Girls suffer: the prevalence and predicting factors of emotional problems among adolescents during upper secondary school in Norway

This longitudinal, quantitative survey examined factors predicting 1077 Norwegian adolescents` emotional problems during Upper Secondary School (grades I–III, approximately 16–19 years old) considering the following research question: “To what extent do students in Upper Secondary School experience emotional problems, and how are these problems predicted by gender, academic/social self-concept, coping beliefs, appearance pressure and school stress?”. The mentioned variables were scrutinized through analysis of frequencies, zero order correlations and structural equation modeling. Results verified previous findings that there was an increase in emotional problems for adolescents, especially among girls. Furthermore, the study results indicate that coping beliefs is a crucial factor when it comes to the perception of pressure and stress, and the subsequent development of emotional health problems.


Introduction
Mental health problems are a matter of major public health concern and are among the leading contributors to the health burden for children and adolescents globally (Erskine et al., 2015;Whiteford et al., 2013). The literature usually separates between mental problems and disorders. Mental health problems refer to symptoms that have significantly negative impact on well-being, learning, daily tasks and socializing with others, while they do not satisfy the criteria for a psychiatric diagnosis (Mykletun et al., 2009). A mental disorder can be assessed when the symptoms are in coherence with the criteria in a diagnostic system, such as ICD10 (World Health Organization, 2018) or DSM-5 (American Psychiatric Association, 2013). Mental health disorders are concluded by health personnel based on diagnostic interviews, while it is common to identify the extent of mental health problems through the use of questionnaires. In the following we refer to mental health problems when we discuss mental health among adolescents.
About 10-20% of children and adolescents in the world suffer from mental health problems (Henderson et al., 2017;Kieling et al., 2011). This can have serious consequences for the quality of life as mental health problems during adolescence are related to weak academic and social functioning (Balazs et al., 2013;Rasing et al., 2020). Studies of prevalence from the 80s into the 21st century suggest a global increase in both mental health problems and disorders, especially internalized symptoms among girls (Bor et al., 2014;Choi, 2018). This tendency is supported by studies from the USA (Merikangas et al., 2010), the UK (Collishaw, 2015) and other countries in the western world (Auerbach et al., 2018). These circumstances have led to strengthened emphasis on how teachers, counsellors and the school as a whole can reduce contextual stressors and facilitate adolescents` ability to cope with life challenges. Based on this, research on risk-and protective-factors connected to coping has increased in the last few decades.
Coping is defined as the "constantly changing cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person" (Lazarus and Folkman, 1984, p. 141). However, coping factors cannot be reduced to purely individual, psychological traits (Condly, 2006;Herrman et al., 2011). Other definitions emphasize the ecological influence on coping, stating that risk-and protective-factors include: "(…) both the capacity of individuals to navigate their way to health-sustaining resources (…) and a condition of the individual`s family, community and culture to provide these health resources and experiences in culturally meaningful ways" (Ungar, 2008, p. 225). This ecological perspective on coping is supported by research indicating that the quality of relationships with significant others is of considerable importance (Currie et al., 2012;Lerner, et al., 2005), in addition to broader environmental factors such as a safe and inclusive classroom environment (Skaalvik and Skaalvik, 2013;Theron, 2016).
In the present study, we will demarcate the scrutiny of risk-and protective-factors to the associations between self-concept, coping beliefs, appearance pressure, school stress, and emotional problems. These are all individual perceptions, and we recognize that coping encompasses a dynamic, bi-directional process in which the environment and the individual interacts constructively over time. In such, contextual factors in both proximal microsystems (e.g. family, school, friends) and in the more distal macrosystems (e.g. politics, culture) affect coping during adolescence (Lerner et al., 2005, Ungar, 2008. However, numerous findings also indicate that individuals` cognitive perceptions and interpretations are important when understanding the development of stress, perceptions of pressure, and mental health problems (e.g.

