Adaptive situations and social marginalization in early adult life: students with special educational needs

In this article is developed a typology of social adaptation that captures differences in special needs students’ coping in early adult life. The typology reflects differences in framing opportunities and restrictions on social network relations in the spare time arena. The article seeks to validate the typology theoretically and empirically and discusses what are the more important factors behind being socially marginalized in spare time in early adult life up to the age of 22. The analysis is based on theory of social network and life course transitions. Empirical analysis of the relationship between four adaptive situations and different types of spare time networks suggests that broad mixed networks have the best potential for social inclusion of these former students. The article discusses how this could be studied further by analyzing collected life course data of the special needs student’s transitions.


Introduction
A main objective of the relationship between school and community in Norway is an inclusive school in an inclusive community based on the principles of equality and equity for all pupils. Inclusion incorporates the quality of the transition between school and community. About 10% of each age cohort, or 6000 of the pupils entering upper secondary school each year, are students with special educational needs. The majority of these students attend vocational classes, and their dropout rate is considerable.
Social network relations are developed during transition from one grade level to another, the internal transitions of the schooling phase of life. Young people spend a considerable part of their lives in school. Hence, social competence is affected and developed by a pupil's social experiences of teaching and learning in school. However, practice, conditions and processes are to a large degree characterized by a segregationist and exclusionary way of thinking that can be traced back to the former special schools of Norwegian tradition (Haug 1998). Overviews of special pedagogical research reveal a lack of condition-and process-oriented studies (Tøssebro and Ytterhus 2006;Solli 2005;Grøgaard, Hatlevik, and Markussen 2004;Haug, Tøssebro, and Dalen 1999). Kvalsund (2004) documents that the traditional organizational approach of upper secondary education has a disabling and excluding function for many pupils with special educational needs, a general feature also reported in other countries (Slee 1996;Skrtic 1991Skrtic , 1995. This knowledge underlines the importance of studying whether this disabling and excluding function of upper secondary school also has consequences for pupils after their transition to adult life. However, longitudinal research on children and adolescents involved in specially adapted teaching programmes is rare in Norway (Myklebust 2004, 118ff;Holmberg 2001;Ogden 1999;Kvalsund and Myklebust 1996;Befring and Soebø 1993).
Upper secondary education has two main aims: to provide academic or vocational qualifications and to encourage social learning and self-understanding. The types of differentiation that upper secondary school pupils experience in special classes compared to ordinary classes have a considerable influence on achieving qualifications and, in turn, on finding a job that provides a living wage (Myklebust 2004, 97Á126, 151Á2, Myklebust 2007. Another issue is how the experiences involved in this qualification process affect social learning. For many pupils, having friends at the beginning of adult life is probably more important than academic learning. However, there is a general tendency in the research literature to describe disabled people's social networks as small and restricted, whereas those of 'normal' young people are described as variable and diverse (see, e.g., Bø and Schiefloe 2007, 116, 138).

Purpose of the study
Analyses of inclusion and disability from a life course perspective most often focus on work, employment and parenting as the important aspects of the transition to adulthood. However, the transition between school and the leisure time arena is hardly discussed (see, e.g., Priestley 2003;Befring and Tangen 2001;Szymanski 1994), and studies of how disabled young people's leisure time networks are affected by the transition between school and early adult life hardly exist. Reviews of research on social networks (Pescosolido 2001;Degenne and Forsè 1999;Walker, Wasserman, and Wellman 1994) show that cross-sectional survey studies dominate the field.
The present study builds on analyses of quantitative and qualitative data on internal transitions between courses and course levels during upper secondary education reported by Kvalsund (2004). In later analyses this will be the source of identifying independent variables of relevance for understanding and explaining social adaptation in early adult life. A main aim of this article is to develop a typology of social adaptation that captures differences in the Special Educational Needs (SEN) students' social coping in early adult life. The categories of the typology are conceived as possible dependent variables. The typology is based on the dimensions of network density and network size, and the associated differences in framing opportunities and restrictions on social network relations in the spare time arena.
The relational qualities of the leisure time network affect the expectations and norms of the adaptive situations. The present study seeks to validate the typology theoretically and empirically by analyzing data on social adaptation in early adult life and discusses what the more important factors behind being socially included or socially marginalized in spare time are in this period of the life course. The purpose of the article is to discusses how this could be further studied by analyzing collected life course data of the SEN student's transitions from school to adult life.

