School district administrators' perspectives on special education policy and practice in Norway and Sweden

The purpose of this study was to examine special education policy and practice from the perspective of school district administrators in Norway and Sweden. Administrators from 266 Norwegian and 262 Swedish municipalities completed a survey concerning: (a) reasons children need special education, (b) common and desired organizational solutions, and (c) the influence of policy on practice. Despite a number of clear differences, findings suggest that Swedish and Norwegian administrators share similar attitudes regarding the provision of special education support. It appears that in both countries inclusive practices are seen as the ideal, yet, Norwegian administrators appear to have a stronger preference for categorical or segregated solutions. However, this finding must be viewed in light of current practices in each country. In particular, we take into consideration data indicating that 17% of Swedish students receive special educational support, as compared to approximately 6% in Norway.

. At the same time, national policies focusing on management by objectives (MBO) or 'goalsteering' have ensured that the state still controls the most important aspects of the system (Moos, Mö ller, and Johansson 2004). In recent years, the state has further regained some of its earlier control, yet the basic approach remains the same: the state steers by setting goals while local communities are responsible for independently finding solutions for ensuring that these goals are met (Bjö rklund et al. 2005;Smehaugen 2007).
As in other regions of the world, the marketization of systems of education has also posed a challenge. Through market forces the idea of a single 'school for all' that also includes children with special needs has been contested (Ferguson 2008). This trend is largely defined by two inter-connected factors: (a) school choice, and (b) increased competition often under the auspices of accountability. In Sweden, the former is demonstrated by the emergence of private charter schools and the opportunity to choose between community schools. In Norway, although there are few private schools, parents may choose to send their children to any number of public schools located within a limited region. Unfortunately, there is little evidence concerning how this type of school choice affects the composition of schools and the services provided to children with special needs. Nonetheless, in some regions there does appear to be an increase in the use of organizational solutions that are similar to the previously common practice of using special schools for educating children with disabilities. That is, resources and special education expertise are pooled together (e.g., special departments, alternative programs) to support concentrated groups of children with similar diagnoses (Egelund, Haug, and Persson 2006;Solli 2005).
The trend towards increased competition and an emphasis on accountability also raises questions concerning the education of children with special needs. One possible line of development is that outcomes and efficiency will come to dominate the discussion. In this case, we are also likely to see the re-emergence of older organizational solutions involving training in basic skills and an expansion of segregated services.
As the pressures on Scandinavian school systems evolve, it is important to investigate the changing situation for pupils with special needs. Moreover, comparisons of different national systems enable the search for uniformities in educational provision across borders and offer the possibility of learning lessons from the different approaches and solutions applied in different contexts (Osborn et al. 2003). National policies intended to regulate how, when, and to whom special education services are provided form the legal basis for the approaches applied in a given national context. The Swedish and Norwegian policies are briefly described in the following section together with some of the challenges inherent in these policies.

Special education policy in Norway and Sweden
The terminology used to describe special education and the children who receive it tends to vary across nations (Meijer 2003;Pijl, Frostad, and Flem 2008). We have chosen to use the terms 'special education' and 'special support' interchangeably in this article, as they most accurately reflect the terms used in the Norwegian and Swedish education laws (described below). From a policy perspective, each of these terms has a legal definition serving several purposes. For the individual child, guidelines serve to guarantee a number of rights or protections, such as: (a) the right to an assessment of individual educational need, (b) access to an individualized curriculum (e.g., IEP), and perhaps most important, (c) a right to additional resources and services (e.g., 'extra' small group teaching). In addition, 'special support' and 'special education' are used to describe both the administration of additional resources for groups of pupils defined to be eligible for such support, as well as the 'special' teaching provided to these students. The latter of these, in particular, is difficult to evaluate Á as indicated by the continuing debate regarding special education's effectiveness (e.g., Cook and Schirmer 2003). We have therefore chosen to focus our study principally on processes tied to identifying children in need of special education and the organizational solutions that are applied in providing this support.
In Norway and Sweden less than 2% of children attend special schools (Solli 2005; Swedish National Agency for Education [SNAE] 2003). Approximately 6% of children in Norwegian schools receive special education as defined by the Norwegian Education Act (Solli 2005). In contrast, 17% of all Swedish students receive this type of support (SNAE 2003). This sizable difference can presumably be attributed to substantially differing laws, regulations and procedures guiding the provision of services to children with special needs in the respective nations. In Norway, for example, evidence suggests that a number of children receive special education without having received a formal statement of eligibility (Dalen 2006;Grøgaard, Hatlevik, and Markussen 2004) and are therefore not included in the Norwegian statistic. In contrast, many of these 'gray zone' children are presumably included in the percentage of children receiving special support in Sweden. Other possible reasons for the sizable difference in special education prevalence include: (a) the Norwegian emphasis on adapted teaching (described below), (b) the Norwegian government's steering of the Educational Psychological Services (EPS) towards system level interventions (Anthun and Manger 2006), and (c) the fact that there are a greater number of large urban areas in Sweden than in Norway. Indeed, evidence indicates that higher percentages of children are identified for special education in urban than in rural regions, and children with minority backgrounds are overrepresented among this population (Blanchett 2009;GSI 2010;Solli 2005).

