Do both ‘get it right’? Inclusion of newly arrived migrant students in Swedish primary schools

ABSTRACT The aim of this article is to advance knowledge on how Swedish primary schools organise education and what strategies they deploy to ensure inclusion and attainment of newly arrived migrant students. The article is based on semi-structured interviews with 30 teachers and school administrators, and one-year of fieldwork undertaken in two multicultural urban primary schools in the Stockholm region. One of the schools initially places students in separate classes, while the other one places them directly into mainstream classes. Both are evoking inclusion and attainment as a reason for using their respective models. As such, do both ‘get it right’? Using inclusion as the theoretical and conceptual framework this article addresses the broader question: How is the meaning of inclusion constructed in the processes of its practical implementation in these two schools? The results show the ambitious tale of inclusion in both schools was, in the process of the construction of its meaning and implementation, reduced to some of its aspects. Teachers and school administrators are allowed to include or leave out of their model whatever they deem necessary, obsolete, expensive or unrealistic and still fitting under the umbrella of inclusion. Sometimes it works, sometimes it does not, and both schools ‘get it right’ and ‘wrong’ in some aspects.


Introduction
Swedish research on student achievements in primary and upper-secondary schools (age 7-18) has for many years corroborated that migrant students arriving after typical age of school entry (age 7), on average, are going to have a substantially lower attainment at the end of their schooling than their Swedish-born peers (SOU 2019:40). 1 With reference to a set of particular challenges many newly arrived migrant students (NAMS), especially refugees 2 , bear with them, such as previous short or disrupted formal education, lack of skills in a dominant language, traumatic experiences, and having parents with limited or interrupted access to education, the expected low achievement is presented as almost inevitable (see also Block et al. 2014). The students' high ambitions and motivation (Svensson and Eastmond 2013) are perceived as a reflection of insufficient information on the structure and operations of education system, rather than a potential fuelling educational success. One of the paradoxes is that schools appear to try to 'correct' the ambitions and dreams of young newly arrived migrants, especially refugee children at upper-secondary level (Fejes et al. 2018), by offering advice that becoming a doctor or a lawyer is virtually impossible, and that they should aim at something more realistic and achievable: often working-class jobs. Obviously, NAMS' backgrounds cannot be altered, but the deficit-based (Uptin, Wright, and Harwood 2012;Migliarini, Stinson, and D'Alessio 2019) educational responses to their particular needs must be critiqued and addressed. A first step in addressing the issue is to understand the way schools currently envision, organise and carry out the inclusive education of these students.
The aim of this article is to advance knowledge on how Swedish primary schools organise education and what strategies they deploy to ensure inclusion and attainment of NAMS. The article is based on an ethnographic study undertaken in two primary schools that enrol students aged 7-15 in the Stockholm region. Both schools are situated in low income, multicultural urban neighbourhoods. Based on public statistics and information gained during the initial fieldwork, we know that the majority of their students are of immigrant origin (being born abroad or having their parents born abroad), both have received NAMS for decades, currently both accommodate similar numbers of NAMS, but they deploy divergent organisational models. One of the schools initially places students in separate classes, while the other one places them directly into mainstream classes.
Furthermore, both evoke inclusion and attainment as a reason for using their respective models. Do both 'get it right'? To address this issue, this study investigates the following research questions: What are the distinguishing features, and the underlying principles of inclusion the schools are building their models around? How is the meaning of inclusion constructed in the processes of its practical implementation in these two schools? Our explicit aim is not to compare the schools and their models, in order to possibly propose which is doing better, but rather to deepen the knowledge on how the interplay between local contextual factors and various organisational approaches condition, shape and justify the variety of practices, their supporting narratives and outcomes.

Newly arrived migrant students in the Swedish context
Parallel with facing an unprecedented arrival of a large group of refugees in 2015, 3 the Swedish education system was preparing for implementation of major policy changes (Bunar 2018). For the first time a precise definition on who and for how long time is to be considered as newly arrived migrant student was adopted (Skollag 2010:800). Thus, a newly arrived migrant student is one who had lived abroad, now lives in Sweden, and has arrived to a school after typical age of school entry (age 7). A student is considered as newly arrived up to four years after arrival. All children, irrespective of their migration statusundocumented, asylum-seekers, under temporary protection have the right to education under the same conditions as Swedish-born children.