3
Girls suffer: the prevalence and predicting factors of emotional… Fletcher and Sarkar, 2013;Jongen, McCalman and Bainbridge, 2019). From the perspective of Lazarus and Folkman's (1984) transactional stress model, adolescents` emotional problems depend not only on the objective challenges or stressors they are exposed to, but also on how they subjectively evaluate the situation and beliefs in their own ability to handle it. Thus, their transactional model of stress claims, in line with other research (e.g. Sowislo and Orth, 2013;Yeoh et al., 2017), that selfconcept and coping beliefs are decisive constructs to predict experiences of stress and subsequent emotional problems. Individuals with high levels of coping beliefs are more likely to view difficult tasks as challenges rather than things to be avoided. Thus, they are more likely to experience less distress and, and subsequently, fewer emotional problems (e.g. Bandura et al., 1999;Orth et al., 2014).
In the following, we will review research on the associations between adolescents' perceptions of emotional problems, self-concept, coping beliefs, appearance pressure and school stress. This literature review will end with a formulation of a primary research question with six related hypotheses. In the next section we describe methodological aspects, including descriptions of the participants and our procedure, instruments and data analysis. Thereafter, we present findings and discuss the self-reported experience of Norwegian Upper Secondary School students in light of earlier research and theory. This is followed by a discussion of the limitations with the study, practical implications and future directions for research. The article ends with conclusive remarks.

Self-concept
Numerous studies have found that mental health is linked to the experience of selfconcept during adolescence (e.g. McCarty et al., 2007;Ohannessian et al., 1999). In the literature, self-concept is generally defined as a person`s perceptions and evaluations of personal qualities, abilities, and knowledge (Esnaola et al., 2020;Ohannessian et al., 2019), in academic, physical, and social domains (Hankin et al., 2015;Van Keyserlingket al., 2019). The literature expresses increasing evidence that supports a multidimensional perspective of self-concept, often based on Shavelson, Hubner and Stanton`s (1976) multidimensional, hierarchical, and domain-specific model of self-concept, divided into academic and non-academic components. Academic self-concept is further divided into self-concepts in particular subject areas (Mathematics, English, etc.) and non-academic self-concept is divided into social, emotional, and physical self-concepts. This is supported by Marsh and Craven (1997) who argue that "specific domains of self-concept are more useful than a general domain" (p. 191) in order to understand the complexity, predict a wide variety of behaviors, relate self-concept to other constructs, and provide outcome measures for diverse interventions in educational settings.
Previous research indicates that self-concept is an important mediator factor for psychological, social, behavioral and educational outcomes, such as personal development, socialization processes, motivation, coping beliefs, perceptions of stress, 1 3 and mental health (Craven and Marsh, 2008;Marsh and Martin, 2011). Taking into account the multidimensionality nature of the self-concept, it is necessary to explore the predictive effects on mental health across several dimensions as the developmental patterns may differ. In this study we will emphasize two of these dimensions: academic self-concept and social self-concept.
Academic self-concept is related to learners` knowledge and perceptions about themselves in overall academic domains (Marsh and Martin, 2011;Wigfield and Karpathian, 1991). Marsh (1989) found that the academic self-concept reaches its lowest point in grades 8 or 9 and subsequently increases through late adolescence, while other studies have found a consistent decline during secondary education (De Fraine et al., 2007;Nagy et al., 2010). Some studies conclude that males score higher than females (De Fraine et al., 2007;Pinxten, et al., 2013), while others find an advantage for females (Young and Mroczek, 2003), or report no significant differences (Esnaola et al., 2018;Marsh et al., 2005). Few studies have investigated the development of social self-concept. However, some studies have found that lower social self-concept is associated with increased emotional problems during mid-late adolescence (Kim-Spoon et al., 2012;Lee et al., 2010).
Based on these findings, it is relevant to underline that Shavelson et al., (1976) emphasized the importance of self-attributions, and suggested that forms of behavior and dimensions of self-concept have correlated and reciprocal relations. This is supported by Marsh and Craven (2006) who reviewed research on academic outcomes and academic self-concept. They found the correlation between academic self-concept and academic achievement (e.g. school grades) to have the strongest association. However, they could not find correlations between other non-academic components of self-concept and academic outcomes, underlining the multidimensionality of the construct (Marsh and Martin, 2011). In sum, the above-mentioned findings indicate that both academic self-concept and social self-concept are directly and positively related with coping beliefs, and negatively related to emotional problems (Craven and Marsh, 2008;Marsh and Martin, 2011).

