What is meant by inclusion? On the effects of different definitions on attitudes toward inclusive education

ABSTRACT

Aiming to further our knowledge about what is meant by inclusion, we examined how various conceptualisations relate to people's attitudes about inclusive education. We assign the varying characterisations of inclusion of specific groups with differing involvement in the education system in Luxembourg, applying an influential systematisation of definitions of inclusion in which four qualitatively different categories are identified. Results of study 1 showed that members of the general population, pre-service and in-service teachers perceive inclusive education in importantly different ways. Although results showed relatively positive attitudes toward inclusive education for the whole sample, attitudes varied by group and in relation to the differential categorisation of definitions. As teachers' attitudes and the extent to which they feel prepared to implement inclusive practice are crucial for the success of inclusive education, the latter aspect is further investigated in study 2. Results showed that teachers with more in-depth understanding of inclusive education reported more positive attitudes and felt better prepared to implement inclusive practices. Implications for education systems and society are discussed.

KEYWORDS:

• Inclusion
• special educational needs (SEN)
• conceptual analysis
• public attitudes
• teacher attitudes
• Luxembourg

Introduction

What is meant by inclusion? Different definitions of inclusive education abound, problematising education research, reforms and implementation of practices. In 1994, inclusion was presented as a promising new pedagogical approach at the World Conference on Special Needs Education in Salamanca (UNESCO 1994). Building on the Salamanca declaration, the global movement of 'Education for All' (UNESCO 2000), states that the inclusion of all children in regular schools is the most effective way to counteract discriminatory attitudes and achieve the goal of inclusive education. In the decades since, inclusive education has become an intensively discussed and empirically investigated concept and pedagogical reform project around the world, even as education systems differ considerably in both their classification and inclusion rates (see Richardson and Powell 2011). Mandated by the 2006 UN Convention on the Rights of Persons with Disabilities (UN-CRPD; UN 2006) that has been ratified by more than 175 countries, inclusive education has become a truly global norm (Powell, Edelstein, and Blanck 2016).

Nonetheless, inclusive education is by no means a clearly defined or universally understood concept, demanding more in-depth, culture-specific research on understandings of inclusion. Among the contributions that attempt to systematise this persistent diversity, the typology of definitions by Göransson and Nilholm (2014) has been prominently discussed. In their analysis of studies published between 2004 and 2012, Göransson and Nilholm recognised that, despite the vitriolic debates about inclusive education, most reviews overlook fundamental differences in definitions of inclusion, especially regarding the goals to be attained, which reflect the range of contexts and education systems. Although their conclusions concerning empirical and methodological shortcomings of inclusion research have been criticised (see, e.g. Dyson 2014; Haug 2014), their provision of a typology of definitions to enable the systematic evaluation of inclusive practices in different contexts has been welcomed. More specifically, Göransson and Nilholm (2014) synthesised key interpretations of inclusive education, leading to four qualitatively different categories (definitions): (A) Placement definition: inclusion as placement of students with disabilities or in need of special support in general education classrooms; (B) Specified individualised definition: inclusion as meeting the social and academic needs of students with disabilities or in need of special support; (C) General individualised definition: inclusion as meeting the social and academic needs of all students; (D) Community definition: inclusion as the creation of communities with specific characteristics. These diverse definitions have considerable consequences for inclusive educational research, reforms and practices, hence warranting empirical analysis. Moreover, without explicitly defining inclusion and embedding this research in specific contexts, scientific progress in identifying successful practices is unlikely.