A mandatory screening assessment of students' previous school experiences and subject matter knowledgeoften managed by teachers in Swedish as a second language, assisted by multilingual personnel based on a template produced by the National Board of Educationmust be conducted no later than two months after a child arrives to a school. The screening assessment consists of three subsequent steps: (a) gathering basic information on a student's background, experiences, the language(s) spoken, number of years in formal or informal education; (b) the attained level of knowledge in numeracy and literacy according to age; (c) the attained level of knowledge in academic subjects. The results are stored in a student's personal file, available for teachers to access when they plan teaching activities (Bunar 2018).
Based on the initial screening assessment and the involved teachers' recommendation, a school principal decides if a student will be placed in an age appropriate class, which is strongly recommended in the legislation due to social reasons (Skollag 2010:800). The principal also decides whether a student will be directly placed in a mainstream or in a separate class (in Swedish förberedelseklass). If the latter is the case, then the student cannot be there longer than two years, which means that regardless the attained level of Swedish they must be completely transferred to a mainstream class. Furthermore, students in separate classes must be offered some instruction together with students in mainstream classes, although it is not regulated how many hours per week and in what classes. But often this requirement is fulfilled through participation in sport, music, and art (SOU 2017:54).
According to the legislation (Skollag 2010:800; Skolförordning 2011:185) NAMS have the right to multilingual classroom assistance (see  if they need it. It means deploying support staff with knowledge in a student's first language, or any other language a student is in command of, to help the student understands the content of lessons, home-work and other requirements (Warren 2017). Although the national legislation does not prescribe the number of hours to which students are entitled, the assistance is often offered one to two hours per week (SOU 2019:18).
Finally, there is a set of other reforms and measures undertaken on the local level. Some examples include professional development of school staff, most often on translanguaging (García, Johnson, and Seltzer 2017) and language and content integrated learning (Gibbons 2013), allocation of additional resources, and even bussing. 4 The ultimate goal of all these measures and regulations is to mitigate inclusion and attainment of NAMS in schools. The Swedish legislation and a wide range of adopted measures could be in international comparison (see Dryden-Peterson 2011;Crul et al. 2019) considered as advanced and for the most part in accordance with research induced recommendations. Sweden appears thus as an interesting empirical case to explore, where a combination of fairly large number of newly arrived, a set of new ambitious policies and a general debate on school segregation and increased inequalities, particularly putting migrant children in a spotlight (Skolverket 2019), occurs simultaneously.

Theoretical and conceptual framework
The concept of inclusion serves as a general theoretical and conceptual framework in the article. Due to limited space, we will below just briefly touch upon some of the concept's, for the purpose of this article, most salient features.
Following Ainscow and Messiou (2018, 2) who claim that 'internationally, however, it [inclusion in education] is increasingly seen more broadly as a reform that responds to diversity amongst all learners', we want to explore how inclusion is used in response to diversity among learners with regard to a particular segment of a student population, newly arrived migrant students in two school contexts. We argue there is a need to infuse often unreflected ideological and distilled theoretical discussions with findings from empirical cases where the meaning of inclusion is being constructed through the very processes of its implementation. This practical meaning of inclusion among the educators and its outcomes shape lives of a whole new generation of young people.
Inclusion is a multifaceted concept, sometimes questioned and more often championed in research and policy alike (see contributions in Schuelka et al. 2019), and sometimes difficult to define. According to Azorín and Ainscow (2020, 59): 'However, capturing the concept of inclusion as both an educational principle and a practical application remains amongst the most challenging tasks related to education'. For many years inclusion and inclusive education have been identified with special education and students with special education needs (Ainscow et al. 2006;Kugelmass 2006). A considerable conceptual and theoretical development have taken place within this framework, which eventually came to include other disadvantaged groups (Ainscow and Messiou 2018;Materechera 2020). As Van Mieghem et al. (2020, 676) put it 'IE [inclusive education] is therefore part of a broad human rights agenda that emphasises the value of educating all students in mainstream education'.