Coping beliefs
In a systematic review and structural equation modeling meta-analysis of 19 studies, Groth et al. (2019) found that coping is predicted by coping beliefs and subsequently predicts mental health outcomes. Four longitudinal studies indicated that coping beliefs were significantly associated with eating problems (Halvarsson-Edlund et al., 2008), anxiety (Lopez and Little, 1996;Weigold and Robitschek, 2011), and depression (Pérez et al., 2009). Furthermore, in light of their findings, Groth et al. (2019) suggest that individuals who report low coping beliefs tend to underestimate the potential positive consequences of their behavior and the likelihood to be rewarded for this behavior. These findings can be interpreted in light of the theory of learned helplessness (Abramson et al., 1978), which claims that learned helplessness occurs when individuals do not experience that highly desired outcomes are contingent on their responses. This may lead to an attribution of challenges and demands as uncontrollable, prompt individuals to use maladaptive coping strategies, such as self-pity,

3
Girls suffer: the prevalence and predicting factors of emotional… and thereby a predisposition toward reactive depression when experiencing failure or negative life events (Lazarus and Folkman, 1984;Rotter, 1966). In this study, we will use the concept of coping competence as an indicator of coping beliefs. Coping competence is defined as a belief in the capacity to "(…) effectively cope with failure and negative life events as indicated by a reduced likelihood of helplessness reactions and fast recovery from any occurring helplessness symptoms" (Schroder and Ollis, 2013, p. 288). From this perspective, coping beliefs (competence) is understood as a component of emotional well-being and resilience against learned helplessness dispositions and helplessness-based reactive depression.
In sum, coping beliefs seems to be crucial for both the individuals` interpretation of the threat level, their own abilities and possible coping strategies. When the expectation of coping is low, various challenges, demands and changes that we are exposed to will eventually be perceived as stressful, while high coping will make us more robust to cope with challenges we face in life and thus reduce the likelihood that we will experience challenging situations as stressful. Based on this, we expect that coping beliefs is directly and negatively related to perceptions of school stress, appearance pressure and emotional problems (Sowislo and Orth, 2013;Yeoh et al., 2017).