In a more recent review, Nilholm and Göransson (2017) concluded that definitions of inclusion broadly used in the literature still continue to lack conceptual clarity. This impacts research findings concerning both the attitudes toward inclusion and the effectiveness of inclusive practices. Inclusive education will be effective when supported by all parties involved, including teachers, students, their parents and the broader community. Indeed, positive attitudes toward students with special education needs (SEN) and their participation in inclusive classrooms are essential (Antonak and Livneh, 2000). When societal attitudes are positive, they facilitate inclusion, furthering acceptance of students with SEN by peers, families and teachers (Morin et al. 2013). Therefore, we explicitly investigate attitudes towards inclusive education in relation to differential definitions of inclusion. We do so in one of Europe's most diverse societies with extraordinary migration and mobility and a multilingual and stratified education system. Also, the concept of inclusion has become a keyword in political and public discussions. In Luxembourg, educational policies encourage schooling of students with SEN in general schools with external support (Limbach-Reich and Powell 2015). Recent reforms implemented to promote inclusion have encouraged team-teaching and differentiation within schools, with schools expected to develop inclusive education development plans (Powell and Hadjar 2018). However, the Ministry of Education continues to support special schools accommodating students with different types of SEN (MENJE 2018). Hence, the implementation of 'inclusive education' reforms has varying elements and the concept remains open for interpretation. This is reflected in teacher training programs in which the next generation(s) of teachers – who must carry out these reforms within schools – are educated. Challenges of definition are a key reason why inclusive education has been implemented inconsistently. Neither in educational policymaking nor programming has a consensus of goals, even after ratification of the UN-CRPD, been achieved (see Powell, Edelstein, and Blanck 2016).

Thus, to improve our understanding of inclusive education and its definitions, our research addresses how crucial groups of actors with varying experiences in the education system characterise inclusive education. We classify the views of these groups according to the typology of Göransson and Nilholm (2014). Authors have argued that the ideas connected with a particular approach towards inclusion influence people's attitudes towards it (see Göransson and Nilholm 2014; Kruse and Dedering 2018). Ultimately, differential beliefs and attitudes will lead to different (teaching) behaviours (Fazio and Roskos-Ewoldsen 2005), hence partially explaining how and why teachers make certain choices in planning and implementing lessons to accommodate a heterogeneous student population. Our first study considered the concept(s) and definitions of inclusion in relation to attitudes towards inclusion in specific settings. We specifically considered definitions of inclusion and attitudes toward inclusive education of different actors in the school system. In applying the typology, we note that we only classify those definitions related directly to the education system (A, B and C), excluding definition D, which applies to the community (Göransson and Nilholm 2014). Furthermore, in the first study, we focused on people's attitudes as reflected in the perceived academic and social benefits of inclusion for students with and without SEN because the UN-CRPD (UN 2006) recommendations not only concern students' academic achievement but also their social relationships and well-being (Ainscow and César 2006). Our second study considered definitions and attitudes of experienced teachers in particular, as they are the key persons to actively create inclusive learning environments (Borg et al. 2011) and hence are largely responsible for implementing inclusive education programs in practice (Meijer, Soriano, and Watkins 2003). We therefore also measured the extent to which teachers felt prepared to implement inclusive practices to accommodate diverse students.

Research design study 1

Questions and methodological design

The first study examines how people view inclusion and if the type of definition affects their attitudes toward inclusion. Respondents were asked to present their own understandings of inclusion in writing. Using people's written responses concerning inclusion may provide a broader view of people's opinions compared to (standardised) questionnaire data and hence may better reflect different understandings of inclusive education in the Luxembourgish context. Written responses may also allow for the detection of differences in perceptions based on differential involvement in the education system (participating vs. professional). In addition, participants were asked to complete a questionnaire on attitudes toward inclusive education.

Data and methods

The participant sample of our first study consisted of 237 people from the Luxembourgish general population (158 female) with a mean age of 29 years (SD= 11.87), 88 pre-service teachers (68 female) with a mean age of 24 years (SD= 3.28) and 45 in-service teachers (32 female) with a mean age of 35 years (SD= 7.17). The members of the general population have had relatively recent experiences as former students as well as potentially being family members of current students. Participants were asked to complete the survey as part of a larger project investigating attitudes toward inclusion in Luxembourg (Pit-ten Cate 2014). Informed consent was obtained for all participants, and ethical guidelines were followed throughout the study.