The most pernicious understanding of inclusion is the one that reduces the concept to just one of its aspects. Providing NAMS with placement in mainstream classes, without undertaking any other systemic changes (Shaeffer 2019), as an undeniable evidence that inclusion is accomplished, is perhaps the most frequent misunderstanding. As reported in international research (Taylor 2008;Block et al. 2014;Grigt 2017;Hilt 2017;Vogel and Stock 2018), inclusion through direct immersion in mainstream classes seems to be a theoretically appealing idea, but almost impossible to build a sustainable teaching strategy on. The critics, and among them many teachers (Juvonen 2015) and parents (Allan and Slee 2019), have pointed out that direct immersion of NAMS in mainstream classes often live up to only one factor, physically shared spaces (Riggs and Due 2010) and essentially lead to exclusion through inclusion. Without tailored support, language and content integrated learning, mentors, limited or non-existent multilingual language assistance, cooperation with parents and local communities, this policy of 'sink or swim' lays the ground for the emergence of zones of exclusions (Zembylas 2011).
The approach labelled as inclusion through exclusion (Budginaitė et al. 2016;Vogel and Stock 2018), placing students in separate classes, has emerged as a main policy response in many countries. The rhetoric cannot obscure the fact that separate classes is de facto exclusion. According to Allan and Slee (2019, 8): 'Inclusive education rejects exclusion and with it segregated schooling. Segregated schooling, on whatever grounds, is a platform for an education in social stupidity'. Furthermore, there are no guarantees that all prerequisites required to justify separation and make this model successful are met. As a matter of fact, the absence of swift and well-prepared transition to mainstream classes and high-quality and challenging instruction as well as the expectations by subject matter teachers that newly arrived after the transition are 'just like everybody else', has been the most posed critical remarks (Bunar 2015;. Arnot and Pinson (2005) argue for a holistic inclusion perspective, meaning a combination of general measures aimed at changing school culture and practices (an ethos of inclusion, a celebration of diversity, and a caring ethos and the giving of hope) as well as targeted measures aimed at individually addressing learning, linguistic and socio-psychological needs of NAMS (Taylor and Sidhu 2012). Similarly, Bunar (2019a) has proposed the concept of support-based inclusion, consisting of: (a) removing barriers or as Ainscow and Messiou (2018) put it, eliminating exclusionary processes (see also Slee 2011). The barriers could be legal, organisational, pedagogical, or related to attitudes; (b) providing shared spaces where NAMS can interact and benefit from relations with non-newly arrived and be socially included (see also Van Mieghem et al. 2020) and; (c) making sure individually tailored socio-emotional and pedagogical support measures are in place (see also Messiou and Azaola 2018). In other words, there is no inclusion without removing barriers and eliminating deficit-based approaches (Uptin, Wright, and Harwood 2012) and exclusionary processes in form of subject matter teachers' disengagement, policy and attitude driven stances on the harming role of a first language, bullying and discrimination, and the absence of tailored support (Pinson, Arnot, and Candappa 2010).
Hence, inclusion, as understood and used in this article, is based on rebuffing the reductionist views of the concept. Instead, we advocate for a holistic or support-based inclusion where shared spaces are accompanied with individually tailored support and efforts to remove barriers in the wider school context.

Method and material
Empirical material was collected by the first author during one-year fieldwork in two primary schools in Stockholm region, here named the Rose Garden school and the Carnation Garden school. In accordance with the study's aim, we sought contact with schools that partly receive a significant number of NAMS in grades 7-9 (age 13-15). And partly, the schools deploying separate classes and direct immersion in mainstream classes respectively. While contacting schools and asking about their circumstances and willingness to participate in the study, we were constantly referred to and thus gradually steered towards a small pool of schools who met the basic selection criteria. As it turned out, all of these schools were located in the region's low income and ethnically diverse neighbourhoods. A very few, if any, of their students have Swedish as their first language. Eventually, two of these schools agreed to participate.