Appearance and school stress
This article investigates the associations between emotional problems and adolescents' perceptions of performance-related stress in two areas: appearance and school. According to a study that compared the student reports on the perceptions of school pressure between 1994 and 2010, the level of perceived pressure did not increase during the period, except for a temporary uptake between 2002 and 2006. However, recent research indicates that the prevalence of adolescent experience of school stress has increased during the last decade (Bakken et al., 2018;Löfstedt et al., 2019), and that there is a clear association between school stress and emotional problems (Låftman and Modin, 2012;Ringdal et al., 2020).
In a systematical review of 33 studies, Lillejord et al. (2017) report mixed findings concerning gender differences in school stress. Some studies find no differences (Coelho and Romão, 2016;Sotardi, 2016), while most studies find that girls report higher levels of stress compared with boys (Goldstein et al., 2015;Seiffge-Krenke, 2012;Sun et al., 2013;Yang and Cheng, 2016). To explain the differences, Lillejord et al. (2017) argue that girls undergo more pubertal changes than boys, emphasize social demands to a greater extent, become more stressed by unclear expectations and more frequently experience school-related burnout. In line with this, Samela-Aro and Tynkkynen (2012) point to research that has found that girls and boys experience stress differently. Girls internalize symptoms such as depression or fatigue and develop a feeling of not coping with pressure. In addition, girls are more likely to experience competitive situations and that they are more exposed to the negative effects of stress, while boys may tend to develop an externalized reaction pattern such as adapting a cynical attitude toward school. This finding is supported by Suldo and Shaunessy-Dedrick (2013) and Wahab et al. (2013). In sum, it seems that girls are more influenced by the school context than boys (Dalen, 2014;Låftman et al., 2013). In the high-performing school context that Låftman et al. (2013) report from, students believe that there is a culture of stress at school and that the students, especially girls, influence each other and exaggerate the stress. This is confirmed by Sonmark et al. (2016), who found that being surrounded by classmates who feel pressured is more detrimental to the mental health of girls than boys.
Furthermore, there is an extensive research tradition on adolescents` attitudes toward their own body and appearance. Studies show that girls have a more problematic body image than boys (Bakken et al., 2018;Skaalvik and Federici, 2015). Also, these gender differences increase beyond adolescence (Bearman et al., 2006;Jónsdóttir et al., 2008), and are more closely related to self-concept among girls than boys (Alm and Låftman, 2018). Dissatisfaction with one`s own body and appearance has a stronger association with mental health problems among girls than boys (Hargreaves and Tiggeman, 2003;Bearman and Stice, 2008). In sum, based on these findings, we expect that both appearance pressure and school stress are directly and positively related with emotional problems.

The present study
The present study was designed to explore how students perceive life in school in general, and the extent of emotional problems in particular. Even though mental health and well-being are becoming a health priority in most western countries, evidence concerning the prevalence and predictive factors over time remains mixed (Cosma et al., 2020;Rasing et al., 2020). Thus, this longitudinal study is designed to contribute to this knowledge gap in the literature. Based on the literature review, the following research question was formulated: "To what extent do students in Upper Secondary School experience emotional problems, and how are these problems predicted by gender, academic/social self-concept, coping competence, appearance pressure and school stress?". Based on this research question and our findings from the literature review, six hypotheses were formulated, and a theoretical path model was specified: • H1: The extent of emotional problems among adolescents has increased (Bor et al., 2014;Choi, 2018). • H2: Academic self-concept is directly and positively related with coping competence, and negatively related to emotional problems (Craven and Marsh, 2008;Marsh and Martin, 2011). • H3: Social self-concept is directly and positively related with coping competence, and negatively related to emotional problems (Kim-Spoon, Ollendick and Seligman, 2012; Lee, Hankin and Mermelstein, 2010). • H4: Coping competence is directly and negatively related with school stress, appearance pressure and emotional problems (Abramson et al., 1978;Schroder and Ollis, 2013) • H5: Appearance pressure is directly and positively related with emotional problems (Alm and Låftman, 2018;Bearman and Stice, 2008).

3
Girls suffer: the prevalence and predicting factors of emotional… • H6: School stress is directly and positively related with emotional problems (Låftman and Modin, 2012;Ringdal et al., 2020).

Participants and procedure
This study was part of a larger data collection and analysis of students` perceptions of their life in school. The sample for the present article comprises 1077 students (56% response rate), 606 girls and 471 boys, through Upper Secondary School (grades I-III, approximately 16-19 years old), from both rural and urban areas in thirteen schools from one county in Norway. Data can be described as a convenience sample (McQueen and Knussen, 2006), and were collected in the fall of 2015 (USI), spring of 2017 (USII), and spring of 2018 (USIII). The participating students were informed in advance that the participation in the study was voluntary and that they were considered to have given their consent by filling in the questionnaire. The data were collected with paper-based questionnaires and administrated by members of the research team according to the guidelines of the Norwegian Social Data Services. The Norwegian Data Inspectorate approved the survey.