Defining inclusion

We asked the participants to define 'inclusion' in their own words ('How do you define inclusion?'). The responses were classified according to the Göransson and Nilholm (2014) typology, whereby keywords were used to facilitate the categorisation process:

1. Placement definition – inclusion as the placement of students with disabilities or in need of special support in general education classrooms

   • Integration or inclusion of students with SEN;

   • Students with and without SEN go to the same school or classroom.

2. Specified individualised definition – inclusion as meeting the social and academic needs of students with disabilities or in need of special support

   • Students with SEN can actively take part in the lessons;

   • Students with SEN get (individualised) support in mainstream classrooms (to achieve their learning goals);

   • Students with SEN benefit from instruction in mainstream classrooms;

   • Students with SEN are respected/valued.

3. General individualised definition – inclusion as meeting the social and academic needs of all students

   • All students actively take part in the lessons;

   • All students receive (individualised) support (to achieve their learning goals);

   • All students benefit from instruction in regular classrooms.

All the definitions were coded independently by five raters. Interrater agreement ranged from 81% to 88%. Disagreements were resolved through discussion. The length of the answers ranged from three to thirty-five words. Examples for answers include 'Children with disabilities go to a regular school' (Category A); 'Children with disabilities actively take part in lessons taught in regular classrooms' (Category B); and 'All students attend their neighbourhood schools in regular classes and are supported to learn, contribute and participate in all aspects of the life of the school' (Category C).

Attitudes

Participants completed the German version of the 'Opinions Relative to Integration of Students with Disabilities questionnaire' (ORI; Benoit and Bless 2014), whereby we replaced the word 'integration' with 'inclusion' in keeping with contemporary parlance. For the purpose of this study, we used only two subscales: (1) Benefits of Inclusion for Students with and without SEN (e.g. 'When children with special educational needs are taught in a regular class, the benefits for the other students more than outweigh the potential difficulties of this practice') and (2) Educational and Social Progress of the Included Students (e.g. 'Regular class teaching offers significantly better learning opportunities for children with special educational needs than a small or special school').

Items were scored on a Likert scale ranging from 1 (fully disagree) to 6 (fully agree). In accordance with the stipulations of the survey authors, subscale scores above the mean of the scale (i.e. 3.5) were interpreted as indicative of more favourable attitudes towards inclusion. Benoit and Bless (2014) reported good internal consistency in their sample of teachers in Switzerland (i.e. Cronbach's alphas of .87 and .84 for the Benefits and the Educational Progress subscales, respectively). In our sample, internal consistency was slightly lower but still acceptable (.73 and .78, respectively). In order to investigate the extent to which people perceive inclusion to be of benefit, we provided an introductory paragraph in which we defined inclusion according to the placement definition of Göransson and Nilholm (2014): 'The following questionnaire asks for your opinion on the topic “Inclusion of children with special needs”. Inclusion means the joint education of all children, regardless of their physical, behavioural or learning disabilities. In 2011, Luxembourg ratified the United Nations Convention on the Rights of Persons with Disabilities, thus committing to the education of students with and with special educational needs in mainstream classes.'

Results

Defining inclusion

The percentage distribution of the three definitions of inclusion for the total sample (see Figure 1) shows that the participants most frequently (43%) understand inclusion according to the narrower 'placement definition' (Category A), while 17% followed the 'specified individualised definition’ (Category B), focusing on the support of students with SEN in general classrooms. In total, 40% understood inclusion according to the 'generalised individual definition' (Category C), relating the concept of inclusion to the support of all students. These results emphasise that the concept of inclusion is understood differently – as within international and national academic discourse – amongst those with various links to the education system in Luxembourg. Furthermore, the majority of participants (60%) related the concept of inclusion specifically to students with SEN (category A and B) whereas 40% widen the concept to include all students (category C).

Figure 1. Percentage distribution of understandings of inclusion for the total sample (N= 370).