Considerations of design, data and methodology
This study is analyzing the SEN student's transition from upper secondary school to adult life in a life course perspective in the light of theory of social networks and social capital. When we speak of 'student with special educational needs' (SEN) in our data material, this is the term used by the specialists in identifying and defining the individuals' characteristics and associated needs of special educational measures. The question of whether people have 'disabilities' has to be balanced against the question of whether social institutions have disabling effects (see, e.g., Barton 2003, 13ff;Williams 2001;Barnes, Mercer, and Shakespeare 1999, 20ff;Clark, Dyson, and Milward 1998, 156ff). However, an overemphasis on the social definition of disability might result in a structuralism just as unfruitful as the traditional medical overemphasis on individual characteristics. When analyzing disabling processes important frame factors and processes of a situation are formed primarily by the stable relational qualities of a person's social network rather than by the characteristics of the person (Wasserman and Faust 1994;Wadel and Wadel 2007). It is also necessary to look beyond the single situations and understand them as part of the persons' life course.
The life course can be regarded as the sum of a person's trajectories, 'composed' of past, present and future events in education, work, family and leisure life (see, e.g., Elder, Johnson, and Crosnoe 2004), manifesting itself as framing opportunities and/ or restrictions on action and interaction, both in the short and long terms. In this study, the short-term perspective on transitions is predominant. To understand the present effects of education and schooling, the researcher has to reconstruct central frame factors and the processes through which they have operated in order to understand the results expressed through the dependent variable. We can only understand properly what happens later if we know what has happened earlier. This analysis uses the concept of transition to capture the short-term part of the life course. Unlike events, transitions can be identified by a marked and permanent change in social norms and expectations (Elder, Johnson, and Crosnoe 2004;George 1993), such as the transition from course level to course level in school or from school to adult life for SEN students. For the reader to evaluate the longitudinal empirical basis for this study as well as the potential of further analysis we present an overview of the data collection process and data sets, see Figure 1.
The studied sample of 760 young people was selected from a population of 2025 students identified by experts as having special educational needs (covering a broad specter of categories) and therefore specially adapted teaching. The sample covered the intake cohorts that entered upper secondary school in autumn 1994 or autumn 1995 in six counties, including Oslo. Prospective data were gathered over the 1996Á 2002 period through questionnaires about the students and their situation that were filled in twice during each school year by the people who were closest to the adolescents in school Á i.e., their form teachers or counselors in cooperation with experts. The parents and youth were also interviewed on the telephone. In addition, audio taped interviews with students, teachers, counsellors, administrators and head teachers at 20 schools were conducted. 1 Data were collected for all 760 students in the sample provided they were full-time students: the 1994 cohort until spring 1998 and the 1995 cohort until spring 1999. The project titled 'Reform 94: Students With Special Needs' began collecting data on SEN students in the spring 1996; The effects of special classes within the 'time-table culture' of the school are documented in Kvalsund (2004) Á a study which contributes as a source of possible independent variables to the empirically grounded conceptual and theoretical validation of the explanatory model of the present article. The data collection effort of the first project was continued in the project 'Adult Life on Special Terms?' with the final data collected from the same youth in 2001 and 2002 (see Figure 1). Cross-sectional data on dominating network types and social adaptation in early adult life is used in the empirically validation of the model and the typology of possible dependent variables. In order to study what happened to the pupils during their internal transitions between course levels in upper secondary school, quantitative analyses of survey data were combined with qualitative analyses of interview data (see Project I in Figure 1) (Kvalsund 2004  During autumn 2001 and winter 2002, interview data Á for this study on leisure time relations and networks Á were received from 494 (65%) of the former SEN pupils. There are only negligible effects of the reduction in number of informants on main variables (such as gender, type of class, functional status, individual adaptation etc.) when the 2002 sample is compared with the original sample (for a more detailed discussion of this, see Båtevik and Myklebust 2007, 9Á13). The reduction from 760 to 494 students can probably be explained by the fact that some of the schools 'misplaced' the lists of pupils' names after they had submitted data for the last time in spring 1999. Even so, a response rate of 65% in longitudinal studies over such a long time period is rare internationally (Myklebust 2004, 31).