Instruments
The dependent variable was measured on a three-point scale ranging from "Not true" (0), "Somewhat true" (1), and "Certainly true" (2). The independent variables, except grades from Lower Secondary School (0-6) and Appearance Pressure (1-4), were measures with items answered on a six-point Likert scale ranging from 1= very untrue to 6= very true, and the reliability was measured by Cronbach`s alpha.

Emotional problems
The dependent variable Emotional Problems scale was constructed from the Strength and Difficulties Questionnaire (SDQ). SDQ (Goodman, 1997(Goodman, , 1999) is a brief assessment of mental health problems that consist of 25 questions covering emotional symptoms, conduct problems, hyperactivity-inattention, and peer problems as well as prosocial behavior. SDQ was originally constructed for children between 11 and 16 but has recently been used in older age groups (e.g. Bøe et al., 2016;Vugteveen et al., 2019). SDQ Emotional Problems were measured by five statements that each were scored from 0 to 2, making a possible score from 0 to 10. Examples of statements: "I worry a lot", "I am often unhappy, down-hearted or tearful". Cronbach`s alpha for the scale was good for both girls (.71) and boys (.73).

Academic self-concept
Academic self-concept was measured by four items based on the "Self-Description Questionnaire III (SDQIII), originally developed by Marsh (1992), and acknowledged as a leading multidimensional self-concept instrument for adolescents (e.g. Boyle, 1994;Hattie, 1992). Examples of statements: "I cope with the challenge of learning new material at school", "When I sit down to learn something really difficult, I can handle it". Cronbach`s alpha for the scale was good for both girls (.89) and boys (.86).

Social self-concept
Social self-concept was also measured by four items based on the SDQIII (Marsh, 1992). Examples of statements: "It is easy for me to make friends", "Most people like me". Cronbach`s alpha for the scale was good for both girls (.84) and boys (.87).

Coping beliefs
Coping beliefs was measured by five items from The Coping Competence Questionnaire (Schroder and Ollis, 2013). This instrument had originally 12 statements, and we used the 5 statements with the highest loadings from the factor analysis. All the statements are formulated negatively and the variable score were turned before we made the scale. Examples of statements: "When I do not succeed right away, I think I will never get it", "When I perform poorly at school, I begin to doubt my abilities". Cronbach`s alpha for the scale was good for both girls (.84) and boys (.87).

Appearance pressure
Appearance pressure was measured using two items based on a scale developed by Skaalvik and Federici (2015). Examples of statements: "I have felt a strong pressure to look a certain way in recent months", "I have felt a strong pressure to dress in a special way in recent months". Cronbach`s alpha for the scale was good for both girls (.66) and boys (.71)

School stress
School stress was measured using three items based on a scale developed for the international survey, Health Behavior in School-Aged Children (Samdal et al., 2016), organized and administrated by the World Health Organization. Examples of statements: "I feel exhausted because of schoolwork", "I`m stressed by schoolwork". Cronbach`s alpha for the scale was good for both girls (.79) and boys (.77).