Next, we considered the definitions of inclusion in relation to the group membership of the participants (participating vs. professional involvement in the education system). As shown in Figure 2, definitions varied as a function of group membership. More specifically, participants showed significant differences in their understandings of inclusion depending on their attained function within the education system. In comparison with pre- and in-service teachers, members of the general population mostly consider inclusion as a concept relating to the placements of students with SEN. In contrast, pre- and in-service teachers indicated that they understand inclusion primarily in accordance with the 'general individualised definition' (Category C) that relates to pedagogical practices. Interestingly, in-service teachers (i.e. participants currently implementing inclusive practice to varying degrees) defined inclusion significantly more often in terms of this category than did pre-service teachers, who are still preparing to pursue their teaching careers (χ² (2, N= 125) = 5.84, p= .04). Emphasising the importance of pedagogical training, people that have only participated in the education system primarily view inclusion in terms of placement – the education of students with SEN within rather than outside general classes – whereas people with professional involvement define inclusion primarily in terms of supporting the needs of all students, implying the creation of inclusive learning environments.

Figure 2. Percentage distribution of understandings of inclusion according to category of definition and group. χ²(4, N= 370) = 119.04, p < .001, Cramer's V = .41.

Attitudes toward inclusion by participant group

To investigate between-group differences in attitudes, we conducted a mixed method ANOVA with ORI-subscale (Benefits of Inclusion vs. Educational and Social Progress) as a within-subject factor and GROUP (general population, pre-service teachers and in-service teachers) as the between-subject factor. The results revealed a significant main effect for ORI-subscale, F(1, 368) = 65.65, p < .001, η² = .16, indicating differences in scores for the two subscales. A t-test analysis revealed that, overall, participants' ratings of the Benefits of Inclusion were higher than their ratings of the Educational and Social Progress, t(369) = 5.81, p < .001, d = 0.30. Additionally, results showed a significant main effect of GROUP, F(2,367) = 40.10, p < .001, η² = 0.18, indicating that there were differences in attitudes between participant groups. Additional t-tests revealed that in-service teachers held significantly more positive attitudes than pre-service teachers, t(132) = 2.16, p = .03, d = 0.45 or the general population, t(281) = 6.78, p < .001, d = 1.35. In turn, pre-service teachers held significantly more positive attitudes than the general population, t(324) = 6.16, p < .001, d = 0.78 (see Figure 3). This finding was further signified by a significant interaction effect of ORI-subscale x GROUP, F(2, 367) = 22.71, p < .001, η² = .11. More specifically, in-service and pre-service teachers provided higher ratings on the Benefits – than on the Progress-subscale, t(44) = 7.34, p < .001, d = 1.27 and t(87) = 9.95, p < .001, d = 1.06, whereas, people of the general population rated both scales equally, t(236) = 0.91, p < .001, d = 0.07 (see Figure 3).

Figure 3. Attitudes toward inclusion as a function of participant group.

Attitudes toward inclusion in relation to the categorisation of definitions of inclusion

To investigate the relationship between the definition of inclusion and attitudes, we conducted a mixed method ANOVA with ORI-subscale (Benefits of Inclusion vs. Educational and Social Progress) as a within-subject factor and CATEGORY (definitions: Category A, B and C) as the between-subject factor. The results revealed significant main effects for ORI-subscale, F(1, 368) = 35.08, p < .001, η² = .09 and CATEGORY, F(2, 367) = 27.31, p < .001, η² = .13. This effect of CATEGORY indicated differences in attitudes for people providing different definitions of inclusion (see Figure 4). More specifically, people defining inclusion according to the 'General individualised' definition held significantly more positive attitudes than the people defining inclusion according to the 'Specified individualised' definition, t(203) = 2.76, p < .05, d = 0.21 and to the 'Placement' definition, t(316) = 6.51, p < .001, d = 0.84. Additionally, people defining inclusion according to the 'Specified individualised' definition held significantly more positive attitudes than the people according to the 'Placement' definition t(217) = 4.00, p < .001, d = 0.64.

Figure 4. Attitudes as a function of the categorisation of definitions of inclusion.