The analyses of one of the main sources of independent variables Á the school context Á and the development of the dependent variable of an explanatory model in this article, are based on robust longitudinal data from two research projects (Project I and II in Figure 1). The project titled 'Reform 94: Students With Special Needs' began collecting data on SEN students in the spring 1996; The data collection of the first project was continued in the project 'Adult Life on Special Terms?' with the final data collected from the same youth in 2001 and 2002 (see Figure 1) Cross-sectional data on dominating network types and social adaptation in early adult life is used in the further empirical validation of the model and the typology of possible dependent variables. Before the empirical validation, this article presents and discusses the theoretical perspectives and explanatory model of the constructed dependent variable. The construction of a model of the dependent variable addresses the question of construct validity. 'Adaptive situations and social relations during spare time in adult life' is developed as the dependent variable, which will be discussed along with the research results and theories on networks and social capital related to individual as well as contextual factors.
Theoretical and conceptual validation. Schooling and social marginalization Marginalization is an intermediate position between inclusion and exclusion (Halvorsen 2000). Being on the margins there is a risk of being excluded into some minority category. Social marginalization may be studied along two dimensions: a substance dimension (e.g. work, education, family) and a process dimension. This article focuses on the process dimension of basic social marginalization Á either having a network of friends or not Á and its subcategories, such as isolation and integration. These are the socially important and existential dimensions of social networks. Below we clarify one of the substantial dimensions of marginalization Á namely schooling and education to which the process dimension is related. Hockey and James (1993) posit a tripartite life course: one phase for education and qualification, one phase for work and adult life, and the last phase for retirement and old age. The three phases seem to represent an imbalance of importance and cultural meaning: by working and living an adult life, an individual has full status as a person and as a citizen of society, whereas children and youths primarily have instrumental value based on what they are going to be in the middle of their life course. Thus, the life course forms a context for how relationships between personal status, dependence and power are socially established. Categories such as child, pupil, student, adolescent, disabled, on welfare, adult or elderly are not just neutral descriptive terms; they also express political, economic and social dependence in the form of expectations and social norms. This indicates that the school phase of the life course is socially constructed (Hockey and James 1993) with marginalizing and possibly disabling features. If children and adolescents do not complete their education at various levels (e.g. by obtaining a craft certificate or matriculation qualifications from upper secondary school), this can be seen as a continuation and intensification of their marginalization rather than as a new situation.
In the first and third phases of the life course, citizens (children and persons of old age) are limited in their social interaction with adults, in the name of effective learning, by an expanding professional system of institutions (see, e.g., Telhaug and Mediås 2003;Edvardsen 1984;Slagstad 1998;Telhaug 1997, 262). Due to such differentiation and disconnection, school becomes an age-divided arena aimed at children both at primary and secondary level. The process of peer socialization may be seen as a special cultural construction, which Á combined with the emphasis on de-contextualized, universal and theoretical subject knowledge Á paves the way for an abstract instrumentalism in relation to adult life, work and mobility. The cultural meaning of schooling and the importance of building pupil identity collectively grounded in community and region have less importance (Kvalsund 2004;Telhaug and Mediås 2003). Furthermore, when peers dominate their social relationships, this leads to uncertainty and instability in the transitions that occur during the adolescent phase of the life course (Hockey and James 1993, 133). These characteristics are institutionally grounded, a quality that makes them far more important in the young person's life than they would be otherwise.
School experiences can have serious consequences for social adaptation in early adult life. Thinking of relevant independent variables, the situation at a given point of time is heavily influenced by what has happened before. The importance of leisure time social networks in early adult life also increases when SEN students face the postponed challenge of obtaining jobs and having families of their own.
Theoretical and conceptual validation. Social adaptation Á networks and social capital: strategic and altruistic perspectives Social networks are forums in which various interactions occur throughout the entire life course, whether they concern friends at school or in one's spare time, assistance in finding a job, a helping hand, company from day to day or someone to comfort. Social capital is a person's network as well as the opportunities that network provides to access whatever is in circulation in the network Á not least, relationally conditioned information, social and emotional support and social company (Wasserman and Faust 1994). Hence, social capital has both an individual and a collective aspect. These concepts help us be concrete in analyzing social adaptation after the transition from school to adult life.