Sweden
Some children in the Swedish school system are eligible for alternative school forms (e.g., special schools) based on a medical diagnosis of disability (e.g., developmental disabilities, deaf/hearing impaired). Almost 2% of the school-aged population attends these schools. In addition, a large portion of children (ca. 17%) are identified as having special educational needs within the ordinary comprehensive school (SNAE 2003). There are also substantial within-country variations in the percentage of students who receive such support; varying on the basis of grade level, geography, and a range of other demographic factors (SNAE).
Special support for these students is defined by the goals of the comprehensive school and, thus, constitutes an educational, rather than medical need. The importance placed on recognizing the problem as educational, rather than one that is localized solely within the child, was notably reflected in the change of the term 'children with special needs' to 'children in need of special support' in the Swedish School Act (Svenska fö rfattningssamling [SFS] 1985[SFS] , 1100. The school ordinance defines four types of special support: (a) native language support, (b) remedial tuition, (c) special tuition, and (d) adapted curriculum. Remedial tuition is by far the most common and, therefore, encompasses the vast majority of children at the centre of our investigation. Children are entitled to remedial tuition under the following conditions: If there is reason to believe that the pupil will not reach the minimum level goals by the end of the fifth and ninth school year or if there are other reasons why the pupil needs support (SFS 2000(SFS , 1108. Thus, according to the law, identification of children with special needs in the Swedish comprehensive school is determined primarily in relation to problems of goal attainment. Of course, the law provides substantial room for interpretation with respect to what goals are at stake (e.g., social, academic) and at what point prior to the fifth or ninth year the child is at risk of not reaching these goals. Children defined as being in need of special support are formally identified in the context of school and receive an individual educational plan (IEP) aimed at addressing the specific difficulties they are experiencing within that context. In addition, the Swedish school ordinance stipulates that organizational solutions in the form of segregated instruction should be avoided whenever possible within the comprehensive school system.

Norway
Over the past several decades the percentage of students in Norwegian schools receiving their education full-time in separate facilities has remained relatively stable, at less than 1% of all pupils (Solli 2005). The few publicly funded special schools are primarily for deaf/hearing impaired students, children with severe emotional disturbance, and a small number of schools for children with multiple disabilities. Less then 2% of Norwegian children attend private schools (Statistics Norway 2008). While all students in Norway have a right to attend their neighbourhood school, families may also choose to send their child to another public school within the municipality. Thus, certain schools may receive higher percentages of pupils with special needs as parents seek out schools with more resources or expertise.
According to Norwegian law, students who 'are not receiving or cannot obtain satisfactory benefit from ordinary tuition have the right to special education' (NMER 1998, para. 5). When it appears that a child is not receiving a 'satisfactory benefit from ordinary tuition', the school is required to apply for an expert evaluation to determine eligibility for special education support. The evaluation is typically conducted by the municipality-based EPS, which is staffed by psychologists, special educators and other educational professionals. Based on this evaluation, a child may be guaranteed the right to special education and an IEP. Thus, as in the Swedish system, the identification process is contingent on an assessment of performance relative to educational goals. Also similar to Sweden, there is significant variation within Norway regarding the percentage of students identified for special education from region to region and across schools and age groups. It is widely assumed that these variations are to some degree attributable to differing interpretations of special education policy (NMER 2004). In particular, it would appear that determining whether a child is receiving 'satisfactory benefit' is a difficult criterion to apply reliably.
Another important factor with regard to special education provision in Norway involves the principle of 'adapted teaching' (NMER 2004). For more than 20 years this concept has been an element of the national curriculum, where it is presented as a basic principle for the education of all children. Bachman and Haug (2006) describe 'adapted teaching' as a politically constructed concept that expands on the ideas of individualized and differentiated instruction and is coupled with the principle of inclusion. The Norwegian Education Act states that all pupils should be provided 'adapted teaching' to meet their individual needs within their local school (NMER 1998, para. 1Á2).