Discussion

The first study focused on the Göransson and Nilholm (2014) typology of definitions of inclusion and surveyed Luxembourgish participants with either a participatory or professional involvement in the education system. Findings showed, in line with previous research (Kruse and Dedering 2018), that participants with differing involvement in the education system perceived inclusion in dissimilar ways. For members of the general population, inclusion primarily concerns the school placement of student with SEN, whereas for pre- and in-service teachers, inclusion reflects the necessity of changing teaching practices to accommodate an increasingly heterogeneous student population. These differences reflect not only the different levels of involvement and experience but also the emphasis on structures (where students are educated) in contrast to the more pedagogical orientation (how are students educated) among teachers. People who participated in schooling may or may not have experienced the inclusion of students with SEN in their own schools or in schools/classrooms of their friends or children. Additionally, they may have followed political debates that have primarily focused on the inclusion of students with learning difficulties and challenging behaviour in general classes (MENJE 2017). In contrast, teachers develop their views of inclusion, especially in terms of the implementation of inclusive practices, based upon training courses or experience teaching students with various SEN. Luxembourg is notable for its hyper-diverse student body, with students of different cultural, ethnic and linguistic backgrounds included in the regular school system for many decades. This diversity has long required teachers to adapt their teaching accordingly. Simultaneously, and following contemporary education policy reforms, teachers are often concerned about their ability to accommodate students with specific learning or behavioural difficulties (Pit-ten Cate and Krischler 2018). Such concerns reflect that teachers often consider the practical implications of inclusion, further illustrated by the finding that none of the in-service teachers in this sample conceptualised inclusion in terms of placement (Category A).

Next, we considered the association between these definitional categories and different groups' attitudes toward inclusion. Results revealed that people defining inclusion according to the 'General individualised' definition (mainly teachers) held the most positive attitudes, followed by those defining inclusion according to the 'Specified individualised' definition and the 'Placement' definition (mainly general population members). It could be that given their professional involvement, teachers consider the broader context of education reforms, which reflect contemporary debates concerning the ultimate aims of inclusion. These aims reflect not only the attempt to reduce educational inequalities (Watkins 2012) and hence may incur a broader definition of inclusion, but also focus on the academic and social benefits for students with and without SEN (Ainscow and César 2006; Szumski, Smogorzewska, and Karwowski 2017). As a result, their attitudes toward inclusion in terms of benefits and student progress may be more positive. Similarly, people with only participatory involvement in schooling may perceive inclusion primarily in terms of the most obvious change in the system – mandated by the UN-CRPD – namely the education of students with SEN in general rather than special settings. Placement in general schools on the basis of a legal right may however not automatically incur positive attitudes, even if the Luxembourgish educational system has traditionally not been as segregated as in the neighbouring countries of Belgium, France and Germany (Limbach-Reich and Powell 2015; Powell and Hadjar 2018).

Investigating attitudes toward inclusive education is motivated by the idea that attitudes can predict later behaviour (Glasman and Albarracín 2006). Consequently, positive attitudes towards inclusion are seen as a crucial condition for inclusive education (Avramidis and Norwich 2002; de Boer, Pijl, and Minnaert 2011). Especially teachers' beliefs and attitudes are critical in ensuring the success of inclusive practices since teachers' acceptance of inclusive education policies is likely to affect their commitment to implementing it (Norwich 1994). In study 2, we therefore investigated the definitional typology in relation to attitudes and readiness for inclusive education in a sample of experienced teachers.

Research design study 2

Questions and methodological design

In study 2, we examined how experienced teachers defined inclusion in terms of Göransson and Nilholm's typology (2014). We further investigated the relationships between the definitions and teachers' attitudes. Given the relationships between attitudes and (intended) behaviour (Fazio and Roskos-Ewoldsen 2005) and because the educational success of inclusion is largely dependent on the expertise and the willingness of teachers to attend to the needs of all students (Lindsay 2007; Watkins 2012), we also investigated teachers' self-reported readiness for inclusive education.