Data analysis
At first, we compared the full SDQ Emotional problems scale data with British norm-data (Goodman et al., 2000), and compared this with a survey from 2000 that included 19 210 adolescents from Norway (Rødje, 2004).
After this, an exploratory factor analysis was conducted on six sum scales (SDQ emotional problems, academic self-concept, social self-concept, coping competence, appearance pressure, and school stress) with a total of 23 statements (see Method for details). Here we found that one item ("often complains of headaches, stomachaches or sickness") in the SDQ Emotional problem scale loaded on the School stress factor and decided to leave this factor out. This resulted in an exploratory factor analysis with a total of 22 statements. The Kaiser-Meyer-Olkin value was .89, exceeding the recommended value of .6 (Kaiser, 1970;1972) and Bartlett`s Test of Sphericity (Bartlett, 1954) reached statistical significance, supporting the factorability of the correlation matrix. Principal component analysis revealed the presence of six components with eigenvalues exceeding 1. The six-component solution explained a total of 68% of the variance. To aid the interpretation of these three components, oblimin rotation was performed. The rotated solution revealed the presence of a simple structure (Thurstone, 1947) with all six components showing strong loadings and all variables loading substantially on only one component.
Furthermore, the hypothesized model of the connections between the variables presented in Figure 1 were tested statistically to explore to what degree it was coherent with the observed data. Structural equation modeling (SEM) in the AMOS 26 program was used to analyze the model with latent (unobserved) variables. None of the error terms were allowed to correlate. The coherence between observed data Fig. 1 Theoretical path model and the hypothesized model is reported as the goodness of fit statistics. The goodness of fit indicators used to assess the model are Non-Normed Fit Index (NNFI, also known as TLI), Comparative Fit Index (CFI), and Root Means Square Error of Approximation (RMSEA), the study also includes chi square, RMSEA ≤ .07, TLI ≥ .90, CFI ≥ .90 are considered as indicators of acceptable fit (Byrne, 2001;Hu and Bentler, 1999). This indicates a plausibility of the associations between the constructs. The model was tested with the whole sample, and then separate analyses were performed for boys and girls, because research indicates gender differences in the importance of the hypothesized relation between constructs. Table 1 shows the distribution of SDQ Emotional Problems through Upper Secondary School for both girls and boys. These are interpreted based on the original threeband categorization from British norm-data (Goodman et al., 2000), and compared with a survey from 2000 that included 19 210 adolescents from Norway (Rødje et al., 2004).

Distribution of SDQ emotional problems
British norm-data indicate that 90% of adolescents will score within the "Normal" (0-5), 5% within the "Borderline" (6), and 5% within the "Abnormal" categories. The 2000 survey from Norway (Rødje et al., 2004), shows that girls (11%) reported abnormal levels of emotional problems almost three times as much as boys (4%). In addition, girls (8%) reported borderline levels of emotional problems over twice as much as boys (3%). Compared to the British norms we can conclude that boys reported less emotional problems than expected, and girls reported about twice as many abnormal and borderline levels of emotional problems.
Our data show that boys score about the same as expected from the British normdata and is distributed with almost exactly the same percentages as in the study from Norway (Rødje et al., 2004). However, when interpreting the scores from the girls in our study we can see another development. The girls report about four times as many emotional problems as the British norms and as the boys in our study. In addition, we can see that the extent is almost doubled compared with the 2004 study and that the emotional problems seem to increase as the girls grow older with 22% of 1 3 Girls suffer: the prevalence and predicting factors of emotional… girls reporting abnormal levels of emotional problems in their last spring of Upper Secondary School. In sum, these findings support H1.

Zero order correlations between observed variables
Zero order correlations between the study variables as well as N, range, number of items, Cronbach`s alpha, statistical means, standard deviations and Cohen`s D for boys and girls are shown in Table 2.
All correlations between emotional problems and the independent variables were significant for both girls and boys. The correlations between the dependent variable emotional problems and the independent variables academic self-concept, social self-concept and coping competence were negative, ranging from r = −.20 to r = −.57. This indicates that the higher the students perceive emotional problems, the lower they perceive academic self-concept, social self-concept, and coping competence. The correlation between emotional problems and the independent variables appearance pressure and school pressure was positive, indicating that the higher students perceive appearance pressure, the higher they perceive emotional problems. Most of the correlations between the observed variables are similar between girls and boys, however the negative correlation between emotional problems and coping competence seems to be stronger for girls (−.57) than for boys (−.38).
In addition, as the distribution of SDQ emotional problems in Table 1 indicated, the means of the dependent variable were significantly higher for girls (3.53) compared with boys (2.01). A similar difference is the mean for appearance pressure (1.94/1.56) and school stress (4.32/3.58). An opposite pattern is shown for academic self-concept (4.57/4.80), social self-concept (4.43/4.63) and coping competence (4.07/4.72), with higher scores for boys compared with girls. Considering effect size, Cohen (1988) claims that an effect of 0.2 is small, 0.5 is medium, and 0.8 is high. The Cohen's D measure thus indicates that the difference effect size is small for academic self-concept (.29) and social self-concept (.22). Medium for coping competence (.62), appearance pressure (.54), and school stress (.65). The effect size is close to strong for the dependent variable emotional problems (.77).