Social capital can be regarded from two points of view, which give the relationships in the network completely different content and direction. First, administering social capital involves strategic actions that follow rational aims or planned actions that calculate the individual social return (see, e.g., Burt 1992;Giddens 1991;Lin 2001;Putnam 2000). According to its instrumental quality, social capital contains information Á e.g. about potential partners or friends, job prospects, exciting further education, special interests and hobbies. The prerequisite for this way of understanding social capital is that everyone thinks in such a strategic way in social contexts, which is hardly the case. Disabled young people are described more often than others as being in social debt, not able to return favours in social exchange relationships. Disabled people are said to benefit more than they contribute, with the result that network members burn themselves out (Bø and Schiefloe 2007, 116). In other cases, pupils actively try to establish new relations in an explorative way without knowing if they are registered and answered by the others. Many disabled adolescents will not be able to keep track of and strategically calculate their future social returns, restricting their abilities in establishing social relations with strategic values.
Second, according to the alternative position, socially altruistic actions and the relationships that arise from them maintain collective rules and qualities of behaviour despite differences between actors Á i.e. we are mutually obligated to each other in a community. Nafstad (2004) speaks of the caring man as a psychological alternative to the strategic, competitive man. Social interaction is communicative and creative, in the sense of not requiring any further external justification. To conceive social interaction or practical communicative action as technical, goal-related rational action, would be a serious instrumental mistake (Skjervheim 1976(Skjervheim , 1992. In a qualitative study, Grue (2001) discusses the leisure time arena of physically disabled people on the basis of interview data and life histories. The size or volume of friendship relations depends on what understanding and empathy disabled people experience from their friends. This is why their friendship networks are small: Friends are selected on the basis of their altruistic qualities. The family as a source of close relations is in most cases based on an altruistic perspective. Thus the independent variables in an explanatory model have to reflect the strategic as well as the altruistic logic. The adaptive situations as dependent variable most likely will have strategic as well as altruistic qualities.
Operationalizing adaptive situations and social marginalization Á network size and network density Without a network, a person appears to others as socially isolated, with few possibilities or little wish for socially valued contact, distinguishing himself as being different and abnormal. Hence, the existential content of social networks refers to two aspects: (1) having a network (or not); and (2) the structural quality of the network (e.g. its size and density). Strategic or altruistic actions are probably subordinate to the existential desire to have a network of friends in one's spare time. The existential aspect and quality of social networks becomes apparent through the use of the concepts of network size and density Á variables that can be operationalized and measured to show nuances in former SEN pupils' active adaptations to adult life. In their research overview, Walker, Wasserman and Wellman (1994, 67) refer to Marsden (1990), stating that data on usual network size and network frequency tend to be highly reliable. Relationships between friends are usually informally egalitarian and mutual, not asymmetrical (Degenne and Forsè 1999, 30). The networks and norms (such as trust, mutual support, solidarity and cooperation) that tie people together are considered benefits for the members of the network. We construct and qualify the validity of the dichotomies in the typology as far as possible on research results and explicit reasons rather than subjective choices.

Network size
The first dimension used in this analysis is network size Á i.e. the number of members of an individual's social network, apart from the person himself or herself. In studying people's social support, Wasserman and Faust (1994, 41 and onwards) referred to these networks as ego networks. Research indicates that networks consisting of many people have a positive influence on prospects for social integration and are also more robust and reliable. The networks can be larger if the youths are in a stable situation Á e.g. in stable school catchment areas rather than areas with a lot of in-or out-migration (Bø and Schiefloe 2007, 136). The transition to adult life represents reduced stability because individuals may have to leave networks (e.g. because of studying, working, establishing a family and possibly moving home, or other changes in their life situation). In a spare time network above a certain size, there are still people left in the network with whom to interact, confirm and develop one's self-image and seek or experience support. Bø (1993, 258Á9) points to a high correlation between success at school and the size of one's network. Hence, it is reasonable to expect that smaller networks will be a regular characteristic in our data on former SEN pupils in upper secondary education because almost 50% of them have dropped out of school (Kvalsund 2004). The networks in this project, however, were identified in answer to the question of with whom those former students usually spend their spare time (Marsden 1990). Asking young people formerly identified as SEN students to identify important persons would lead their thinking toward family members, according to established tradition. Walker, Wasserman and Wellman (1994, 62) refer to research by Wellman and Gulia (1993), who report that larger networks have a greater potential number of people who can provide social support and that in such networks each individual tends to be more supportive. Young people generally have rather large networks Á between 11 and 15 important people Á and studies show that the young have daily contact with half of their important network members (see Bø and Schiefloe 2007, 137Á8;Walker, Wasserman, and Wellman 1994, 55). A Norwegian study reports that 55% of girls and 45% of boys in lower secondary school usually are members of a group of adolescents who spend their time together (Øia and Bakken 2002, 76, 79). In fewer than one in five groups, two or three of the pupils are closer friends within the group, and about one in three adolescents (boys as well as girls, not disabled pupils) is usually together with one or two closer friends. Disabled people are generally described as having problems creating and maintaining their social networks, resulting in small networks restricted to family members and service providers within the state welfare system (Bø and Schiefloe 2007, 116). Based on the preceding discussion and literature on the subject (see, e.g., Bø and Schiefloe 2007, 116, 138;Walker, Wasserman, and Wellman 1994, 55), we categorize informal support networks as large if the more frequent contacts consist of at least three members apart from the adolescent herself. The network members can be friends, schoolmates, neighbours, workmates, family members of different age and gender (see Figure 2). In this way, we can describe and analyze the meaning of size variations in disabled people's networks. Studying these structural qualities is important because small and large network sizes seem to communicate very different norms and values.