Data and method

The second study's sample consisted of 42 in-service teachers (39 female) with a mean age of 36 years (SD= 6.88). All of them reported having experience with students with SEN in their classrooms. Teachers completed the survey as part of a larger study concerning teachers´ competence and attitudes toward inclusive education in Luxembourg (Pit-ten Cate 2014). All participants provided informed consent and ethical guidelines were followed throughout.

Definitions of inclusion

We asked the in-service teachers to define 'inclusion' in the context of education in their own words. Again, we gathered open written expressions to investigate qualitative differences in the conceptualisation of inclusion.

Attitudes

Participants completed the German version of the Questionnaire on Attitudes towards Inclusion for Teachers (EFI; Seifried and Heyl 2016). The EFI consists of two student-focused subscales: 'Promoting academic competencies in inclusive classrooms' (e.g. 'Children with special needs would ultimately receive better support in an inclusive classroom') and 'Social inclusion in school' (e.g. 'I estimate the changes in the social climate through the inclusion of children with special needs as positive'). In our opinion, the items and measured construct of these two subscales are quite similar to those of the ORI-subscales used in Study 1. In addition, the EFI contains one teacher-focused subscale: 'Individual readiness for inclusive education' (e.g. 'I can imagine teaching in an inclusive class in the coming school year'). Items were scored on a Likert scale ranging from 1 (fully disagree) to 6 (fully agree). Average subscale scores above the mean scale score of 3.5 were used to indicate a more favourable attitude toward inclusion. Seifried and Heyl (2016) reported good internal consistency in their sample of teachers (i.e. Cronbach's alphas of .90, .85 and .81 for the Promoting competencies, Social inclusion and Individual readiness subscales, respectively). In our sample, we found quite similar values (α = .88, .87 and .80, respectively). As in study 1, we provided an introductory paragraph in which we defined inclusion according to the placement definition before teachers were asked to complete the questionnaire.

Results

Definition of inclusion

We used the same operationalisation of the typology of definitions as in Study 1. Interrater reliability for the categorisation ranged from 79% to 85%. The distribution of the three categories of definitions of inclusion (see Figure 5) shows that most teachers (67%) characterised inclusion according to the 'general individualised definition', conceptualising inclusion in terms of meeting the academic and social needs of all students. In addition, 10 respondents (24%) employed a concept according to the 'specified individualised' definition and only 4 (9.5%) used the 'placement' definition.

Figure 5. Percentage distribution of understandings of inclusion according to the category of definition (N= 42).

Relationship between the category of definitions and attitudes toward inclusion

To test for the influence of definition CATEGORY on EFI-subscale ratings, we conducted a mixed-method ANOVA with EFI-subscale (promoting competencies vs. social inclusion vs. individual readiness) as a within-subject factor and CATEGORY (Placement vs. Specified individualised vs. General individualised) as the between-subject factor. The results revealed significant main effects for EFI-subscale, F(2,40) = 12.32, p < .001, η² = .39 and for CATEGORY, F(2, 40) = 8.84, p = .001, η² = .31. The main effect of EFI-subscale reflects mean score differences for the three subscales. Post-hoc paired t-test revealed that ratings on the social inclusion scale were significantly higher than ratings on the individual readiness subscale (t(41) = 5.90, p < .001, d = 0.88) and the promoting competencies subscale (t(41) = 7.29, p < .001, d = 1.10). Ratings did not differ between the individual readiness and the promoting competencies subscales, t(41) = 0.75, p = .46, d = 0.10 (see Figure 6).

Figure 6. Teachers´ attitudes as a function of their definition of inclusion.