The fieldwork commenced in the autumn of 2018, with the first author's observations of the schools' daily life, participation in staff meetings, gathering and analysing secondary material and conducting a series of informal talks with school personnel and students. An essential objective was to acquire an understanding of schools' policies, models, their supporting narratives, formal and informal instructional practices, relations between different actors, generally, and in particular with respect to addressing educational needs of newly arrived. The initial fieldwork elicited also a number of important questions, that during the spring of 2019 were used as a template for semi-structured interviews with school personnel and NAMS. As Hammersley (1998) argues, learning directly, 'on the spot', about the field studied and simultaneously engaging in an ongoing methodological and theoretical reflections, is the most proper way of moulding scientifically interesting and practice-oriented significant questions. Our theoretical reflections embraced, from the outset, the notion of inclusion and the interest in how it was understood, constructed, and utilised in these two schools.
In total 42 respondents were interviewed in both schools. 12 were newly arrived children (age 13-15) and 30 were teachers in both mainstream and separate classes, school management, and multilingual classroom assistants. Particularly important are the interviews with school leaders, which Van Mieghem et al. (2020) in their analysis of literature on inclusive education identified as one of the research gaps.
For the purpose of this study, and given the limited space, the focus is only on interviews with school personnel. Additionally, we make a use of insights the first author gained during the fieldwork, to when appropriate corroborate or question the interviewees' descriptions and arguments.
All interviews and fieldwork notes were transcribed and integrated into the Nvivo software to elicit analytical categories and subcategories. Subsequent reading and rereading of the categories and subcategories was undertaken against the backdrop of previous research, the concept of inclusion, and the study's objectives.

Separate classes for a 'safe start' in the Rose Garden school
The model used by the Rose Garden school could be labelled as separate classes. Although, in the school's local documents and guidelines, as well as in interviews with the school staff, inclusion emerges as an ultimate goal and, which is particularly interesting, as an overarching operational strategy. Both are succinctly summarised in the generic concept of 'safe start', recurrently highlighted by the school staff, encompassing the engineering of social environment (smaller classes, children with similar background, fostering dense relations); pedagogical interventions in terms of providing skilled teachers in second language acquisition, and multilingual classroom assistants; and bridging strategies between separate and mainstream classes.
Upon the arrival, all newly arrived children in the Rose Garden school are being screened, based on the material from National Agency for Education. The screening result is used to make decision in which class a child will be placed, although one of the school's priorities, as put by its assistant principal, is to make sure the children are placed in an age-appropriate class. Additionally, the results are used as a template to build subsequent pedagogical strategy tailored for individual needs of every child. These strategies are defined in individual study plans, by law required to be produced for all NAMS in the three final years of primary school, describing how a student will be enabled and supported to overcome obstacles to achieving final grades and eligibility to national educational programmes at upper-secondary level. As one teacher, who primarily work in separate classes, put it: We use the screening results when we plan our education. Because the students have diverse experiences, they are maybe not used to our way of working. Maybe they had learned a lot of text by heart, maybe that was important in their previous school, but now they ought to reflect more about the texts we study. We notice pretty quickly that some students are not used to this, and it's good if I as a teacher know why it sometimes goes as it does. Or why it takes more time in some cases. So, it's important to take the screening results into account when you plan your lessons. And that you have a competence to work with newly arrived, for example in second language acquisition.
The screening is thus utilised as a multipurpose tool that first and foremost equips the teachers with knowledge and understanding about who their new students are, what they need and what they already can. But, equally important is that the screening results, through intermediary function of individual study plans, are expected to propel changes and clarify responsibilities among all teachers in the school. In other words, the encompassing strategy is based not just on a compelling notion of what a newly arrived child has to do and improve, but also, what teachers ought to do and improve to meet their needs.
In the Rose Garden school all newly arrived children are initially placed in separate classes, but they are also assigned a place in a regular class. Maintaining relations between the two organisational entities and frequently overbridging them through social and pedagogical practices is one of the most salient distinguishing features of the school's inclusionary strategy against the backdrop of its per se exclusionary organisational model. And this is a significant move forward, since the previous system was organised around three self-contained entities. In the words of the assistant principal: We started three and a half years ago, when I became an assistant principal, we started to transfer them according to individual schedules. Before, they first attended a reception class during one year, and then a separate class for a year, and then out to mainstream class. Even though I worked here at the time as a teacher in mainstream class, I could be very surprised, it was poorly prepared when they would be transferred to mainstream classes. One day, a child who had been at school for two years and you barely seen him, just sat there in your classroom. And you ask: 'Hi, are you a student here?' Yes, ok. If you are going to be here, welcome. They were very isolated, they had their own activities, it worked very well, but it wasn't part of this school, it felt like they were their own island, that no one knew anything about, except a few teachers who directly worked with them.