Network density
Given a certain number of individuals in an ego network (g), the density of the network (d) is defined as the ratio of the number of actual relationships (L) to the maximum number of relationships that are possible, or [g x (g Á 1)]/2. The network density may vary from 0 (no relationships) to 1 (all possible relationships have been realized). With a value of 0.5, half of the possible relationships are active (see, e.g., Wasserman and Faust 1994, 101ff). The information in this analysis reflects the answers to questions about with whom the former students are usually together in their spare time and who of these persons usually interact. Data on usual interaction is found to be clearly more reliable than comparable data from questions on specific persons and specific points of time (Marsden 1990). Network size and network density are calculated in SPSS (Norusis 1999).
The contact between the people in the network can be more or less interwoven or integrated. Wasserman and Faust (1994, 62Á3) report that with all relations counted, except the one between the person (ego) and the others (alter), active networks have a density of between 0.3 and 0.5. In other words, there are usually relationships between one third and one half of the members in an ordinary network. Hence, there is a basis for talking about circles within networks, with the potential to form interpersonal relationships. When the network covers several contexts, however, there seems to be a significant decrease in social density (Degenne and Forsè 1999, 51, 55). This article focuses on the spare time context. Because network density in an ego network is an expression of interaction between the other supportive persons, it is reasonable to consider a normal network density at the same level as for ordinary networks. Taking account of this and the research based statement by Bø and  Schieflo (2007) that disabled people generally are described as having problems creating and maintaining their social network, we therefore choose network densities above 0.4 or more precisely Á at least 0.4 Á to be high. That level was chosen as a reference point to distinguish between high and low social densities in our analysis. However, network densities as low as 0.2 or less has been found in the data. Integrated networks have a relatively high number of members, and consequently, they invite communication and can challenge the need for coordination. In this way, the quality of the network increases the access to support and may compensate for threats to and changes in the life situation of the person concerned. With just a few people interwoven into the network, low network density may represent a process of looking inward, screening and bringing or accepting intimacy and isolation to one's life situation with close relationships.
Bonding and bridging networks Putnam (2000, 22Á3) distinguishes between bonding and bridging networks. Bonding networks (in our study defined by low network size, high network density) are characterized as inward looking, with supportive care, security, closeness, solidarity and homogeneity; they refer to a special variety of social capital that may be found in a family with responsibility for disabled children. This represents the inner area of the spare time arena, where people are on home territory or perhaps with a group of very close friends. Still, the same factors that bind people together inside a network may have negative effects for those excluded from the network. Internal loyalty may become too strong and experienced as isolation, sectarianism and especially egoistic actions. Such problems may be found, for example, within families and ethnic groups and can have a disabling effect (see, e.g., Bø 1993, 145;Bø and Schiefloe 2007), referred to by Woolcock (1998, 171 and onwards) as 'amoral familism'. Hence, more social capital is not necessarily for the best as the small, close and bonding family network in some cases may function as a sheltered environment for violence against the interests of the disabled.
In bridging networks (in our study defined by high size and low density), however, a restricted but rather large number of parallel relationships may occur between different people one at a time, establishing social bridges to other people. Bridging relationships are directed outwards. They are uncertain bonds as to whether they are answered and therefore can carry the substance of the transaction (e.g. friendship), but they might succeed. This is of the same kind as Granovetter (1973) stated Á the strength of weak ties. The bridging dimension is related to network size and may change over time from bridging to bonding qualities or vice versa, depending of changes in the social situation.