Concerning the main effect of CATEGORY, attitude ratings varied as a function of definition of inclusion. Teachers defining inclusion according to the 'General individualised' definition held significantly more positive attitudes than teachers defining inclusion according to the 'Specified individualised' definition, t(203) = 2.76, p < .05, d = 0.21 or the 'Placement' definition, t(316) = 6.51, p < .001, d = 0.84. Additionally, teachers defining inclusion according to the 'Specified individualised' definition held significantly more positive attitudes than the teachers that gave a 'Placement' definition t(217) = 4.00, p < .001, d = 0.64 (see Figure 6). The interaction effect of CATEGORY and EFI-subscale was not significant (p = .60).

Discussion study 2

The second study focused on the typology of inclusion definitions (Göransson and Nilholm 2014) among experienced teachers in Luxembourg. Findings showed that attitudes varied as a function of categorisation (i.e. Category C – the general individualised definition – was associated with significantly more positive attitudes than Categories A and B). Most importantly, the individual readiness to teach in inclusive classrooms was only positive for those teachers defining inclusion according to the 'general individualised' definition. Hence, teachers concerned about the needs of all students rather than seeing inclusion as primarily including or meeting the needs of students with SEN show more willingness to implement inclusive practices. Consistent with previous findings, we show that teachers who have an open perception about inclusive education are more confident in their own abilities to implement it (Buell et al. 1999).

In general, teachers showed positive attitudes toward the social inclusion of students with SEN, and neutral attitudes toward the promotion of academic competencies of all students. Additionally, they showed a moderate individual readiness to implement inclusive teaching. This finding is in line with previous research, stating that most teachers express positive attitudes toward inclusive education in general (Avramidis and Norwich 2002; de Boer, Pijl, and Minnaert 2011), but only moderate willingness to teach in inclusive classrooms (Lautenbach and Antoniewicz 2018; Mousouli et al. 2009). Mousouli and colleagues (2009) concluded that many teachers support inclusion theoretically, but still do not fully understand the key ideas and their implications. While they may well be against discrimination and therefore understand the aim of inclusion, they may consider students with SEN as forming a defined, separate group and therefore be concerned about how to organise inclusive schooling and teaching, whereby they may not realise that their own attitudes and willingness to change might be more important than methodological recipes or school restructuring.

General discussion

This contribution presented two studies of how crucial groups of actors in the education system – people in the general population who have participated in schooling in Luxembourg, pre-service and in-service teachers – characterise inclusion. Their definitions reflect the main contemporary conceptualisations of inclusion. Their attitudes toward inclusive education relate to these varying understandings. More specifically, although it has been implicitly assumed that definitions of inclusion are related to attitudes, our findings provide empirical evidence of the existence of this relationship. Our findings demonstrate the importance of clarifying how inclusion is defined and applied in different studies in order to better interpret differences in empirical findings relating to inclusive education (see also Nilholm and Goransson 2017).

Results showed that inclusion remains an ambiguous concept, with definitions that vary according to a person's experience involved in the education system, whether as a (past) participant or as professional. The general population conceptualises inclusion primarily according to student placement. In this regard, Felder (2018) emphasised that different understandings of inclusion have varying implications and consequences for action. If inclusion is merely conceived as students with and without SEN being taught in the same school building, then the task of education reforms of policy and practice would require reallocations and redistributions of resources that would allow for the placement of students with SEN in general schools. While this definition secures the right of students with SEN to attend general schools, it does not mean that each and every student will have access to and be supported in actively participating in the (school) community. If, however, inclusion is seen as an active form of participation of every student in a common learning frame and in diverse classrooms – as defined by most pre- and in-service teachers – the task would rather be to facilitate respectful interpersonal relationships and equality of provision (Felder 2018). Successful inclusion is about valuing diversity and encouraging differentiation rather than uniformity and curriculum commonality (Ainscow et al. 2006). Education has a key role in facilitating the acceptance of diversity and utilising varying strengths and perspectives.

As Luxembourg has one of the most heterogeneous populations in Europe accompanied by a relatively low rate of segregation (only 1% of all students attend special schools), the inclusion of students with differing needs seems to be culturally endorsed and promoted at least by those directly involved in the education system. This is emphasised by our results implying that especially experienced teachers have a more in-depth understanding of inclusion, more positive attitudes toward it, and more willingness to implement inclusive practices.