Purpose of the study
The purpose of this study was to examine and compare school district administrators' perspectives on special education policy and practice in Norway and Sweden. In addition, we sought to summarize and compare the forms that special support takes in each country and the importance placed on the different approaches reported. We argue that this comparative approach provides a useful means of examining policy and practice by providing a better understanding of the manner in which national policies are recontextualized at the level of the municipalities. The concept, 'recontextualization' implies that policy is interpreted and transformed with regard to local circumstances rather than being implemented in a straightforward manner. Thus, we suggest that there are substantial degrees of variation at lower levels of the policy chain. In this way, our study strives to compare the recontextualization of what has been taken to be similar national policies of Sweden and Norway centring on the concept 'one school for all', a concept which has been challenged in both countries in latter years. The relevance of the study to policy is thus that it makes it possible to explore and compare what happens with overarching policy notions when they are played out at lower levels of the educational system and, most important, the consequences for children in need of special support. The comparison between Sweden and Norway provides a unique opportunity to compare how one overarching idea becomes recontextualized in two different school systems and even to question if these systems still can be seen as an expression of a similar idea.
However, there are also some interesting differences between the two countries in educational policy as regards children in need of extra support. As described above, the concept of adapted teaching, suggesting that everyday education should be tailored to the needs of all pupils, has been very influential in the Norwegian context. On the other hand, Norway has a more centralized support system. In Sweden, SENCOs (Special educational needs coordinators) have a large responsibility for special educational support and are most often placed in the schools. Moreover, SENCOs are supposed to work with the whole learning environment of the children. Thus, there are differences between the two countries that might have implications for how special education practices are realized despite the common notion of 'one school for all' (see Emanuelsson, Haug, and Persson 2005, for a more in-depth discussion of these differences).
One means of improving our understanding of these phenomena is by examining the perspectives of school district administrators who hold key roles in determining the manner in which special education policy is both interpreted and implemented. These professionals are among the best situated to provide a general picture of common practices across multiple schools within their districts. The value that administrative leaders place on different practices can also provide an indication of the perceived benefits and constraints associated with different organizational solutions. From a broader perspective, comparative research of this nature serves to highlight the distinctive regional features of wider (e.g., Scandinavian) trends and contributes to the validity of interpretations concerning the origins of features that are unique to a given region. To guide us in this process, we developed the following overarching research questions: (a) What differences and similarities can be seen in the manner in which special support, as reported by school district administrators, is provided in Norway and Sweden? (b) What perspectives appear to dominate local school administrators' interpretations of policy and practice in each country?

Analytical perspective
We argue that one can discern two quite different views on special education, termed the categorical and relational perspectives (Persson 2003). The basic assumptions of these different views are summarized in Table 1. The model takes its point of departure in two 'ideal types' (Weber 1949). These ideal types represent polarized perspectives, or one-dimensional points of view, not to be found in reality. As such, they can be understood as describing two poles at either end of a continuum that can be used to examine special education approaches that more or less reflect a given one-dimensional perspective. Emanuelsson, Haug, and Persson (2005) used these two perspectives as a theoretical basis for their analysis of Norwegian and Swedish special education. The perspectives can be applied to, among other things, the ontology of special needs, approaches to difference, provisions of support, understanding of special education competence and reasons for the existence of special needs. Within a categorical perspective children's difficulties can be explained by individual diagnoses describing pathological 'deviations' and with the use of 'special' provisions as the main educational strategy. In contrast, the relational perspective refers to special needs as social constructs in which inclusive provisions form the main basis of education. In Table 1. A categorical and relational view of special needs (Persson 2003, 277).

Categorical perspective
Relational perspective

Ontology of special needs
Special needs refer to actual characteristics of individuals this study, the two ideal-typical perspectives provided a useful analytical tool and theoretical framework for designing the investigation and interpreting findings. Based on this model, we propose that municipalities indicating a dominance of the categorical perspective are more likely to: (a) perceive children as needing special support because of their individual shortcomings, (b) view the use of medical diagnoses as important for providing special support, (c) report preferring and using organizational solutions such as separate instruction and individual assistants in the classroom, and (d) rely on steering documents with individualistic perspectives. On the other hand, a comparatively more relational perspective would be indicated by: (a) the attribution of educational problems to factors outside of the child, (b) downplaying the importance of medical diagnoses, (c) organizational solutions such as extra teacher resources, and (d) a reliance on steering documents with more inclusive approaches.
In connection with this approach, it is important to distinguish between the formulation and realization of political intentions. Furthermore, it is important to note that this study deals with how participants in the two countries understand and formulate special educational provisions taking place in their schools. It is therefore likely that their answers reflect a mixture of both intentions as well as ongoing practice.

Survey instrument
As part of a previous investigation in Sweden (Nilholm et al. 2007), researchers designed a survey of Swedish municipalities' practices in which the following themes were addressed: (a) distribution of resources, (b) identification of pupils, (c) personnel practices, (d) organizational solutions for provision of support, (e) influence of policy, (f) influence of different stakeholders, (g) prevention and (h) outcomes. The survey instrument contained 30 questions grouped on the basis of these themes (see Nilholm et al. 2007 for a description of the survey). The theoretical model describing categorical and relational perspectives (Persson 2003) was used as a framework for designing questions with forethought to later analysis. The questionnaire was then translated into Norwegian and reviewed independently by three researchers in the field of education in Norway. In this study, we examine findings pertaining to five questions chosen for their explanatory power in gauging responses in accordance with our theoretical approach and overarching research questions: (1) What are the most common reasons for children to need special support?
(2) How relevant are medical diagnoses for the provision of special support?
(3) What organizational solutions are most common for providing special support? (4) What organizational solutions are most desired for providing special support? (5) What guiding documents are important in working with children with special needs?
Questions employed a 4-point likert scale (e.g., 1 0 'not at all common', 4 0'very common'). The number of rateable items per question ranged from 1 to 9. Participants were also given the option of marking 'no opinion/do not know'.