Structural equation model
The relations between the variables were further analyzed by means of SEM analysis for latent variables using the AMOS 26 program as shown in the theoretical model in Figure 1. The model had a satisfactory fit to data for both girls and boys: CFI = .945/933, TLI = .931/915, RMSEA = .050/.053. Figure 2 shows the SEM between academic self-concept (US I), social self-concept (US I), coping competence (US II), appearance pressure (US III), school stress (US III), and emotional problems (US III), and Table 3 shows the direct, indirect and total effects, for both girls and boys. Figure 2 and Table 3 indicate that the correlation between academic self-concept and social self-concept are positive (r = .51) for both girls and boys. Academic selfconcept and social self-concept in the first year of Upper Secondary School has a

3
Girls suffer: the prevalence and predicting factors of emotional…   Table 3 indicates that the two self-concept variables have indirect effects on appearance pressure, school stress and emotional problems. However, we could not find any direct significant relationships between the self-concept variables and the perception of appearance pressure and school stress. In addition, of the two selfconcept variables, it is only social self-concept that has a direct negative relationship with emotional problems (β girls/boys = −.17/−21). Coping competence has a significant direct negative relationship with appearance pressure (β girls/boys = −.50/−.39), school stress (β girls/boys = −.51/−.36), and emotional problems (β girls/boys =−.31/−.18) for both girls and boys, respectively. In addition, Table 3 indicates that coping competence also has a considerable indirect effect (β girls/boys =−.31/−.26), on emotional problems through appearance pressure and school stress, resulting in a total effect of β =−.61 for girls and β =−.44 for boys. The model explains 25% (girls) and 15% (boys) of the variance in appearance pressure, and 26% (girls) and 13% (boys) of the variance in school stress.
Both appearance pressure (β girls/boys =.25/.30) and school stress (β girls/boys =.36/.39) have a significant positive relationship with emotional problems for both girls and boys, respectively. In all, the structural equation model explains 63% of the variance in emotional problems for the girls, and 49% for the boys.