Results and discussion
Social isolation is a marked characteristic of the SEN pupils' social adaptation in early adult life. The main aim of this article was to develop a typology of social adaptation in early adult life and discuss nuances of the concept of social marginalization. What is the pattern of social adaptation when we combine data on network size and network density and analyze the resulting adaptive situations by the kind of network that is the more dominant? The situational categories of the typology represent an invitation to conceive marginalization as a process that can change content Á from isolation to bridging and bonding action and vice versa. For example, to stay in a bonding situation can be necessary for some time to establish the security in order to dare the more self confident bonding actions later in the process. The marginal situations can be relative stable as well. This dynamic quality of marginalization and the aim of validation of the typology as well as using it in empirical analysis, require the present sections of the article, results and discussion, to be integrated.
The network data collected for this analysis offer the possibility for reflecting on relational qualities, which is regarded as a more fruitful and advantageous approach than a more narrow explanation model on individual characteristics (see, e.g., Kvalsund 1994, 89Á108;Wadel and Wadel 2007). By combining the results of the two dimensions, network size (high and low) and network density (high and low), a fourpart typology of adaptive situations is presented, representing varying content (amount and types) of social capital and clear differences in expectations for the individual adolescent (see Figure 2).
Among the adolescents with experience as SEN pupils in upper secondary school in our survey, 3% to 4% (17) are completely alone or experience isolation and loneliness in their spare time, and 14% usually interact with just one other person. Thus, almost one fifth are alone or have just one person with whom to interact in their spare time. Another 23% of the former SEN pupils have regular contact with just two people. In other words, a total of 40% of the adolescents in the survey usually have contact with two or fewer people in their spare time.

Adaptive situations Á the network categories of the dependent variable
The people with whom the adolescents have relationships and how those patterns of relationships develop may differ. The social networks can thus have different compositions. We have data from the SEN pupil her/himself Á or from parents/ guardians in cases where the pupil cannot answer Á on the relationship between the person (ego) and the others with which he or she usually interact in spare time. A network consisting of people from the family in which the adolescent grew up (parents or siblings) or from the person's own established family (partner, spouse or children) is referred to as a family network. If it consists only of close friends, it is called a network of friends. However, a network may also comprise other adolescents from school or work, social workers, friends or family and relatives. If members from at least two of those categories are included in a network, it is referred to as a mixed network. (Pure school networks, pure job networks or networks containing only social workers were so few in number that they were not included in this analysis). Those categories allow us to study the social adaptation of the disabled in more than one arena, thus giving the possibility of judging social inclusion in a broader context than the group of young people of the same age. Thus, the typology of adaptive situations addresses the validity of social marginalization or inclusion.
By combining the results of the two variables, network density and network size, we can analyze differences in the distribution of those three types of networks among the four adaptive situations shown in Figure 2: socially isolating situation, socially exploring and bridging situation, socially intimate and bonding situation, and socially inclusive situation. The network of friends (in general, close friends) is the dominant network category, making up more than two thirds of the identified networks. The pure family network is the smallest category, only 7% of the networks, whereas mixed networks constitute 24% of the reported networks (see Figure 3). Schools are documented to be basic arenas of peer socialization and the (existential quality of) social networks, which points to the importance for adolescents of interacting with friends in their spare time. In particular, for many former SEN pupils, that importance yields a transition into further, or possibly reinforced, marginalization from upper secondary school to adult life. It is worth noting that all of the types of networks cover all four adaptive situations.
What embedded patterns of expectations and norms can be identified by analyzing the adaptive situation of leisure time? Three of the adaptive situations in which former SEN students find themselves after the transition from upper secondary school (situations I, II and III in Figure 3) represent a reinforcing contrast to the aim of social inclusion and normality in adult life.