Participants
The head school district administrator (i.e., municipal supervisor for compulsory schooling; superintendent) in each municipality was asked to complete the survey. It is clear that the Norwegian (n 0266) and Swedish (n 0262) administrators who completed the survey have varying professional roles and responsibilities from region to region, both within and between the two countries. However, from a broad perspective, the position's duties tend to be fairly consistent. These duties include management of school districts' budgetary resources, supervision of school personnel, and ensuring that education is provided in accordance with applicable law. Participants were instructed to consider all questions in relation to public schools for all compulsory grade levels. It is important to note that there are substantial variations in size across municipalities in both countries and between the countries themselves. The largest Norwegian school district comprises an urban population of 538,411; the smallest municipality is in a remote region of the country with a population of only 214. In contrast, the largest municipality in Sweden has a population of 758,311 and the smallest has 2652 inhabitants. The average population of municipalities is approximately 10,800 in Norway and 31,000 Sweden. The total population of Norway is approximately 4.6 million inhabitants, whereas Sweden has a population of over nine million.

Data collection
The survey was distributed to all municipalities in Norway (431) and Sweden (290). Response rates were 61.7% and 90.3% for Norway and Sweden, respectively. In Norway, the surveys were mailed to municipalities in three consecutive rounds from September 2006 to June 2007. Data collection in Sweden was completed by January 2007 and involved two separate mailings. Swedish municipalities who did not respond to the surveys by mail were contacted by telephone and asked again to participate. More resources were available for data collection in Sweden than in Norway. This allowed the researchers to devote additional work-hours to contacting Swedish municipalities by telephone. This procedure was also deemed less feasible in Norway given the larger number of municipalities in that country. The different procedures used in each country may account for the higher response rate in Sweden than in Norway.

Analysis
As this study is both exploratory and comprehensive (all municipalities in both countries were asked to participate), our analysis is descriptive. Survey items were entered into a separate database for each country and analysed using SPSS software. All calculations are based on valid responses; missing responses and answers of 'no opinion/do not know' are not included in the analysis. In Norway, one item had a substantially lower response rate than all other items in either country: 47% of Norwegian participants marked 'no opinion/do not know' when asked to rate the importance of the Salamanca Declaration. Interestingly, this item also had the lowest number of responses in the Swedish survey (85%). Excluding this item, the percentage of valid responses for the remaining 31 items (across five questions) ranged from 83.1% to 98.5% of the total sample.
We chose to collapse ratings of 'very' and 'somewhat common' in order to offer a general indication of commonly reported practices and for ease of comparison. Percentages for these items are presented graphically (Figures 1 and 2) and descriptively in the following section. In contrast to perceptions of 'commonness', we conjecture that ratings of 'desirability' and 'importance' are more indicative of personal judgements or beliefs. As such, it is important to consider both strong negative or low ratings (e.g., not at all desired) in combination with very positive or high responses (e.g., highly desired). Therefore, in order to provide an estimate of the overall strength and direction of these beliefs, we calculated means and standard deviations for these items (Tables 3 and 4). To facilitate comparison, we then applied cut-offs to indicate high, moderate and low levels of perceived importance and desirability. Low mean ratings are those falling below the theoretical neutral (2.5) on the 4-point scale. We also provide summative data based on percentages for these questions in the following section. Responses to the question pertaining to the relevance of a medical diagnosis were concise enough that findings for all levels of the scale are presented (Table 2).

Reasons for requiring special support
Participants rated six potential reasons for a child to need special education support (Figure 1). In Norway, 97.3% and in Sweden, 74.0% of municipalities reported pupils' 'individual deficits' to be a common reason (i.e., 'very' or 'somewhat' common) for children to need special support. Similarly high ratings were reported for the item, 'school's goals are too difficult' (81.2% and 74.6% of Norwegian and Swedish municipalities, respectively). 'Difficulties in the home environment' was also rated as a common reason in a relatively high percentage of Swedish municipalities (76.7%); whereas Norwegian participants' ratings on this item were slightly lower (62.7%). The perception that 'classes function poorly' and that 'teachers were underqualified' were less often reported as common reasons for requiring support in Norway (35.6% and 20.3%, respectively) than in Sweden (39.0% and 48.7%, respectively). The greatest difference between the two countries was with respect to the item 'schools are poorly adapted to handle differences', for which 34.2% of Norwegian participants and 72.4% of Swedish participants rated this as a common reason for needing support. Survey participants were also asked about the relevance of a medical diagnosis (Table 2) for providing special support. The majority of municipalities in Norway (73.4%) indicated that a medical diagnosis was 'somewhat' or 'very' relevant for determining the provision of special support. This percentage was substantially lower amongst Swedish participants (37.5%).

