Long-term dropout from school and work and mental health in young adults in Norway: A qualitative interview-based study

Abstract School dropout is related to difficult life trajectories in Western society. Developing effective preventive interventions is urgent. Nevertheless, few studies have interviewed unemployed young adults in the aftermath of school dropout to understand their experiences with influential factors. We interviewed seven former students two to five years after they had dropped out and seven same-aged students in their final year at college. The participants were given qualitative semi-structured interviews focusing on questions about what kept them on track and what pushed them off track when struggling to complete school. The participants were also clinically interviewed, drawing on the Mini International Neuropsychiatric Interview. The analysis revealed that the students who had dropped out described a larger number of mental health problems and problems of a more serious nature than the college students did. The participants who had dropped out also described less access to resources and social support. The clinical interviews supported the impression given in the qualitative interviews, that those who had dropped out were more burdened by mental disorders than the college students. The college students described comprehensive social support to play a major role in their coping with school and mental health problems. The former students who were unemployed and who had dropped out described internalizing mental health problems in combination with a lack of social support as important influences in their dropping out from school and employment, indicating the importance of further exploring the role of internalizing mental health problems in school dropout processes.


PUBLIC INTEREST STATEMENT
The importance of mental health and social support for school dropout We interviewed seven unemployed youths twofive years after dropping out from high school, and seven same-aged college students who had graduated high school. Those who had dropped out described a larger number and more serious mental health problems and less access to social support than the college students. The college students described that social support played a major role in their coping with school and mental health issues. The participants that had dropped out and were currently unemployed described internalizing mental health problems and lack of social support as important to their dropout processes. It might therefore be of interest to investigate further the relative occurrence of internalizing disorders in the group of dropouts who struggle the most to get back on track. Such knowledge could be essential in helping schools and mental health care services in preventing dropout.

Introduction
Education is a strong predictor of health (Freudenberg & Ruglis, 2007). Completing high school opens the gateway to higher education and thus to better-paid jobs. Higher education also contributes to better-informed health choices, and is associated with higher levels of social support (Freudenberg & Ruglis, 2007). Consequently, dropping out of school may reduce access to several health-promoting resources, and increase the risk of negative health behaviours, such as the use of tobacco, alcohol, illicit drugs, and attempted suicide (Bachman et al., 2008;Maynard, Sala-Wright, & Vaughn, 2015;Swain, Beauvais, Chavez, & Oetting, 1997;Townsend, Flisher, & King, 2007;Wichstrøm, 1998). In addition to the consequences for the individual, low educational attainment is also associated with substantial negative consequences for society. There is, for example, an estimated lifetime cost of about $240.000 for each student who drops out of high school (Levin & Belfield, 2007). These costs are due to subsequent lower tax contributions, higher crime prevalence and dependence on social welfare subsidies (Levin & Belfield, 2007;Rouse, 2007). Therefore, it is crucial to gain an understanding of the consequences of the dropout process on the transition to employment and adult life.
Surprisingly, dropout research has rarely focused on the transition period between school and employment (Maynard et al., 2015;Ramsdal, Bergvik, & Wynn, 2015;Ramsdal, Gjaerum, & Wynn, 2013). Most studies have sampled adults or younger adolescents. Thus, our knowledge is limited concerning what happens with respect to the school dropout processes in this transition period, which has been named "emerging adulthood" (Arnett, 2000). Moreover, the consequences of dropout depend on whether the dropout is temporary or permanent. Students who have dropped out temporarily seem to be more like students who have stayed in school in terms of demographics and school performance, and they seem to have a resiliency that enables them to bounce back (Dillon, Liem, & Gore, 2003;Entwisle, Alexander, & Olsen, 2004). Nevertheless, the amount of time spent out of school seems to be an important factor in explaining the chances of bouncing back Polidano, Tabasso, & Tseng, 2015;Ramsdal et al., 2013). Long periods out of school may decrease the chance of re-engagement, because these students also have a higher risk of anxiety and depression. Students with low educational attainment have a higher risk of these mental disorders (Bjelland et al., 2008;Chazelle et al., 2011;Wang, Schmitz, & Dewa, 2010). Moreover, the transition into adulthood seems to be a critical juncture in the course of psychopathology and mental health (Schulenberg, Sameroff, & Cicchetti, 2004). Roughly 50% of all lifetime mental disorders emerge by the mid-teens, and 75% by the mid-20s (Kessler et al., 2007). Consequently, a considerable number of adolescents and emerging adults develop a mental disorder during their school years and later during the transition into employment. Summing up present knowledge on mental health problems in young Norwegians, Sletten and Bakken (2016) found these problems to be increasing. In a study of students who had dropped out of high school, as many as 20% reported that their dropping out of school was due to mental health problems or psychosocial problems (Markussen & Seland, 2012). As a result, it is of crucial importance to know more about the mental health of those who have dropped out in emerging adulthood, particularly those that struggle the most to get back on track. For dropout prevention and re-engagement interventions to be effective, they should target this group of off-track dropouts in particular.
