchers and as individuals." (Emg_FG4_I19) Regarding the role of Pusher (2.2%) the informants referred to the need to guarantee that everyone involved could participate following the steps of the project.
"When we give some information to the students, we meet them so that they understand it, because if you organise a session so that they can make decisions, but you give the same information to the students, families and teachers . . . you lack something that is necessary in order that they can make that decision with a deeper understanding." (P_FG4_I22) From this role, they considered it important to create spaces and channels that facilitate the meeting and participation of all the participants in a clear inclusive attitude.
The role of Observer (1.1%) takes this emancipatory idea expressed by the researchers to the limit, when the development of the action is led by the actors: "In the last research project, the students were able to lead the last stage of knowledge mobilisation in the Service Learning Project. And it was them, tutored and supervised by their teachers and accompanied by their families. We were observers, simply at their service [. . .]  This implies an exchange of roles and a strong emancipatory positioning. On the other hand, the respondents mentioned other roles that were not aligned with the idea of knowledge mobilisation and action research processes. For example, the Resources Supplier (6.5%). This is based on the idea that the actors in the field (teachers) lack time or reading habits, and therefore, it is the task of the researcher to facilitate the relevant material: 'I had time, or more time than a teacher, to read those contributions and provide a conceptual-theoretical frame to those practices.' (RS_FG1_I4).
The same applies to the role of Manager/ Bureaucracy (Emg.MB) (4.3%) which was considered as an emerging category: 'There is an institutional role that demands from us . . . everything related to the managing deadlines [. . .]. You must keep an eye on the deadlines, when the aids are published. I'm like a researcher-manager.' (Emg.MB.FG1_I3).
Finally, two new emerging categories could be related to knowledge mobilisation strategies in inclusive education: The role of In-service and initial training adviser (Emg A) (3.3%): 'It is important to keep with the area of training, teachers initial training, that is to say, those teachers who in the future will run educational centres or will be in the classes with the students must know this research. Therefore, the knowledge acquired during the research on inclusive education must be directed to the initial training programmes and curricula.' (EmgA.FG2_I10) And the role of Educational administration adviser (EmgA.EA) (2.2%) which is related to building partnerships with the political and administrative field: 'I mean, researchers must also provide decision-makers with this knowledge and body of positive thinking towards inclusion.' (Emg A.EA.FG2_I10) In both cases, the knowledge generated by inclusive education research needs to be mobilised to have an impact on contexts beyond the practical: the academic context (training) and the policy context (administration).

Discussion of the results
In this section, we will try to problematize the roles that inclusive education researchers play in the AR process, observing the reflections, dilemmas and contradictions that they have. However, trying to listen to our inner voice, first of all, we are going to address the big question that has arisen from the research approach itself. Thus, the first discussion is generated by the assumption of the proposal made by Wang and Mu (2013) as a data analysis model. In the interest of guaranteeing the scientific rigour of the study, we may focus too much on the superficial structure of the AR process, by assuming the correspondence between roles and stages. Although the stages of AR are widely defined in the literature with the contributions of Dewey or Lewin (Elliott 1991;Schön 1983;Stenhouse 1975), it is true that neither knowledge mobilisation nor the construction of the different roles has a linear nature. In fact, some roles, such as the one of accompanying or affective caregiver, were not considered linked to a specific stage of the process. Nevertheless, although we are aware that this may be considered by some people as a limitation of the study, the self-criticism of it leads us to formulate the first conclusion: the construction of roles has a non-linear nature. Researchers play their roles not depending on the stages of the process but on their convictions about the nature of the AR itself and its epistemological basis, with a clear objective of learning and improving.
In the second discussion, we want to place value on the tensions and dilemmas that arise from the results obtained. Regarding knowledge mobilisation, the roles expressed by researchers are situated in different conceptions of what knowledge mobilisation means.
We could identify four perspectives: (a) Mobilisation as transmission of knowledge from the scientific and academic context into practice The role of Expert is not without dilemmas and tensions. Assuming this role means considering the rest of the actors involved as passive subjects who receive theory. According to Stuardo (2017), it involves a unidirectional transfer of knowledge. Although most of the researchers do not identify themselves with this role, they assume it. If they do it at the beginning of the project, they question how this can influence on the orientation of the demand and problem delimitation. According to Bergeron (2014), the researcher must know how to leave room for the professionals when he or she takes his/ her place as researcher-trainer-expert, which, many times, causes dilemmas between the role that the actors give him/her and the one he/she aims to play.
(b) Mobilisation as a bi-directional, interactive and participatory process between researchers and practitioners in which the shared construction of knowledge takes place.
In AR, especially in problem definition and action planning, the role of the researchers as Partners and Facilitators is essential; helping the actors to explore their professional thoughts and perspectives, on the basis of critical reflection, collaborative learning and shared analysis of their own practices (Ainscow et al. 2016;Murillo and Duk 2018;Moliner et al. 2017) in order that they are able to make decisions and propose solutions for the problems they have (Arnaiz, Haro, and Mirete 2017;Moliner et al. 2017).
On the other hand, in action implementation, it is a dramatic shift, and the researchers place themselves as learners, faced with the dilemma between doing, not doing and letting others do. The actors play a prominent role, relegating the role of the researchers to a secondary place in the process, from which they value the work of the actors, and realize that teachers have their own contextualized knowledge, their own voice, and construct meanings (Iliško, Ignatjeva, and Mičule 2010;Koutselini 2008), which they value as a learning opportunity. This involves the forging of a new teaching identity, pedagogically competent, capable of investigating and reflecting on his/her teaching practice, together with other professionals, and who is aware of the social and ethical issues of his/ her profession (Durán and Giné 2017;Parrilla 2009). This means the legitimacy of the 'other' as constructor of knowledge, practical knowledge, is valued, and it is assumed that knowledge is mobilised at any moment and is constructed among all.