Interviewer: Why did you carry out this reform?
Assistant principal: Mostly, so they will be part of this school. They must be included, that is good. Inclusion is, you know, kind of a strange word, that has been misused, but it is about making a priority based on individual needs.
Apart from getting a glimpse into the complexity of how the widely championed concept of inclusion can be perceived by the school staff ('a strange word', 'misused'), the assistant principal laid out the core problem with separate classes. They tend to be isolated ('their own island') from the rest of the schools, and their students are of little concern for other teachers and other students.
For many years, a powerful perception that the only way to meet educational needs of NAMS was in secluded separate classes, where they can be shielded off from the chaos of schools' everyday life, was dominant. Ultimately, it served to present and normalise exclusion as intrinsic to the only possible system. Furthermore, exclusionary practices that perpetuated the system and its outcome were misrecognised (Bourdieu 1991) as a necessary feature of what it means to cater for NAMS' educational needs. Keeping them in separate classes until they are 'ready' concerning the language and subject matter knowledge for mainstream classes, making sure they do not interact 'too much' with other children, thereby risking provoking traumatic flash-backs, focusing almost solely on second language acquisition, understood as 'a key to integration', are just some of these perpetuating exclusionary practices.
But, to reiterate the assistant principal, 'they must be included'. Instead of abolishing separate classes altogether with reference to isolation, something that a number of Swedish schools have done (Bunar 2015), including the other one in our empirical example (see next section), the Rose Garden school refurbished the system, by adhering to three principal strategies. The first one was to make sure newly arrived will be immediately transferred to mainstream class in those academic subjects where their attained level of knowledge in Swedish language was deemed to meet a minimum criterion. Transfer to mainstream classes is not a sudden or a definitive step, but rather incremental, erasing or at least blurring the previously impenetrable boundaries between separate and mainstream classes.
The second strategy was ensuring institutionally sanctioned cooperation between teachers from separate and mainstream classes is in place. This was achieved through a didactic collegium where teachers jointly, on a weekly basis plan their lessons.
The third strategy was to physically place newly arrived in a separate classroom in the middle of the school. The interviewed school staff claimed it to be a significant symbolic message, that these students are not something that should be 'hidden away' in a classroom at the end of corridor, or in a separate building, but rather made visible. The hope was also that this placement will mitigate informal meetings between the students in separate class and other students. Indeed, during the fieldwork, the first author could detect how both students and teachers from separate class regularly shared school spaces with other students and teachers, getting engaged in dense social relations.
Evidently, the separate class-model is here understood as a necessary prerequisite for achieving a 'safe start', although primarily through deploying pedagogical interventions, what Göransson and Nilholm (2014) label as pedagogical inclusion. Meaning, the opportunity to learn and acquire appropriate knowledge. Important to underline and based on interviews and informal talks with school staff during the first author's fieldwork, this model is beside being deemed as the only possible, widely accepted and recognised among school staff as timely and appropriate.
On the other hand, we have learned that the bridging strategies (also to a great part pedagogically justified), and physical placement of separate class are the main strategies for, to paraphrase Göransson and Nilholm (2014), social inclusion. Meaning overcoming obstacles for the newly arrived children's full participation in social relations with their non-newly arrived peers.
Notwithstanding the importance of these strategies, there is another circumstance, that more than school orchestrated interventions, promote peer relations. And that is the school's multicultural composition and its local context. Since virtually all of its students are of migrant background, with shared first language, country of origin and even ethnicity as the newly arrived, a sense of solidarity and friendship is being forged transcending the migration induced boundaries between newly arrived and other students. It is here that dense social relations between migrant families and their solidarity in the local community (Dahlstedt 2018) are penetrating social relations among children in school, alleviating social inclusion. Migration that causes institutionally imposed boundaries through a separate class-model is, in the context of this school, simultaneously helping to erase them.