Common organizational solutions
Figure 2 presents ratings for common organizational solutions. Five items were more frequently rated as being common (i.e., 'very' and 'somewhat') than uncommon (i.e., 'not very' and 'not at all') in both countries. These include: (a) separate group instruction for less than 50% of the school day (Sweden 0 72.6%; Norway 085.7%), (b) extra teacher resource in the classroom (Sweden 058.3%; Norway 0 96.6%), (c) general education instruction with advising from a special educator (Sweden 080.6%; Norway 090.3%), (d) use of assistants in general education (Sweden 069.1%; Norway 095.5%), and (e) weekly individualized instruction with a special educator (Sweden 0 83.1%; Norway 0 74.9%). Swedish municipalities reported a slightly higher use of 'segregated instruction for more than 50% of the school day' (22.2%) and 'integration in special groups' (12.4%) than did Norwegian municipalities (18.0% and 11.4%, respectively). In contrast, 'adjustment of class size' was more frequently rated as a common solution in Norway (33.1%) than in Sweden (19.6%). With the exception of two items, reported differences between the Norwegian and Swedish responses were less than 13.5%. These two items were 'extra teacher in the classroom' and 'assistant in the classroom' for which the difference between the two countries was 38.3% and 26.4%, respectively. Table 3 presents means and standard deviations for desired use of organizational solutions in the two countries. Average ratings were categorized into three groups based on the following cut-offs: (a) low desirability 0 B2.5 (below theoretical neutral), (b) moderate desirability 02.5Á3.0, and (c) high desirability 0 3.0. In general, respondents from both countries indicated similar responses on this question. There were two clear exceptions to this finding. First, 'use of assistants in the classroom' was rated more desirable in Norway than in Sweden, where only 29.6% of respondents rated this option as desirable (i.e., 'very' or 'somewhat') compared with 94.4% of municipalities in Norway. A sizeable difference was also found regarding the 'adjustment of class size' item, where 75.4% of Swedish municipalities rated this as a desirable solution, in contrast to 49.8% of Norwegian respondents. In both countries, 'use of separate groups 50% of the day' and 'integration in special groups' (i.e., groups of pupils with similar difficulties integrated in general education together) were rated as having low average desirability. However, a larger percentage of municipalities rated 'use of separate groups for greater than 50% of the school day' as desirable in Norway (27.6%) than in Sweden (10.7%). In addition, whereas both countries average ratings for desirability of 'separate weekly special education instruction' were in the moderate range, findings indicate a greater desirability for this solution in Norway than in Sweden. Norwegian participants rated this as a desirable option in 79.3% of municipalities as compared to 60.7% of Swedish respondents.

Importance of policy documents
Means and standard deviations for participants' ratings of the importance of various policy documents are presented in Table 4. Both in Norway and Sweden, school district administrators gave high ratings for four of the six document types connected to within-country policies. Greater than 95% of ratings for these four items in both countries were coded as important (i.e., 'very' or 'somewhat'). The only exception to this finding was with respect to 'local school policy', for which a smaller percentage of Norwegian (78.3%) than Swedish participants (96.0%) rated this item as important. The importance of 'state guidelines and reports' was rated higher in Sweden than in Norway, with respectively 84.8% and 56.1% of participants rating this item as important. In contrast to these findings, considerable differences appeared in comparisons of ratings for the importance of international policy documents. All three of these items received higher ratings in Sweden than in Norway. Whereas the percentage of respondents reporting these documents to be important in Sweden ranged from 85.3% to 95.7%, responses on the high end (i.e., 'very' and 'somewhat important') of the scale were infrequent among Norwegian responses (range 24.6% to 38.6%).

Discussion
The purpose of this study was to examine the perspectives of school district administrators concerning the education of children with special needs in Norway and Sweden. In addition, we sought to identify differences and similarities between the two countries with respect to participants' responses about the form that special education takes and the value placed on different approaches. It is perhaps this second objective that is most open to criticism, as a major difficulty of international comparative research is determining the true source of differences and similarities observed (Artiles and Larsen 1998). In investigations of this type, the breadth of contextual factors suggesting alternative explanations for findings is almost without end. It is also clear that participants' ratings of policies and practices are not a direct measure of actual practices. Bearing this in mind, we have tried to make sense of these data in a cautious and reasoned manner.