Therefore, these dilemmas lead us to consider that AR puts pressure on the researchers' roles as they aim to 'transform power relations towards greater democracy'. (Greenwood and Levin 2007, 73).
Moreover, when talking about research processes in inclusive education a very relevant role is highlighted: the role of Affective Carer. Although the results suggest that the researchers refer to the personal emotional care that they show for the actors, or among themselves, it is important to stress the sense this role acquires when the relationship established between both of them is a one of equality, reciprocity and horizontality. The affective dimension, when dealing with processes that promote inclusion, refers, according to Hopkins and Stern (1996), to the development of mutual commitment (willingness to help others) and of sympathy (enthusiasm and affection for others).
In addition, the results enable us to discuss another role presented: that of Co-author, which, according to Wang and Mu (2013), becomes more visible during the dissemination stage, when considering the publication and co-authorship of the research report. Nevertheless, this role is close to that of Co-researcher, with the difference that this affects all the stages of the process, from the initial proposal until its dissemination, in the case of a collaboration between researchers and practitioners. At this point, it is interesting to distinguish between 'participation' and 'collaboration'. According to Talajic (2013), research participation is often advocated as an approach to increase the researcher's knowledge and skills in research. However, research in collaboration means that researchers contribute equally as pairs and co-learners, in collaboration with the actors (Huang 2014).
This opens up the vision of AR as an appropriate research methodology for knowledge mobilisation in the framework of inclusive education. Kemmis and McTaggart (2005) describe a number of broad types of AR including participatory research, critical action research, classroom action research, action learning, action science, soft systems approaches and industrial action research. The results show that the researchers who have participated in this work describe their roles from a participatory perspective, expressing their willingness to adapt to the needs of the actors and the context; with the result that the process of negotiation of the demand becomes relevant. The negotiation of the demand brings us back to the zero point of the research, prior to the initial stage: the first contact and the shared negotiation of the demand. According to Ander-Egg (2012), at this stage the aim is to explore the demand, working together in an attempt to identify the project's problem and objectives. This involves, as Parrilla et al. (2017) say, researchers sharing a research agenda and putting themselves at the service of the participants. Thus, researchers are positioned in a more inclusive research perspective that requires researching 'with' the actors instead of 'about' them.
(c) Mobilisation as an emancipatory perspective, as the possibility of empowering actors to direct their lives and processes of school transformation In inclusive research the participants become co-researchers, active subjects of a process that also belongs to them, since it fully affects them. This is a research that dignifies them and releases them from the mere status of object/ patient of other interests. According to Bergeron (2014), inclusive research implies the development of inclusive practices as a process to redefine the dominant social culture and it is essential that the researcher gives the professionals considerable autonomy to lead the AR process. Emphasising the transformative and emancipatory role of inclusive education, we can find reported situations in which actors are given the role of researchers or co-researchers. This allows the empowerment and emancipation of vulnerable groups (Alba and Nind 2020), which is indispensable for a more inclusive AR.
(d) Mobilisation as the impact of knowledge generated by research in other contexts such as the academic and political spheres.
Two roles emerging from the study, In-service and initial training adviser and Educational administration adviser, highlight the idea that the knowledge generated by inclusive education research needs to be mobilised to impact different contexts (academic and political). In these cases, the effects of research are indirect and gradual, typically occurring over time as idea get taken up and mediated through various social processes. Fortunately, according to Levin (2011), governments are increasingly paying attention to 'evidence-based decision-making', including the establishment of new policies.

Conclusions
A first conclusion of the study leads us to rethink how the roles adopted by researchers are configured, not only due to the responsibilities undertaken, but also the researcher's positioning when choosing AR as the method for investigating and approaching the educational reality. A research that, to be inclusive, must be, in turn, more participatory, transformative and emancipatory. It must be linked to the problems and needs of the groups and entities involved, and put at their service, that is, it must be transformative, and based on real problems detected, such as situations of exclusion and injustice. This research urges researchers in inclusive education to assume a transforming researcher's role, in line with the PAR processes, which breaks the gap that exists between educational theories and practice, and that is consistent with the precepts of inclusion. A role that models democratic procedures that are fully inclusive and gives a voice to all research participants. A researcher role committed to the critical transformation of reality.
As a second conclusion, we find that researchers are confronted with the roles they have traditionally adopted and which still persist in action research processes. Regarding knowledge mobilisation processes, researchers take on roles associated with different conceptions of what knowledge mobilisation is, which leads to contradictions and dilemmas. On one hand, roles more associated with a model of knowledge transmission (such as expert or resource provider) persist. On the other hand, they express roles in line with a process of co-construction, interaction and emancipation (such as partner, facilitator, coresearcher) more in line with the AI processes they are developing.
In order to handle such dilemmas it is necessary to: recognize them and face them; share them with the team and other researchers who participate in the Research Networks on Inclusive Education; learn to leave room for the professionals and practice actors, namely, exploring, valuing and appropriating strategies for knowledge co-creation and inclusive research; and in coherence, keeping an eye on the principles of inclusion (equity and social justice), incorporating them into one's own research practice. According to Skipper and Pepler (2020), embracing a knowledge co-creation model involves identifying and involving the partners in an early stage of the research process and working collaboratively with them in order to co-create knowledge in each of the stages of the research process. In short, the adoption of action research processes in the field of inclusive education demands that the researchers adopt roles that are in accordance with the latest three KM perspectives: co-construction of knowledge, an emancipatory approach, and impact in academic and policy contexts.
We consider that PAR is the appropriate framework for developing these approaches in the field of inclusive education, and that research on inclusive education in Spain must continue progressing in this direction.