Effective interventions would require more knowledge of students' own dropout experiences, their experiences of how dropout processes may influence their transitions into employment, and the role of mental health in these processes. Therefore, we interviewed young adults facing challenges of getting back on track two-five years after dropout. We asked them: When you had difficulties staying in school, which factors contributed to push you out of school (temporarily or permanently), and which factors contributed to keep you in school? To get a better understanding of the characteristics of these dropout stories, and how they might differ from the stories of those who do not drop out, we also interviewed students who had moved on to college after having struggled to complete high school. In addition, at the end of the qualitative interview, we also asked directly about experiences with mental health problems. We wanted to know about their present mental health. Therefore, we also chose to interview all the participants clinically. In this way, we included two different sources of information on mental health, their subjective experiences through dropout or graduate processes and beyond, and a structured diagnostic assessment of their present mental disorders.

Method
This study combined qualitative interviews with clinical interviews. The purpose of this strategy was to obtain the informants' own narratives regarding their mental health and the influence of any mental problems on their school completion or dropout from school. In addition, we wished to diagnose any present mental disorders, drawing on clinical assessments made by a clinical psychologist supported by a structured clinical tool (the M.I.N.I.) in order to obtain valid psychiatric diagnoses. In this manner, we could compare the informants' own narratives to their psychiatric diagnoses and thereby get a more comprehensive picture of their mental health challenges and the implications these challenges have for their school performances. By including different sources of data (narratives and structured clinical interviews) on the topic of the informants' mental health, we achieved data triangulation; a method used to increase the validity of qualitative approaches (Hansen, 2006).

The recruitment of the participants
The Norwegian Labour and Welfare Administration (NLWA) has a central role in the Norwegian public welfare system, organized by local authorities and the central government. The NLWA provides services such as unemployment benefits, work assessment allowances, sickness benefits, and pensions. These services are in other countries typically covered by the social services, disability insurers and employment agencies. The NLWA informed potential participants about the study and asked for permission to pass on their phone number so that the researchers could contact them and ask for their consent to participate. Four men and three women were subsequently interviewed. College students in their third year of social work studies were recruited through a short introduction to the study in their classroom and were encouraged to contact the researchers by e-mail or phone afterwards. Three male college students and four female college students, aged 18-25, contacted us by e-mail and were subsequently interviewed. None of the students lived with their parents, and they were all financially self-dependent. Only one of those who had dropped out lived with a parent, but all of them were financially dependent on assistance from the NLWA. The recruitment and the interviews were completed in the period 2013-2015. The interviews were performed at the local university college, their length varied between one and a half and two hours, and they were audiotaped.
Recruiting long-term school dropouts was a complicated process. These young people had experienced years of social and academic failures at school, and revisiting these narratives was challenging for them. Thus, we were unable to include the participants as co-researchers. However, we did give advice about how to connect with free health care services, and offered to contact these services for them if the results should make such contact desirable. After the interviews, we also contacted some of the participants by phone to exchange information and secure a common understanding of the results.
The Regional Committee for Medical and Health Research Ethics (2013/101 REK Nord) approved the study. Participation was voluntary, and the participants all gave written informed consent.

Qualitative interviews
The qualitative interviews of the 14 participants in this study were conducted at the local University college. The interviews were semi-structured and started with questions on demographic information. The interview guide was organized according to the following themes: educational resilience, mental health, somatic health and well-being. Questions on resilience were most numerous and focused on subjects like, who or what had been most helpful in their struggle to stay in school? What was the main problems/obstacles when trying to cope with school activities? Where there any turning points, for better or worse, in their school engagement? What influenced their intention to stay or drop out of school? Who cared and got involved when they were at risk of dropping out? What helped them endure school and what made them want to give up and was any kind of help or intervention offered to help them stay in school? Thus, we presented the participants with several occasions for spontaneously sharing their experiences with mental health problems during school years. Finally, at the end of the interview, we asked in particular about somatic and mental health problems, and feelings of well-being. The same questions were given to all 14 participants. All interviews were performed by one of the authors, Gro H. Ramsdal, who is a trained clinical psychologist, and has specialized in adult psychology. The qualitative interviews lasted one and a half to two hours and were followed by the clinical interviews. The interview sessions was finalized by inviting the participants to give feedback on their experiences with the research situation.