Interpretations
Despite a number of clear differences, our findings suggest that Norway and Sweden share many approaches to the manner in which special support is provided. From an international perspective, the Swedish and Norwegian educational systems are considered fairly inclusive. That is, in Norway and Sweden segregated solutions in the form of special schools, special classes or other forms of segregated instruction are uncommon compared to other countries (Meijer 2003;Vislie 2003). Indeed, findings from this study support the conclusion that schools in Norway and Sweden strive to accommodate learning differences to a large extent within general education settings and that solutions consistent with a relational perspective are both practiced and valued. However, our findings also show that these countries rely on solutions that are more indicative of a categorical perspective as well, and although these solutions are not necessarily seen as ideal, they are considered important.

Reasons for requiring support
Both Swedish and Norwegian participants rated categorical explanations for children to need support (i.e., individual deficits, difficulties at home) as being fairly common. However, as compared to their Norwegian counterparts, Swedish participants placed more emphasis on relational explanations for students' difficulties. One exception to this finding is that the item, 'school's goals are too difficult', was frequently reported as being a common explanation for providing support in the Norwegian sample (81.2%). It is possible that participants interpreted this statement as meaning: 'the school's goals are too difficult [because the child has individual deficits]', thus, reversing the polarity of the ideal type that the item was intended to represent. This interpretation is reinforced by the rather low ratings in the Norwegian sample given for other relational explanations of special needs (e.g., schools are poorly adapted, teachers are under-qualified) and the fact that 'individual deficits' was the most commonly cited reason for providing support. In addition, Norwegian participants placed significant emphasis on the importance of a medical diagnosis for providing support. Thus, at first glance it appears that a categorical perspective for ontology of special needs is more prevalent in Norway than in Sweden. However, we speculate that this finding may be attributable to the different populations that were the subject of surveys in each country. In Sweden, 17% of students receive special support in the mainstream school system, whereas in Norway this group comprises only about 6% of children.
Approximately 2% of Norwegian children (ages 0Á15) receive social benefits due to disability, injury, or illness (Statistics Norway 2009), and prevalence estimates suggest that a significant portion of these children are likely to have been diagnosed with an intellectual disability (Lorentzen 2008). Thus, we estimate that as many as 1 out of 3 Norwegian students receiving special support also have a medically diagnosed disability or condition. If this estimate is accurate, it is possible that Norwegian participants simply placed more emphasis on medical diagnoses because children with severe or multiple disabilities, who are identified early in life and require long-term medical support, comprise a larger portion of the students that they had in mind when answering survey questions.
Participants in both countries were less likely to attribute the need for extra support to poor teaching or lack of classroom accommodations than other factors. School administrators quite logically may be less critical of their own role or the role of teachers in preventing children from requiring special support. Presuming that they believe schools and teachers have done their best with the resources they have been given, they may reason that the root of the problem is connected to factors outside of the educational system itself. Swedish participants, however, appeared to be much more critical of teachers' qualifications than were Norwegian participants. Nearly half of Swedish participants rated teachers' lack of qualifications as a common reason for children to need special support, as compared to just 20% of Norwegian municipalities. Due to the larger populations of Swedish municipalities, Swedish participants may have been more removed from actual practice, and/or had weaker personal ties to individual schools or teachers, and therefore felt that they could be more critical of school practices without directly criticizing their own performance.