The socially isolating situation (I)
The first type of adaptive situation shown in Figure 2 refers to relationships in which adolescents have very few individuals in their ego network, encompassing social isolation. The adolescents usually have no one to be with in their spare time, or only one or two people who may be the one person they can turn to Á the last resort. Social isolation frames and restricts action and interaction and makes adolescents vulnerable. There is little social capital in this adaptive situation on which to rely. With no relationships between the people in the network, the interweaving of the network is weak; i.e. the network density is low. The adolescents are on the borderline of social isolation and are socially marginalized. This combination of relational qualities is referred to as the socially isolating adaptive situation, and applies to 29% of the adolescents in our data. The family network is the one most clearly associated with the tendency to keep the adolescent in a socially isolating situation, shut up in a very small and cramped social room, with little social variation, asymmetrical social relations because of dependence and restricted flow of information. In almost half (49%) of the family networks, there are relationships between two or fewer people that do not involve the adolescent (see Figure 3). This may not be because the members of the network want it that way, but rather because the situation in forces them to. Some of these families may be under pressure and exhausted after many years of uphill struggle in the social welfare sector. The adolescent might end up in isolated helplessness. This situation could turn out to be a cover for violence and misuse of the disabled. The risk of social isolation clearly decreases with other types of networks: to 28% for networks of friends (with two friends at most) and to 18% for mixed networks (e.g. one person from the family and one social worker, or a workmate and one family member, either a parent or a sibling), see Figure 3. So, larger social networks probably will have preventive results.

The socially exploring and bridging situation (II)
The second type of adaptive situation reflects a rather low social density, or interweaving in the network (less than one third of the members of the network have relationships with each other). At the same time the network is rather large (three or more people in addition to the adolescent). This means the adolescent has parallel individual relationships with several people who otherwise have little or no contact. In our data, people in the networks have contacts in some cases, but the combination of low density and high size make parallel relations predominant. These ego networks may also have a structure of great value to the egos insofar as they can control the information in the network. These frame factors give opportunities for independent social exploration, which means expectations of one-way attempts to establish social relations that lead out into society (e.g. away from the family, out of a cohabiting partnership or perhaps out of a situation as the sole breadwinner) towards new friends. These are the expectations of social bridging, and 16% of the adolescents in our survey are in socially exploring adaptive situations (see Figure 2).
The strength of this type of network structure is that it has the potential for outward-looking interaction, with possibilities for increasing social capital in the direction of integration and inclusion. The situation can change in character and develop to become socially integrating and inclusive if lasting relationships are established between the other members of the network. Conversely, it can also lead back to a socially isolating position if people leave the network without being replaced or keep the person in the socially isolated situation if relational action stays unanswered. The percentage of this type of adaptive situation rises with increasing outward-looking orientation in the network: the family network (9%), the network of friends (16%) and the mixed network (26%), as shown in Figure 3. The increasing percentage by increasingly open network types confirms the outward-looking norms and qualities Á i.e. the social bridging and represents a further validation of the typology. It gives data on an important meaning category Á outward-directed independence Á of social adaptation for the formerly SEN pupils, now adolescents in early adult life.

The socially intimate and bonding situation (III)
The third type of situation arises when a network comprises few people but there are strong, close or intimate relations between them. This may be social relations to parents or selected relatives. There may also be an intimate friendship between siblings or close friends (e.g. two girls, two boys, or a boy and a girl). This situation is inward looking and restricting, and it protects against the world outside. This restricted framing of the situation develops the relationships between the few people through intimacy and social bonding, for better or for worse. For certain periods, the only need may be to create intimacy Á to have someone to trust, someone with whom to feel secure. The phenomenon is well known not only in the case of close friends and in families but also in mixed networks. One example can be observed in shared houses for intellectually disabled in a municipality. Scarce resources create situations where the ego actively tries to exclude the others from contact with a personal assistant. This exemplifies a socially intimate adaptive situation, with a closely interwoven core of very few individuals: two or three, including the adolescent. In this type of social interaction, the adolescent may become encapsulated. About 12% of the adolescents in our data fall into this type of adaptive situation (see Figure 2).
As might be expected, this situation is found most often in pure family networks (23%) because the family is the primary group for the adolescent (see Figure 3). Moreover, social intimacy is less common in networks of friends (13%), though it may exist as, for example, the 'bosom friends' variety of network. Intimate networks can also be found as mixed networks (6%), most commonly involving strong ties to a friend and a family member, but also personal assistants. The pattern of decreasing percentage by increasingly open network types represents at the same time a further validation of the typology. Empirically it documents that social adaptation have a very different content and quality compared with socially bridging situation. It also tells us that developing socially bridging action in many cases has the dissolving of a socially bonding situation as its prerequisite.