Qualitative analysis
The interviews were fully transcribed and analysed. The analysing process was based on a qualitative methodology inspired by concepts from grounded theory (Gammon, Johannsen, Sørensen, Wynn, & Whitten, 2008;Glaser, 1992Glaser, , 2001Glaser & Strauss, 1967;Wynn, 2004;Wynn, Karlsen, Lorntzsen, Bjerke, & Bergvik, 2009). First, the interviews were transcribed word by word, read and reread-one at a time-thus giving room for open coding before the next interview was performed. By open coding, we mean staying close to the transcribed wording while describing, "what is going on" in our data (Glaser & Strauss, 1967). The interviews with the students were analysed first because they were recruited more quickly and interviewed first. Consequently, the open coding of these interviews formed a background for the analysis of those who had dropped out. Secondly, we examined the open codings looking for common features and differences in content to form higher level concepts describing important contributions, positive and negative, relating to their struggle to stay in school. Each transcript was colour-coded, enabling us to make constant comparisons (Glaser & Strauss, 1967) of related text samples within and between interviews. From the open codings, a set of higher order concepts emerged describing negative contributions to the students' struggle to stay in school: depressive moods, sleeping problems, being bullied, social anxiety, low self-esteem, problems making friends, eating problems, lack of motivation, attention problems and learning difficulties. All these concepts seemed to converge on a common category here called "mental disorders". The open codings describing positive contributions to the struggle to stay in school combined into concepts like parental ambitions and school engagement, students' school engagement, support from parents, friends, teachers and extended family. These concepts seemed to converge into a category here called "access to support and resources". The content of the interviews was constantly categorized and recategorized throughout the research process. Through the research process, memos were produced to analyse, compare and connect themes and categories. Thus, the categories were tested and refined until saturation occurred and new interviews no longer produced new information about what contributed to making them stay in school and what contributed to making them leave school prematurely.

Clinical interviews
The clinical interviews were administered immediately after completion of the qualitative interviews and were scored consecutively and plotted into a table (Table 1). We applied the Mini-International Neuropsychiatric Interview (M.I.N.I.). This is a short structured diagnostic interview developed for DSM-IV and ICD-10 mental disorders covering some of the most common axis-I mental disorders, including affective disorders (major depression, bipolar disorders), anxiety disorders (agoraphobia, panic disorder, social phobia, generalized anxiety disorder, post-traumatic stress disorder), obsessive-compulsive disorder, alcohol and substance-related disorders, psychotic disorders and eating disorders (anorexia, bulimia). One axis-II disorder, namely antisocial personality disorder is also briefly covered (Sheehan et al., 1998). An interview based on the M.I.N.I. will typically take 25-35 min to complete, and may aid the clinician in setting more valid diagnoses.

Results
The analysis resulted in two categories, "mental disorders" and "access to support and resources". The following presentation will show how mental disorders and access to support and resources were central to our findings. In the qualitative interviews, the participants were asked questions about what helped and what interfered with their coping at school. Two kinds of interference seemed to be of particular importance to coping in these dropout and graduation processes: mental health problems and learning difficulties. Mental health problems, in particular, were described in relation to the dropout processes as well as in relation to the graduation processes. Nevertheless, there seemed to be some essential differences between these two kinds of processes.
First, the students who had dropped out reported more diagnosed mental disorders during the qualitative interviews than the students (Table 1). With two exceptions, they had received a first psychiatric diagnosis already during their school years and only one of those who had dropped out had never been diagnosed with any mental disorder. All but one of those who had dropped out never finished their second year in high school, although some of them tried more than once. Even though some of the students who had completed high school described struggling with mental health problems during their school completion processes, only two of them were actually diagnosed with a mental disorder while still at school.
Second, there appeared to be differences between dropout and completion processes in the descriptions they gave during the qualitative interviews of the symptoms they had experienced during their school years. Those who had dropped out described a higher number of symptoms and more serious symptoms than those reported by the students. While the students talked about loneliness, stress and lack of motivation, those who had dropped out talked about phobic anxiety of being with peers, extreme underweight, self-harm and acting out.
Third, both groups completed a clinical interview (M.I.N.I.) enabling us to diagnose their present axis-I mental disorders. The structured clinical interviews showed the same pattern as the qualitative interviews, indicating that also presently those who had dropped out fulfilled the criteria of more mental disorders than the students did (Table 1). Four students with mental health problems described how these problems had played a significant part in their dropping out of school. When asked what influenced the decision to leave school prematurely, one of them replied: Well, really the same mental problems that I'm still struggling with now.
To explain what had happened, this former student went on describing four years as a victim of school bullying. In this process, a feeling of uneasiness developed when being around peers, particularly at school. The transition to high school presented the former student with the insurmountable challenge of unfamiliar peers in new environments, causing dropout within a month: Something about being at school really bugged me, I became extremely stressed, and I developed headaches and anxiety.