Organizational solutions
Participants' responses concerning organizational solutions were similar between the two countries. Although separate teaching arrangements were common, findings suggest that the use of these solutions is not without restraint. That is, while periodic, segregated solutions (e.g., separate instruction B50% of the school day; weekly special education) were commonly reported, only about one in five municipalities in either country rated the use of separate special classes for more than 50% of the school day as a common approach. Furthermore, these arrangements generally apply to a very small group of children (Solli 2005). In addition, approaches consistent with a relational perspective, such as special educators advising classroom teachers, also appear to be frequently employed in both countries. Thus, we conclude that Norway and Sweden frequently engage in a mixture of practices that lean towards both categorical and relational perspectives.
In general, organizational solutions associated with a relational perspective received higher ratings of desirability in both countries than did those associated with categorical approaches. It is noteworthy that the two categories with 'moderate' average ratings in both countries (i.e., weekly special education instruction; separate group B50% of the school day) were reported as being among the most common organizational solutions. Despite their frequent use, these procedures appear to be less desired compared to other solutions; yet they are clearly not rejected altogether. We surmise that, given their high ratings of desirability, participants see relational solutions as the ideal (e.g., multiple teachers, special educators acting as advisors) but view some segregated solutions (e.g., intermittent special education classes) as necessary and, indeed, acceptable.
With respect to common approaches, the greatest difference between Norway and Sweden concerned the use of extra teachers and assistants in the classroom. In both cases, these practices were reported as being less common in Sweden than in Norway. When we examine average ratings for desired solutions, the most noticeable differences were in the areas of adjustment of class size and, once again, use of assistants. These data suggest that use of assistants in general education is a more desired practice in Norway than in Sweden, whereas the opposite appears to be the case for adjustment of class size.
Over the past several years, Norway has moved towards more flexible ways of organizing instruction, including multiple teacher models (Dalen 2006). National requirements concerning the size of classes were dissolved in 2003, giving schools considerable flexibility with respect to grouping arrangements (NMER 2004). It was argued that schools should be free to use extra resources on groups of children and subject areas that were especially demanding (NMER). Findings indicating that the use of instructional assistants and multiple teachers are both highly desired and frequently practiced in Norway suggest that these approaches play an important part in how Norwegian schools organize instruction. This may explain why multiple teacher solutions are more commonly reported in Norway than in Sweden and why there appears to be less concern for adjusting class size among Norwegian participants. As noted previously, these varied approaches may also help to explain the large difference in percentage of students found eligible for special education in the two countries.
It is interesting that while extra teacher resources are highly desired in both countries; the use of instructional assistants received much lower ratings of desirability in Sweden than in Norway. Whereas Norwegian participants appear to view these two approaches as complementary, educational leaders in Sweden may see them as being in conflict with one another. That is, Swedish participants may see themselves as facing a choice between using resources to either increase the number of instructional assistants or bring in more qualified teachers, but not both. Given that choice, it would not be surprising if they chose the latter.
In summary, findings concerning the use of special educators and multiple teacher solutions suggest a belief in the merit of relational solutions in both countries. In Sweden, a strong desire to adjust class sizes is further evidence of support for inclusive efforts. Evidence indicating the importance placed on instructional assistants in Norway is somewhat more difficult to interpret. If assistants in Norwegian classrooms are tied to individual pupils they may represent a largely categorical solution, which may in fact result in segregation from other children (Snaefridur and Rannveig 2009). Nonetheless, support via instructional assistants rather than traditional, segregated classes suggest an effort to integrate students with special needs. Thus, a desire to use instructional assistants is not easily definable as a purely categorical perspective. As is the case with much of these data, the choice of organizational solutions cannot simply be described in terms of inclusion or non-inclusion; rather different solutions appear to be used to handle dilemmas of individual student differences given particular national and local contexts.

Policy documents
Participants' ratings of the importance of various policy documents were similarly high for policies that address education on the national and local levels within each country. However, a striking contrast between the two countries emerged in comparisons of international documents, such as the Salamanca Declaration and the United Nations Convention on the Rights of the Child. Average ratings concerning the importance of these documents were much higher among Swedish participants than among their Norwegian counterparts, who indicated that these policies have little relevance for their work with children with special needs.
The Salamanca Declaration, in particular, is an international agreement with a strong emphasis on inclusive principles (UNESCO 1994). It is unlikely that Norwegian participants' low ratings of relevance for this document denotes a rejection of these principles, as the Norwegian Education Act and National Curriculum also place significant importance on inclusive goals. Rather, we speculate that Norwegian administrators' responses can be attributed to a lack of familiarity with the international documents themselves, and/or the assumption that these policies have limited impact on the day-to-day services provided to children with special needs in their communities. Indeed, the fact that 47% of Norwegian participants indicated 'no opinion/do not know' when asked to rate the importance of the Salamanca Declaration supports this conclusion. Given the relative size difference of municipalities in the two countries, Swedish participants may simply hold a more politically and/or internationally oriented position than do Norwegian educational leaders. Though the reasons are unclear, these data suggest that where international policy is concerned, local Norwegian school officials are either more isolated or more introverted than are their Swedish neighbours.