The socially inclusive situation (IV)
The fourth and final category of adaptive situation occurs when the network is large, and there are many relationships between its members (apart from the adolescent) Á i.e. there is high social density. This socially inclusive situation contains the most and best prospects for being socially included in a fellowship and represents the greatest social capital of the four adaptive situations. This situation, which applies to 43% of the adolescents in our survey (see Figure 2), includes combinations of several people in mutual relationships. Social capital is developing, however, and as such, there is far greater potential in this adaptive situation than in the other three. When several people cooperate, more ideas and initiatives arise, and the chances of still further contact and participation in youth culture and society increase.
Mixed networks (50%) are the most open and inclusive, followed by networks of friends (43%); family networks (20%) emerge as the least inclusive (see Figure 3). Hence, the differentiation into three different network types reflects a validation of the combination of the two variables, network density and network size, to form a network typology of social marginalization.

Social Marginalization
The socially isolating situation and the socially intimate and bonding situation have a more defensive, protective and excluding character Á both from an individual or a collective point of view Á than the other adaptive situations do. Hence, those two categories of adaptive situations may be referred to as providing processes of social security. The socially exploring situation and the socially inclusive or integrating situation are quite differently framed; the situations are offensive and open Á the bridging situation on the personal level and the integrating situation on both the individual and the collective level. People in those adaptive situations take greater social risks and are more explorative than people in a isolating or bonding situation. The analysis of the relationship between adaptive situations and network categories suggests that broad mixed networks have the best potential for social inclusion (see Figure 3). The relative strength of the different network types in each adaptive situation of the typology validates the content and process of the adaptive situation. At the same time, it is important to remember that as members of that sort of network group, adolescents may see their disabilities through the eyes of the other group members and so be reminded of being different and accepted. Despite this social inclusion is hardly a secure, stable state Á perhaps it is more like a risky process of social negotiations?
Regarded as a whole, the socially isolating, intimate/bonding and exploring/ bridging adaptive situations can be seen as meaning categories subordinated social marginalization. Those situations may also reflect different attempts to balance the search for security against the search for risk, enlightening the dynamics of the marginalization processes. Three of the categories of adaptive situations Á the socially isolating, socially exploring and bridging and socially intimate and bonding situations Á at the same time represent a subtle differentiation of the concept of social marginalization. As many as 57% of the adolescents who attended upper secondary school on special terms experience a transition into a clear socially marginalizing situation at the beginning of their adult lives, leading to adult life 'on special terms'.

Closing reflections: on further research
The different adaptive situations discussed here are not absolute or permanently established. They may change depending on the individual's interactions and contact with others, or on how they choose or are forced to act given the circumstances of the spare time arena. Establishing intimacy and creating security may be a phase before adolescents move on to social exploration, as a strategy of action that in turn leads to more socially inclusive adaptive situations. Further research on that issue is needed.
Many adolescents follow different roads through education and work. Social density presupposes group stability, which is contrary to the prevailing logic of a steady flow through upper secondary education. This informal part of young people's lives is most affected by transitions rather than by group stability. The adolescents are expected to find new friends, obtain a job, perhaps start a family, move to a new place and so on. Hence, social density in early adult life (as opposed to the network size) seems to be influenced, to a less degree, by the adolescent's social experiences in upper secondary education. Density within the ego-network, however, is probably more determined by the contemporary contextual situation and the people who are interacting, than by the ego-persons earlier experiences. This points to a conclusion that network size can be used as the main indicator of whether the experiences of former SEN pupils affect their spare time networks (see Kvalsund and Bele [in press]). Moreover, new data gathered through telephone interviews in spring 2007 could enlighten the qualities of adult life of the former SEN students, who are now 30 years old. Displacements and changes between the isolating, bridging, bonding and inclusive adaptive situations over time Á tracing social resilience Á would be an important topic of further research. This subject invites further analysis and follow-up longitudinal studies, qualitative as well as quantitative.
Note 1. Because the Norwegian Data Inspectorate did not give permission to identify and interview the sample of students from the intake cohorts, interviews were conducted with other students evaluated by experts as having special educational needs Á students from selected schools in selected counties of the total material. Those students came from schools, which gave no information on pupils' disabilities in advance. However, we have no indications that allow us to be persuaded that the situations in the schools from which we have qualitative as well as quantitative data are different from the schools where we have only quantitative data (see Båtevik 2002;Båtevik and Myklebust 2007;Kvalsund 2004;Kvalsund and Myklebust 1998).