Two other former students were also struggling with anxiety in social situations, making their presence at school a painful experience. One had been troubled by bullying and the other by social exclusion. Nevertheless, they endured their hardships and they kept coming to class all through elementary school. For the one who had been bullied, the number of people present in class was the decisive factor. She was able to cope with the small number of pupils in her elementary school class, but more or less gave up on her high school class of 30 students. Reflecting on her failed attempts to graduate from high school and her problem with being present at school she observed: That's not my kind of setting, too many people for me to … Well, fact is, I really don't think I would be all right and do well in a class of more than 15.
While this young woman could achieve academically when she was able to get to school and stay in class, the participant who had felt socially excluded had another problem that influenced her experience of being in the classroom. She had problems with reading and developed a phobic reaction to all tasks that involved making a presentation in front of her class. Nevertheless, she coped with her reading problems until she started having tests and getting grades in middle school. This was the first time she realized how much she was lagging behind academically. From that point in time, she described school as an uphill struggle, draining her motivation to achieve and her joy of living. Even though her social life had improved dramatically in middle school and she was in with the popular kids in high school, this could not compensate for her academic problems. When asked how the learning difficulties affected her life she answered: I thought I would end up on the street or something like that, being such a failure … I was so fed up.
She described how her strong feelings of being "useless" and "worthless" peaked during her second year in high school and caused her to drop out. Thus, all three former students cited so far described how their problems with being present at school influenced their decision to drop out. Another young woman, however, focused on factors outside school when reflecting on her first dropout experience: The first year I had to quit school due to love. I was very unstable, up and down. /…/it was also due to my background and how I grew up, such things.
She talked about close relationships and how her family life had affected her dropout process. Her experiences had made her life chaotic and insecure, reducing her capacity to trust people. This lack of trust was exemplified in her descriptions of several heated conflicts in class with teachers and fellow students, where she reacted with stress, moodiness or intense anger, sometimes resulting in her shifting to another class or dropping out altogether. Describing her attitude toward other people in communication she revealed her sensitivity to criticism, she said: However, do not talk to me in a rude way because then I immediately become sort of against you. /…/ If I don't like you I just end it.
She described herself as hypersensitive to all kinds of injustice adding that she had had enough of people looking down their noses at her when she grew up. Thus, overtime, family difficulties came to influence her relationships at school in disruptive ways. Her lack of trust in other people's good intentions made it difficult to establish the sort of stable positive relationships with teachers and peers that she needed to cope with her academic challenges.
In the dropout group, two of the participants reported symptoms of mental health problems only after dropping out. Nevertheless, one of them described being bullied at school and struggling with learning difficulties. During high school, he was less monitored and quickly got massively involved in out of school social activities like partying and drinking. Before the end of the first year, his absenteeism forced him to quit. After dropping out, he had problems finding a job he really liked. In the same period, he developed anxiety attacks and was treated for depression. When asked if he had problems with anxiety during his school days, he replied: Yes that may well be, taking into consideration that I was bullied I think it started there.
Another young man also seemed to develop his mental health problems in the aftermath of dropout. He was a bright student, but teachers kept commenting that he did not fulfil his potential although he had quite decent grades all the way into high school. He moved several times during elementary school and ended up feeling socially marginalized. During the second year of high school, he had a conflict with one of his teachers. He described that this conflict caused him to lose his academic self-confidence, and consequently, he was not able to uphold his apprenticeship contract. Without documentation of formal competence, he was unable to fulfil his ambitions in working life and had problems keeping a job.
Although these two former students were not diagnosed with a mental disorder during their school years, they described some early problems of school adaption. Perhaps they had difficulties verbalizing these problems at the time, so the problems remained vague. However, several years later, they did put into words how the lack of formal competence made it difficult for them after school dropout and how this eventually resulted in depression, anxiety and dependence on social services. Finally, one of those who had dropped out had no history of symptoms of mental disorders, although he had some learning difficulties. He had also struggled to get a job and to stay in jobs.
In spite of the burden of mental health problems described in the qualitative interviews, only two of the students who had dropped out had been in contact with mental health services prior to the dropout event. One of them said that "my mother helped me get in touch with mental health services". Another described that the school nurse took contact because she had observed the student`s extreme weight loss and therefore helped this student to get in contact with a psychologist. The others who had dropped out and had mental health issues sought help only after the dropout event but before the interviews due to situations and events that they experienced as a life crisis. Three of the college students had been in contact with school mental health services during middle or high school. None of them were in any kind of treatment for mental health problems at the time of the interviews.
Among the students, however, only two had a history of being diagnosed with a mental disorder. However, several students had experienced some symptoms of mental health problems in middle and high school, causing their grades to drop and temporarily draining their motivation to graduate for a limited period. One young man, for example, was a bright introvert student struggling with loneliness. He had more success befriending his teachers than his classmates. Gradually through mid-adolescence, these problems made him feel increasingly sad, apathetic and exhausted. It was not until his first semester in high school, that his grades dropped dramatically. Talking about factors weakening his decision to graduate, he remarked: Well, there have been periods when I have been kind of depressed and stuff, by social things, with my schoolwork going down the drain.