Implications
A site of tension in all educational systems concerns the distribution of power between the state, local leadership and schools. In the Scandinavian context the state has sought to regain control of education from the local authorities since the reforms of the 1990s (Moos, Möller, and Johansson 2004). In part, this development appears to be connected to the government's confidence in the benefits of competition, accountability and efficiency. The increasing focus on evidence-based practices and accountability through testing can be understood as one way in which control over local practices can be achieved. In both countries, this phenomenon is further evidenced by the outcry over poor performance on international assessments (e.g., PISA) and the recent emphasis on basic skills in educational reform (NMER 2004;Skolverket 2007). Thus, we might expect several changes which are intended to make the system more competitive. One conclusion being that efforts to accommodate and include students with special needs will be sacrificed in order to achieve higher outputs (Meijer 2003).
Seen in this light, evidence that local school administrators' strive to adapt school environments for children in need of extra support are reassuring. To the degree that desired practices can be interpreted as the direction of future practice, both countries may be headed towards more inclusive solutions involving flexible grouping arrangements, multiple teachers, and a more collaborative role for special education professionals. Indeed, this conclusion is in keeping with the trend towards inclusion that is visible on a global scale (Ferguson 2008;UNESCO 1994).
At the centre of this study are pupils who, by definition, are perceived as demonstrating unsatisfactory performance in relation to predetermined educational goals and standards. On the one hand, the tendency for Norwegian participants, in particular, to attribute the need for support primarily to children's individual deficits may imply a lack of attention to the role of schools in shaping or failing to prevent learning and behavioural problems. On the other hand, the findings that Swedish participants frequently reported poorly adapted schools (72.4%) and teachers' lack of qualifications (48.7%) as a common reason for children to need special support may suggest that many Swedish students may be experiencing unnecessary failure and exclusion as a result of schools' failure to accommodate these children.
In some ways our findings suggest that the international discourse of inclusion has a stronger hold on the educational establishment in Sweden than in Norway. Evidence from this study pertaining to the reasons that children need special support and the value placed on various organizational solutions bolster this conclusion. Given the global movement towards inclusion, perhaps this is connected to the apparent isolation or lack of awareness of Norwegian participants with respect to international policy. However, this pattern also has to be understood against the backdrop of the different contexts, and perhaps more importantly, current practices in each country. Taking these factors into account, another interpretation is worthy of consideration.
It is possible that Norway, at least at the organizational level, already has in place many practices that are sought after by participants. Indeed, extra teacher resources, adjustments of class size and special educators working with general educators were all more frequently rated as common practices in Norway than in Sweden. At the same time, the emphasis in Norwegian policy on flexible teaching, grouping arrangements and adapted teaching (Bachman and Haug 2006;NMER 2004), suggest a strong ideological support for inclusive practices. Moreover, Sweden's more frequent use of separate placements (e.g., special schools) and larger proportion of students identified as in need of special support also suggest a greater prevalence of organizational solutions leaning towards the categorical perspective.
Assuming this interpretation is accurate we can infer what it would imply from a broader theoretical perspective. It seems likely that the value placed on categorical explanations and preferences for segregated solutions such as those seen in Norway, increase when fewer children are identified as needing special support. In other words, as the proportion of children receiving special education decreases, the learning difficulties of this group will naturally become more extreme Á or different Á from the norm. Conversely, as the proportion of students receiving special education grows, as the Swedish case illustrates, the desire to attribute difficulties to relational explanations (e.g., the failure of schools to accommodate difference) and seek out more inclusive solutions, may increase.

Limitations and suggestions for future research
In comparative research, there is seldom an objective point of reference for which to evaluate the meaning of differences or similarities observed. Often the best we can expect is to draw conclusions based on the relative presence of a phenomenon in one group as compared to another. Correspondingly, similar findings from two regions may be caused by completely different phenomena and different observations may be caused by related phenomena (Artiles and Larsen 1998). One proposal for dealing with these issues is to use several sources of information to better understand the meaning of observations within the contexts from which they were drawn. A limitation of this study is that we did not examine contexts (e.g., local school policies) more closely and collect information from a wider range of sources.
Responses from school district administrators represent only one indicator of the perspectives and practices of their respective educational systems. We have attempted to extrapolate the information they provided and apply it to these systems and, indeed, the countries as a whole. It is far from certain that other groups (e.g., teachers, parents) would hold the same attitudes or understandings of the policies and practices examined. Therefore, we suggest that future comparative research along these lines include input from diverse stakeholders, closer consideration of local and national policy and, where possible, evidence derived directly from practice.
Another limitation concerns the type of information collected from participants. The instrument we used is based on a theoretical model (Persson 2003) in which individual items are intended to solicit information about larger constructs as well as provide straight-forward answers to basic questions about the situations in each country (e.g., prevalence of different solutions). Due to the explorative nature of the study, we chose not to combine multiple items to measure or compare theoretical constructs for fear of losing important information in this process. Rather, we have chosen to present these data descriptively giving as much detailed information as possible. Future research might examine our application of theory in relation to administrators' perspectives in greater detail. Such an approach could yield insights into the relationships between perceptions as well as assess the validity of this approach and our interpretations.

Conclusion
Findings from this study suggest that (a) Norway and Sweden share many approaches to supporting children with special needs, and (b) inclusive organizational solutions are largely preferred by school administrators in both countries. However, participants also appear to value solutions that are typical of traditional segregated services, leaning towards the 'categorical' side of our theoretical model. In addition, whereas Swedish participants placed more emphasis on 'relational' explanations for students' difficulties (e.g., 'schools are poorly adapted'), the child's 'individual deficit' was the most commonly cited reason for providing special education in Norway. We speculate that many of the differences we observed reflect the composition and size of the different populations of students receiving special support in each country. For example, the emphasis on 'individual deficits' in the Norwegian data may be due to the much smaller percentage of children who receive special education in Norway compared to Sweden. This conclusion underlines the importance of viewing contextual factors and examining alternative explanations in comparative research. That there is substantial ambiguity and variability in the way that countries define and provide special education is wellestablished (Meijer 2003;Pijl, Frostad, and Flem 2008). This study represents a small step towards clarifying that picture.