Among the implications are the need to sensitise the general population about the human right to inclusive education, and diverse definitions and understandings thereof, by discussing its meanings and its potential consequences for all students. Good practice examples could be presented to diverse communities to clarify that schooling should support the learning progress of every student, whatever their background and abilities. Moreover, positive teacher attitudes toward inclusive education could be promoted further in the context of an education system that provides conditions necessary for good practices in this field. Recently, adaptations have been made in response to the ratification of the UN-CRPD in 2011. The newest education law, foresees further changes in Luxembourg's education system: Curriculum restructuring (e.g. focus on individualised learning processes and social skills); more help should be made available for diverse learners from support teachers (e.g. members of multi-professional teams); and more time should be allocated for preparing (inclusive) educational activities and creating and developing opportunities for partnerships between teachers, students, multi-professional teams and parents (MENJE 2017, 2018). Yet, with its hyper-diverse multicultural and multilingual population, Luxembourg still faces considerable challenges in equalising learning opportunities. Hence, curricular reforms would also need to consider teacher training. Teachers need to acquire knowledge of inclusion and its diverse principles and consider broader issues, such as migration background, language(s) spoken at home, socio-economic status and gender, which are closely linked to the identification of special educational needs (classification) and the recommendation to attend special schools (allocation) (see Limbach-Reich and Powell 2015; Limbach-Reich, Powell, and Brendel 2015). The research presented here for the first time surveyed different groups in Luxembourg about their ideas concerning inclusive education and their attitudes toward such transformation of education systems. Our finding that the definition of inclusion is related to attitudes provides empirical support for a notion that has often been implicitly assumed. This finding also highlights the need in research, policymaking and educational practice to provide a clear definition of the constructs that are investigated or discussed (for both research and implementation purposes). This is particularly the case in multilingual Luxembourg as all schools must develop their own plans to develop inclusive education.

Limitations

Some considerations of our study's limitations, which could be addressed in future research, are warranted. Participants were recruited as part of a larger project. The general population sample represented a relatively young group, which had recent experience in the education system. Although this may have ensured participants were familiar with current educational practices, they may not be representative of the Luxembourgish society as a whole.

Similarly, pre-service and in-service teachers were recruited based on voluntary participation and therefore may have resulted in a sample for which educational reforms and especially the drive toward more inclusive schooling was of particular interest.

The concept of 'attitudes toward inclusion' is complex, and the measurement of these attitudes is no simple task. The ORI (Benoit and Bless 2014) and EFI (Seifried and Heyl 2016) questionnaires are among the most frequently used instruments over the last few years that reflect good psychometric properties. Even though the ORI questionnaire was explicitly designed to investigate teacher attitudes, this questionnaire has already successfully been used in a study involving people from the general population (Krischler et al., 2018). Nevertheless, several other questionnaires exist that could be used to investigate (public) attitudes toward inclusion (for reviews, see Antonak and Livneh, 2000; Rubergand Porsch 2017).

Furthermore, responses to the explicit measure may be affected by social norms (De Houwer 2006), whereby the positive attitudes toward inclusion may in part have resulted from a desire to conform to the social desirability standard (i.e. as the inclusion of all students is defined as a human right, participants may have felt it would no longer be socially acceptable to openly speak up against this movement). Given the social sensitiveness of the inclusion debate, it may, therefore, be relevant to complement the use of self-report measures with measures of implicit attitudes that are less prone to social desirability (Fazio and Olson 2003; Houston and Fazio 1989).

Conclusion

Our results show that important differences in the conceptualisation of inclusion exist. These differing definitions are also important when considering people's attitudes toward inclusive education policy and practises in various cultural contexts. Different conceptualisations affect not only inclusive policies and practices within the education system but rather illuminate the positioning and production of disability in society as a whole.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The study was funded by the Fonds National de la Recherche (FNR), Luxembourg, grant number [C14/SC/7964914/INCLUS▀].