However, he did get professional help and gradually he became socially included in his high school class. Within a year, his high school grades were back to normal. He remarked that during his school and college years, coping with schoolwork remained a source of joy and self-confidence.
One of the college students who had a history of a mental disorder prior to high school graduation also had difficulties with social inclusion. This participant was bullied both physically and psychologically during middle school, causing social withdrawal. Eventually, at the beginning of high school, child mental health services diagnosed the student with depression. The student described the social problems at school: I was afraid to walk past people my own age. Particularly young … oh god I really hate that! Nevertheless, the student enjoyed coping with schoolwork, loved art classes and had some support in therapy. After a year at home and receiving strong family support, the student joined a new class, developed a close friendship and eventually completed high school.
The three remaining female college students, on the other hand, had little overall problems with social acceptance. They felt socially included from elementary school and all through the middle years. They reported meeting and socializing with their friends as one of their top motivators to come to school. When starting high school, two of them experienced changes in their school environment that disturbed their prior feeling of social self-confidence. They both described a marked drop in their well-being due to these changes and reported a decreasing motivation to focus on their schoolwork followed by a marked drop in their grades. One of them said her demotivation and concentration problems had started when her parents divorced. Describing the going back and forth between her parents she said: Well, that was actually the real reason that she (her psychologist) advised me to leave home /…/ I had frequent stomachaches and was very … stressed and nervous all the time.
Although the parental divorce came at a challenging period in her life, she had by then established a stable sense of coping with several school subjects. This was a common experience among these three female college students. They described their first 7-10 years in school as learning processes characterized by positive coping experiences, skilled and supportive teachers and positive peer relationships. They also described highly internalized parental expectations concerning the completion of high school resulting in both motivation and demotivation. When they could not live up to these expectations, they sometimes gave up trying for a while. One of them said it like this: I felt that I hadn't managed to live up to what was expected of me … so I simply stopped trying altogether, just to spare myself the disappointment.
Nevertheless, their inner drive to succeed academically seemed to catch up with them and get them back on track. The female college students all described that when grades dropped and absenteeism increased in high school, they sought out the help and support they needed to overcome their problems and graduate.
The two remaining male college students differed from the female students in one important aspect. The men had learning difficulties. Both suffered through years of low grades and academic disappointment without reporting symptoms of depression, anxiety or externalization. The first reasons they gave for still being able to graduate coincided with that of the female college students, they all enjoyed school socially and thus were motivated to attend. The second reason these two men gave for completing school, was that they received help and support: one by his parents and the other by his mother and an academically successful friend in his high school class. This kept them academically motivated. One of them said: I am a bit proud of being where I am today, actually. I think it mattered somehow that they (parents) pushed me a lot to help me graduate.
Summing up descriptions from both groups, it seems that both in the graduate group and in the dropout group some participants described struggling with mental health problems. Nevertheless, the four students who had dropped out and had been diagnosed during their school years described more long-term problems. Their problems started in kindergarten or elementary school and intensified into a downward spiral resulting in comprehensive absenteeism and eventually high school dropout. Furthermore, their previous mental health problems still seemed to disturb their coping with present-day challenges in education and employment. Prior experiences of bullying in elementary school, for example, still strongly limited present day employment choices of one who had dropped out: I have problems with situations involving young people.
For two of the former students, the mental health problems described emerged in the aftermath of leaving school. Their mental health problems were described as resulting from the frustrations and challenges related to finding satisfying employment and becoming economically independent without any formal education.
The students, on the other hand, described their mental health problems more as temporary midadolescent challenges they had coped with and successfully overcome, at least for now. Even one of the graduates with a prior history of depression reported an overall sense of coping and satisfaction in her third year at college: I'd say life's never been better I feel really fine … like it at school, like it at home.
The differences in present mental health described by those who had dropped out and students during the qualitative interviews were in line with the differences found in the structured clinical interviews. The latter showed that those who had dropped out had more mental disorders than the students did two-five years after planned school completion (Table 1).
Surprisingly, the help and support triggered by the less debilitating and more short-termed problems of the students was comprehensive. There was a marked difference when we compared this support to the support triggered by the more debilitating long-term mental health problems of those who had dropped out. The participants also described how there had been far more resources available for the students who had struggled to cope with mental health problems and learning difficulties before gradation than for those who had actually dropped out of school. The students described help coming from parents, other relatives, friends, boyfriends, teachers, therapists, mental health nurses, student counsellors and other adults in their network. When counting all these examples of help described in the qualitative interviews, we found a total of 47 examples across all interviews. The participants who had dropped out of school provided only 13 of these examples. This means that the students who had dropped out provided 28% of the examples of help and the college students provided 72% of the examples of help described in the qualitative interviews (see Table 2).