Inclusion (for some) in the Carnation Garden school
In 2016, the newly appointed leadership at the Carnation Garden school announced that the old system catering for educational needs of NAMS would be changed, with inclusion as its leading principle. Inclusion was perceivedas reiterated in the interviews and informal talks with school staffas a product of two combining factors, which we briefly label 'the inclusion formula': direct immersion in mainstream classes and duly supported in second language acquisition. While the extent and the quality of the language support could fluctuate and only affect some students and teachers, direct immersion was a flagship of the changes, visible to and affecting virtually everybody in the school. What are the distinguishing features and the underlying principles of 'the inclusion formula' induced changes?
First, the old organisational model was dismantled. According to an interviewed teacher in Swedish as a second language (SSL), who together with a principal was an architect behind the reform, it was not even clear what the previous system was: separate classes, direct immersion or something in-between.
Teacher 1: I started here two and a half years ago, at the same time as our principal. After a couple of months, we realized there was a need for a reorganization of the reception. It was very messy. There was no clear read thread and it was even unclear whether we had separate classes or not. The teachers barely knew themselves, and I and the principal agreed that the students would directly go to ordinary classes. We didn't want separate classes and we wanted to have a clear way of working, who does what. To have it clear for ourselves and for our staff, they will know what it is, what it means to receive a newly arrived, who they are, whose responsibility it is. Both the principal and I firmly believe and are committed to our way of doing it. We chose it deliberately because we believe it's good, the best for our students and that it's clear, so everyone feels I'm interested and I want them to succeed.
The Carnation Garden school did not even attempt to overhaul and adjust the obviously ill-devised and poorly functioning system based on the separate class-model. Instead, the decision was made to throw it out altogether and replace with direct immersion. Inclusion, predictability and everyone's responsibility for newly arrived migrant students' learning were launched as a legitimate ground and the backbones of the new model.
The second aspect of 'the inclusion formula' was to expose NAMS to a Swedish language speaking environment. This environment was not deemed to exist in separate classes. Additionally, mainstream classes were conceived of as the only way to offer NAMS a cognitively challenging education.
Apart from general language acquisition strategy, through continuous exposure to language environment of mainstream classes, a more tangible strategy was to provide extended instruction in SSL. The school formed a team of three teachers in SSL that through dense cooperation with ordinary teachers in SSL provide support to both teachers and students. According to an ordinary teacher in SSL the cooperation with the SSL-team and their extended instruction works very well. However fruitful and well-functioning this cooperation is, a number of interviewed teachers pointed out its greatest weakness. There is virtually no cooperation between the SSL-team and other, subject matter teachers. The entire focus is on language learning and not on general learning. One particularly severe consequence thereof is that students are provided uneven instruction quality in different subjects. Some teachers are familiar with and interested in using content and language integrated learning (Gibbons 2013) and thereby able to reach out to their students with limited skills in Swedish, and other are not. Some teachers have been through professional development about what it means to teach in diverse classes, some have not.
A solution the school adheres to, in order to mitigate the variety of support provided by subject matter teachers, is the use of multilingual classroom assistance. The problem, as identified by interviewed teachers and corroborated during the first author's fieldwork, is that the municipality decided all newly arrived children, irrespective of their individual background and needs, are entitled to only one-hour of multilingual classroom assistance per week. The primarily reason was financial. Notwithstanding its importance, even if it is one hour, this decision goes against all research-based recommendations (SOU 2019:18) that newly arrived children cannot be treated as a collective, but rather as a heterogenous group.
Given all these circumstances a theoretically sound model based on an attractive formula is facing unsurmountable obstacles. But, its most formidable 'adversary' emerges from internal opposition. The majority of school staff, interviewed and engaged in informal talks during the first author's fieldwork, and even some from the school management, are opposing the current system with direct immersion in ordinary classes.
Teacher 3: I don't like it, I don't like to have a student that just has arrived to Sweden jumps into a classroom where they [teachers] expect everybody speaks Swedish. I think that's wrong, it slows down their learning, it slows down my work as well. I think they must have a separate class for themselves and focus on that and on Swedish.