When the students struggled, many people cared and got involved. When those who had dropped out struggled, on the other hand, there seemed to be less involvement. Only two of them had been in contact with mental health services during school years and experienced some support there. The others seemed not to have been aware that their mental health issues were signs of a mental disorder and thus did not seek help until their mental health problems caused some kind of life crisis in the aftermath of dropout. None of them mentioned financial problems as a reason for waiting that long to get help. This maybe because mental health services in Norway are available at a very low cost. When asked about who got involved and helped them with their problems in school, most who had dropped out answered: 'my mother'. Then, as we explored the potential involvement of others, several answered like this young man: No … no one that I can think of at least.
Having support from their mothers was of course emotionally important. Even so, their interviews showed how their problems with learning difficulties, eating disorders, bullying, and social phobia, were more than their mothers could handle on their own.

Discussion
In the qualitative interviews, all but one of the seven who had dropped out of high school described symptoms of mental disorders. Four of them reported long-term processes where symptoms had developed through their school years. They described intensifying symptoms during middle or high school. Finally, their symptoms manifested themselves in several diagnoses in the clinical interviews. Two other students in this group experienced symptom development over the two-five year period after dropout. They had both been able to hold odd jobs, but somehow they were not motivated to stay in these jobs, and they were presently unemployed. With one exception, all those who had dropped out had struggled for years to establish adult autonomy and financial independence without succeeding in the long run. The students, on the other hand, reported stress, lack of motivation, periods of family distress and other social challenges triggering moderate degrees of anxiety and depression. These processes had temporarily affected their motivation, their presence at school and their grades negatively. Nevertheless, they coped by accessing support from parents, teachers and health services, and by joining summer school and remotivating themselves, thus getting back on track and completing high school within a year or two. Several years later, in their final year of college, their previous school problems had manifested themselves in a quite limited number of diagnoses as assessed by the M.I.N.I. when compared to those who had dropped out. Thus, it was the intensifying symptoms of mental disorders accumulating over the school years or in the aftermath of dropout that were described as making adult autonomy and economic independence difficult to attain.
However, data from a qualitative study with a small sample of participants makes generalizing difficult (Hansen, 2006;Wynn, 1995). Such generalizations are made even more difficult by the fact that only a few studies have tried to estimate the proportion of high school dropout attributable to mental disorders (Breslau, Lane, Sampson, & Kessler, 2008;Kessler, Foster, Saunders, & Stang, 1995;Vander Stoep, Weiss, Kuo, Cheney, & Cohen, 2003). Although there is some agreement that dropout is one of the consequences of mental disorders, the estimates of the proportion of termination attributable to mental disorders vary between 10% and 46%. So far, the extent of the problem has not been established. This makes it more challenging to interpret the meaning of the present data as to the scope of the problem suggested.
Furthermore, it is important to notice that our sample of youths that had dropped out is not typical of the entire dropout population of Norway. Our participants represent young men and women who are not back on track or in stable regular jobs, and who are registered with social services several years after they dropped out of school. Our results may suggest an overrepresentation of mental disorders within this group, but statistical studies with representative samples are needed to address this question. These results may indicate that psychopathology is an important factor in the developmental trajectories of one group of youths that drop out of school, namely those struggling the most to get back on track after high school dropout.
Hence, our data seem to indicate that in this group in particular there may be quite a few young adults struggling with mental disorders. For the four participants in our sample, the development of mental disorders seemed to be associated with the dropout process itself. According to Breslau's literature review (2010), findings regarding health effects on dropout were mixed. However, he concluded that studies testing for such causal effects indicated that psychopathology was the health factor most likely to have an effect on dropout. Breslau (2010) differentiated the disorders related to dropout into four groups: internalizing, externalizing, attention deficit/hyperactivity and substance use disorders. Since these disorders may have distinct effects on educational processes, their tendency to co-occur becomes a challenge. Nevertheless, the suggestion that there is an impact of mental health on dropout is also supported by the findings of Vander Stoep and colleagues (2003). They reported that students with emotional/behavioural disability had a far lower proportion of school completers than other disability groups, i.e. mental retardation and multiple handicaps. Furthermore, Breslau and colleagues (2008) observed that the presence of a single disorder was not related to school dropout, whereas the presence of three or more disorders was so related. In line with this finding, four out of seven who had dropped out in our study had been diagnosed with three or more disorders, while none of the students had been diagnosed with more than one. It was nevertheless a bit puzzling that none of those who had dropped out had externalizing diagnoses or ADHD diagnoses during their school years. This was unexpected, since there is accumulating evidence suggesting that the relationship between mental disorders in general and high school dropout may be caused by the effect of a few specific disorders, namely externalizing disorders and ADHD (Breslau, 2010).