Teacher 4: I found it incredibly tough, frustrating, that I didn't have time to help them and I didn't even know how to help them. It's difficult, I don't know, I thought I probably didn't have skills enough and then I heard from other teachers that they also find it tough. And I believe I see among some students, they think too it's tough.
A multilingual classroom assistant put it in an even more dramatic way, arguing that 'sometimes it feels like a child abuse, because they are so different' (see also Juvonen 2015). The question is then, who benefits from the model and who is defending its existence? The principal and a rather small group of teachers in SSL, who found the model tough from the beginning, but with a significant pay-off after a while, are the primary apologetics.
Teacher 1: It's a challenge, but especially with students with strong school background we achieve great results. Both development of language skills and subject matter knowledge goes very fast. And of course, we get a lot of support from multilingual classroom assistants.
Interviewer: And how does it go for students without proper educational background?
Teacher 1: Hm, I mean, yes, that's different, I mean a bigger challenge for us, they can feel a bit lost in the classroom.
In other words, the model is devised to meet educational needs of NAMS with proper educational background and knowledge in English, enabling initial communication, only. The biggest losers in the practical implementation of the Carnation Garden school inclusive model are NAMS with limited school background and those without any knowledge in English.
As observed by the first author during fieldwork, informal peer-groups were not primarily formed on the basis of newly arrived versus non-newly arrived scale. Even this school is placed in a multicultural community where family ties based on shared first language and ethnicity lay significant ground for social relationships among students.

Conclusions: the construction of the meaning of inclusion and its implications
The article has provided an insight into how two Swedish primary schools address educational needs of newly arrived migrant students. Both schools present inclusion as a defining principle of their (divergent) organisational models, deployed measures, and ultimately as a prerequisite for students' academic achievement. It could be argued that both schools are responding to growing diversity among their students, with respect to new arrivals, by adhering to inclusion. Given the differences in their approaches, the generic questions we posed were: Do both 'get it right'? What are the distinguishing features, and the underlying principles of inclusion the schools are building their models around? How is the meaning of inclusion constructed in the processes of its practical implementation in these two schools?
Inclusion is a comprehensive reform incorporating a holistic approach (Pinson, Arnot, and Candappa 2010) and support (Bunar 2019a) that timely and properly responds to diversity and diversity of needs amongst learners (Ainscow and Messiou 2018). A truly inclusive educational reform always takes its point of departure in a careful consideration of how to meet the needs of every newly arrived migrant child. Needless to say, and in line with Allan and Slee's (2019) arguments, this approach rejects segregated schooling 'on whatever grounds'.
One aspect of the construction of the meaning of inclusion is the way the school staff understand and deal with the nature of challenges they are about to address, having NAMS in their classes. We have learned from previous research (Uptin, Wright, and Harwood 2012;Bunar 2015;Migliarini, Stinson, and D'Alessio 2019) that one significant aspect of that understanding is a deficit-based perspective, meaning NAMS are not recognised as competent and resilient, but almost exclusively seen as a problem and with problems, especially for subject matter teachers in mainstream classes. Participating in ordinary lessons is not regarded as a cause for adjusting the content and form of the teaching, but rather as a disturbing moment affecting the learning environment of all students. Or as one interviewed teacher in the Carnation Garden school put it, 'it slows down my teaching as well'. As evident from our material, the deficit-based perspective is present in both schools, but it is not the only one. A number of examples, in particular from the Rose Garden school, demonstrated that the national legislation requirements, in form of the mandatory screening, individual study plans and the need for bridging strategies between separate and mainstream classes, are positively affecting what the local practices should and must 'include in inclusion'.
Another aspect of the construction of the meaning of inclusion is the way school staff understand the prevalence of and position themselves in relation to three interrelated contextual perspectives: (a) the contextual opportunities, i.e. available resources and their distribution, professional development possibilities, the presence of multilingual personnel, (b) contextual requirements, i.e. whether professional development in content and language integrated learning, working in diverse classes, and working together with parents are mandatory, and (c) contextual limits, in terms of how flexible inclusion is perceived to be, how much can it be bent and still called and accepted as inclusion in the local context. Do recall, the schools are obliged to follow the national legislation purporting to provide a template for inclusive education.