In the present study, those who had dropped out had been diagnosed with internalizing disorders such as anxiety, depression, and eating disorders. It must be noted that the clinical interviews (M.I.N.I.) did not examine the presence of conduct disorders or ADHD, but we could find little in their descriptions of dropout processes suggesting that these types of disorders were present. Although internalizing disorders seem to have a weaker effect on school dropout (Esch et al., 2014), a systematic review confirmed the role of internalizing problems as a risk factor in high school dropout (Melkevik, Nilsen, Evensen, Reneflot, & Mykletun, 2016). A recent meta-analysis of long-term studies found small but significant associations between internalizing disorders and subsequent school attainment, with more consistent associations for depression than for anxiety (Riglin, Petrides, Fredrickson, & Rice, 2014). Complex interactions emerged between types of internalizing disorders, school attainment and other factors. For example, lower school grades were associated with higher levels of depression. However, an association between lower school grades and anxiety was only found when including other factors such as age. Anxiety is more detrimental to school achievement in later adolescence than in early adolescence. This signifies the importance of investigating effects of depression and anxiety separately. Similar complexities surfaced in a nationally representative sample of American adults showing significant associations between school disengagement and two specific mood disorders (major depression, bipolar disorder) but only one anxiety disorder (specific phobia) (Vaughn et al., 2011). These suggestions seem to agree with the descriptions given in our study both by those who had dropped out and by the students, of the complex processes linking internalizing symptoms to school attainment. For example, the students described how their fear of failure inspired them to get back on track, while those who had dropped out described how their anxieties overwhelmed and demotivated them. Students leaving school prematurely due to anxiety also seem to have a higher risk of lifetime diagnoses of social phobia and a number of other diagnoses than those completing their education (Van Ameringen, Mancinci, & Farvolden, 2003). Thus, the combination of low education and internalizing disorders like social phobia might affect students' ability to get back on track.
One possible interpretation of our data may be that among those who had dropped out, and struggled for years to get back on track, internalizing symptoms might be more frequent than in the general dropout population. However, as this is a qualitative study, we can only hypothesise about this issue and it should be further addressed through statistical studies. Four of those who had dropped out in the present study described school trajectories characterized by vulnerable life situations. These vulnerabilities were described in the form of complex experiences of unstable childhood environments, loneliness, bullying, learning difficulties, lack of support from teachers and no one to talk to about their problems. They explained how these factors interacted, and described how these interactions contributed to a downward spiral of decreasing self-esteem, motivation, concentration and school engagement and increasing symptoms of internalizing disorders. These four participants described their processes as cumulative. Early reduction of self-esteem due to bullying was experienced as gradually developing into extreme shyness and social phobia, while overall feelings of being inadequate due to learning difficulties were experienced as developing into later depression. In line with these descriptions, an increase in internalizing symptom trajectories seems to be associated with a significant decrease in academic performance and school attainment (Patalay, Deighton, Fonagy, & Wolpert, 2015;Veldman et al., 2014). The research literature does not necessarily indicate that internalizing disorders cause people to drop out. What it does suggest is that when internalizing symptoms intensify overtime as reactions to difficult life situations, these disorders may become markers of long-term cumulative dropout processes. In agreement with these results, the participants in our study described how struggling with an increase in internalizing symptoms made their presence at school an unpleasant experience. Accordingly, those who had dropped out reported that, as problems piled up overtime, intensifying mental health problems contributed substantially to absenteeism, and was described as an obstacle to school completion. Consistent with this explanation, Suldo, Thalji, and Ferron (2011) found that internalizing problems predicted a decline in school attendance, thus supporting the notion that students with anxiety and depression tend to avoid being at school. Since absenteeism is strongly related to school dropout (Allensworth & Easton, 2007;Balfanz & Byrnes, 2013), any factor that substantially increases absenteeism may influence school attainment. Consequently, dropout and mental health issues may be associated due to their common association with a range of social and contextual factors potentially causing both of them (Fergusson, McLeod, & Horwood, 2015). Nevertheless, our data seem to suggest that internalizing symptoms work as a moderator, thus strengthening the link between childhood adversity and high school dropout.
Two students with symptoms of internalizing disorders became aware of their problems in the aftermath of their dropout processes. Their descriptions of the post dropout period were dominated by experiences of how the absence of formal education made coping with adult working life difficult. One of them could not access the jobs he preferred, while the other was repeatedly excluded from promotions and was vulnerable to downsizing and reorganizations. Both these young men described how internalizing symptoms developed gradually, as experiences of failure accumulated. Because they could not find stable, meaningful employment and become financially independent, they were unable to fulfil essential adult roles. In line with these descriptions, a Norwegian study based on a nationwide sample showed that sense of mastery surfaced as a strong mediator between low educational attainment and mental health problems (Dalgard, Mykletun, Rognerud, Johansen, & Zahl, 2007). The descriptions from our study suggested, however, that dropping out of high school in itself had contributed to the development of their mental health problems. These descriptions seem to be in accordance with studies showing that internalizing disorders may develop as a consequence of school dropout (Esch et al., 2014). Every additional year of delayed high school completion was found to be associated with increased levels of internalizing symptoms ). Thus, both the level of academic attainment and the time students take to graduate from high school seem to affect the level of internalizing symptoms in adulthood. This conclusion was supported by the fact that there were no differences in symptom levels between those with elementary school as their highest attainment and those with delayed high school completion . The study suggested that even temporary delays in school completion might have negative effects on future mental health. Consequently, whether the participants in our study have dropped out temporarily or permanently, we would expect them to have a higher level of internalizing symptomatology in adulthood than the participants who graduated within a normative timeframe. Data from the qualitative and the clinical interviews were in line with these expectations.
Summing up so far, our study suggests that the development of internalizing disorders is one marker of young people struggling the most to get back on track and remain on track in the aftermath of high school dropout. Whether the internalizing symptoms intensified during the dropout process or after the dropout incident, those who had dropped out described their symptoms as contributing to long-term problems with re-entering education or remaining employed. In addition, we observed that the presence of mental health disorders combined with limited access to important resources were the factors that substantially differentiated the two groups (Table 1). Those who had dropped out described problems that went unnoticed by the teachers and the other students. Their internalizing symptoms and learning difficulties were therefore not given much attention in class and thus, they had to cope on their own most of the time. They experienced only rare, arbitrary and short-term interventions. Moreover, they had few other resources available when struggling to stay in school. One example of this was the fact that among the six participants experiencing mental health problems, only one described outreach initiatives from mental health care services while in school. For several of the others, help was accessed only after the dropout event. These descriptions are in line with a study of Norwegian adolescents indicating that even among individuals with the highest symptom-loads of anxiety and depression, help-seeking is still low (Zachrisson, Rødje, & Mykletun, 2006). The researchers argue that the ability of the adolescents and their parents to recognize symptoms as signs of internalizing mental disorders is the first "filter" on the way to treatment. Externalizing problems, however, seem easier to recognize. Nevertheless, Frøseth and Markussen (2009) observed that students with less serious behaviour problems, accessing fewer resources, more often dropped out, while there was less dropout in students with the more serious behaviour disorders. They commented that the low dropout rate among students with the more serious disorders most likely was due to the support they received through accessing extra resources. Furthermore, 45-60% of students with externalizing disorders received clinical health services, while only 18% of those with an anxiety disorder, and 38% of those with any other mood disorder received treatment for their diagnosis (Merikangas et al., 2011). Thus, students with internalizing disorders seem to be underrepresented in school mental health care. A part of the problem may be that a high proportion of teachers seem unable to identify high levels of depressive symptomatology in students (Auger, 2004). This may be of particular importance because children depend on close teacher-child relationships to openly express feelings and concerns, and thus evoke the necessary help and guidance (Birch & Ladd, 1997). Internalizing students can be quiet in class, and thus may not get much attention (Henricsson & Rydell, 2004). They may thereby be unable to establish the close teacherchild relationship necessary to access the resources that could be provided by the teacher. There are also lower levels of interrater correspondence when rating child-internalizing problems than when rating externalizing problems (De Los Reyes & Kazdin, 2005). These studies suggest that there is a problem with discovering and assessing internalizing symptoms accurately. Furthermore, some children with internalizing problems seem to come to class lacking sufficient social support resources. Internalization may be associated with disturbances in the child's relational network (Hammen, 1999). Finally, resources like early identification of problems and provision of support are important to academic outcomes (Gnambs & Hanfstingl, 2015;Riglin et al., 2014;Saele et al., 2016). Consequently, young people with internalizing disorders like the ones in our study, seem to be at a disadvantage that may affect their school attainment and eventually their work employment.

Conclusion
This study indicated that internalizing symptoms might be more problematic for some students than others, making it vital to identify those with internalizing disorders who are at risk of dropping out and spending years struggling to get back on track. In the present study, students describing a combination of intensified internalizing symptoms and a lack of appropriate adult support, failed to establish autonomous adult lives years after they had dropped out of high school. Studies with statistical designs are needed to investigate further the relative occurrence of internalizing disorders in the population of those who have dropped out of school and employment and are struggling to get back on track during emerging adulthood.

Funding
The publication of the article was supported by the Open Access Fund of UiT The Arctic University of Norway.