In the case of the Rose Garden School and the Carnation Garden School this aspect of the meaning of inclusion has been constructed based on generally weak contextual opportunities, weak contextual requirements and very flexible contextual limits. Together with the earlier mentioned deficit-based perspective, and the incorporation of national legislation demands, thus constructed meaning has been translated into a set of local policy actions and their justifying narratives, summarised as 'safe start' and 'the inclusion formula'. Following Migliarini, Stinson, and D'Alessio (2019) it could be argued that 'safe start' and 'the inclusion formula' exemplify the gap between international human rights law and normative national policy advocating inclusion education, on the one hand, and the exclusionary school practices, on the other hand.
The policy actions and narratives of 'safe start' and 'the inclusion formula' are very broad, enabling them to accommodate a wide range of present progressive practices and local policies. But they also appear to be able to absorb critical perspectives raised by badly needed but absent, practices and policies. The inclusion through exclusion model of the Rose Garden School is justified by the need for a 'safe start'. Fully including some students into mainstream classes from the beginning is not even considered. One assistant principal, interviewed in the study, even found inclusion a misused and strange word. On the contrary, as it turned out, the exclusion through inclusion model of the Carnation Garden School is justified by the need to eradicate the previously poorly functioning model with separate classes. Very few other reforms have been conducted to support direct immersion. Additionally, there is a broad disagreement among the staff in this school about the appropriateness of the general model. The opposite voices are not demanding more inclusive reforms, but on the contrary, the return of segregated schooling.
Evidently, the ambitious tale of inclusion in both schools was, in the process of the construction of its meaning and implementation, reduced to some of its aspects. Teachers and school administrators are allowed to include and leave out of their model whatever they deem necessary, obsolete, expensive or unrealistic and still fitting under the umbrella of inclusion. Sometimes it works, sometimes it does not, and both 'get it right' and 'wrong' in some aspects.
In order to reach beyond the arbitrariness of the models' configuration and operations, and to promote inclusion and attainment, the construction of the meaning of inclusion ought to start with and maintain focus on individual needs of every child, in an environment with strong contextual opportunities and requirements as well as with defined boundaries of inclusion. Not everything is inclusion, just because someone claims it to be. The holistic approach, rejecting segregated schooling, and supportbased inclusion are some promising theoretical perspectives worth considering building a sustainable policy on.

Notes
1. See Budginaitė et al. (2016) regarding similar results in international research. 2. Migrant students is a broad category encompassing all students that for various reasons have migrated to Sweden: as children to labour migrants, for family reunion or fleeing wars and persecution in their native countries. Refugee students are also migrants per definition, who have applied for protection in the country, but their life circumstances are often much more dire than other migrant categories (see Bunar 2019b for a deeper discussion on legal and practical definitions and their implications). 3. Approximately, 73,000 children under the age of 18 applied for asylum in 2015. 4. In the city of Stockholm, the local government decided that no newly arrived migrant students will be enrolled in any of the 16 most multicultural public schools in order to prevent further segregation.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Denis Tajic is a PhD candidate at Department of Child and Youth Studies, Stockholm University, Sweden. His ongoing research is about inclusion of newly arrived migrant students in primary schools and the role of various organisational and support models the schools are deploying to meet educational needs of these students. Denis has previously worked as a multicultural classroom assistance for newly arrived migrant students and as a head of the Language development unit at Södertälje municipality.
Nihad Bunar is a professor at Department of Special Pedagogy, Stockholm University, Sweden. His main field of research is education and migration. He has an extensive publication track record in Swedish and English. He has also served as a special investigator for Swedish government writing a law proposal on the use of first language and language support teachers for migrant children. Nihad has also worked as an expert and advisor for the European Trade Union Committee for Education (ETUCE) and the European Federation of Education Employers (EFEE); Organization for Security and Cooperation in Europe (OSCE) on preparing training manual for education of refugee and asylum-seeking students; European Commission, DG Education and Culture; Network of Experts on Social Aspects of Education and Training, NESET I and NESET II. Some of his latest publications include: Bunar, N. 2019a. Promoting effective integration of migrants and refugees in education: experiences from Spain, Serbia and Belgium. Brussels: Education International Research.