Discussion
The primary goal of this study was to further explore the prevalence of emotional problems, and a proposed pathway between emotional problems among Upper Secondary School students, and their perceptions of academic self-concept, social selfconcept, coping competence, appearance pressure and school stress. The following research question guided the study: "To what extent do students in Upper Secondary School experience emotional problems, and how are these problems predicted by gender, academic/social self-concept, coping competence, appearance pressure and school stress?" Based on this research question and our findings from the literature review, six hypotheses were formulated, and a theoretical path model was specified. Our findings mainly confirmed previous studies.
The first main finding is that the prevalence of emotional problems among adolescents has increased considerably compared with the 2000 survey from Norway (Rødje et al., 2004). This supports H1 and previous research (Bor et al., 2014;Choi, 2018). However, even though it is expected, the large effect size difference between boys and girls is somewhat surprising. The zero-order correlations between the dependent and the independent variables reveal that the relationships are somewhat the same for both girls and boys, with the exception of the association between emotional problems and coping beliefs (competence). This relationship is stronger for girls compared to boys, indicating that girls may be more vulnerable for developing learned helplessness (Abramson et al., 1978;Schroeder and Ollis, 2013). This notion is supported when analyzing the Cohen's D effect size difference between the genders. Girls perform significantly higher than boys in Lower Secondary School, but still they have lower academic self-concept and social self-concept. In addition, the coping belief, measured as coping competence, is considerably lower for girls compared to boys, while the perception of appearance pressure and school stress are higher to the same degree. This imbalance may explain the high effect size difference in emotional problems. In sum, this finding corresponds well with other research that has found that girls are at greater risk of developing emotional problems, reporting lower levels of self-concept, coping beliefs, appearance pressure and school stress (Caprara et al., 2010;Muris et al., 2015).
The other main finding is that the proposed pathway in the theoretical model presented in Figure 1 had a good fit with the data, especially for girls, as it explained 63% of the variance in emotional problems. This indicates that academic self-concept and social self-concept, coping competence, and the perceptions of appearance pressure and school stress predicts emotional problems among adolescents in a prominent manner. In support of earlier research, the correlation between the selfconcept variables was high (Marsh and Craven 2006;Marsh and Martin, 2011). In addition, Shavelson et al.`s (1976) theory on the multidimensional nature of selfconcept was underlined by our finding about the difference of whether, and with what strength, academic self-concept and social self-concept predicted the other variables in the model. As expected, the self-concept variables predicted the adolescents' perception of coping competence in USII in similar strength for both girls and boys. This supports H2, H3 and previous research (Craven and Marsh, 2008;Marsh and Martin, 2011). Furthermore, it was expected that social self-concept in USI was significant for both girls' and boys' perceptions of emotional problems in USIII. However, in contrast with previous findings on the longitudinal effects academic self-concept has on emotional problems (Craven and Marsh, 2008;Marsh and Martin, 2011), we could not find any significant direct negative association to emotional problems, even though there were some indirect effects through coping beliefs. This finding is surprising and needs further scrutiny in future research Another interesting finding from the structural equation modeling is that coping competence has a significant negative relationship with appearance pressure, school stress and emotional problems for both girls and boys. However, the strength of these direct and indirect associations is somewhat stronger for girls compared with boys, indicating that coping competence seems to be important for both genders, and most important for girls, when it comes to predicting perceptions of pressure and negative health outcomes. These findings support H4 and previous research (Sowislo and Orth, 2013;Yeoh et al., 2017). In addition, in support of H5, H6 and previous research (Alm and Låftman, 2018;Bearman and Stice, 2008;Låftman and Modin, 2012;Ringdal et al., 2020) appearance pressure and school stress had a direct and positive relationship with emotional problems. However, it was more surprising that these associations were somewhat stronger for boys compared with girls, even though the differences were small.
In sum, the findings can be discussed in light of other studies that have shown that the level of coping beliefs plays a significant role in how different kinds of perceived demands are experienced (e.g. appearance pressure and school stress), and affects the risk that stress leads to emotional problems (Caprara et al., 2004;Shelley and Pakenham, 2004). Several studies conclude that increasing individuals` coping beliefs and expectations of coping may reduce the negative effect of the various stressors we are exposed to (e.g. McKay, Dempster and Byrne, 2014). In a longitudinal study, Burger and Samuel (2017) investigated how perceived stress and coping beliefs play a role in influencing adolescents` satisfaction with life. They found that perceived stress was negatively correlated with life satisfaction, whereas the opposite was true of coping beliefs. In addition, the individuals` basic expectation of coping further moderated the negative effects of stress. Thus, a high level of coping beliefs neutralized the negative effects of stress on adolescents` satisfaction with life. Furthermore, our coping beliefs today are related to the coping experiences we have had in the past (Groth et al., 2019;Samdal et al., 2017). A situation where we are exposed to stress stimuli can result in both positive and negative response outcomes. Positive outcomes represent situations in which individuals experience mastery. They have experienced being challenged, but also that they can cope with it. Negative outcomes are characterized by individuals experiencing that they do not cope. At worst, they will experience helplessness ("I can`t do anything about the situation"), and hopelessness ("things go wrong anyway and it`s my own fault") (cf. Abramson et al., 1978;Schroder and Ollis, 2013). A number of such negative outcomes contribute to both low coping beliefs and sustained stress activations, which in turn can lead to negative health outcomes.

Perceptions of internal/external demands/resources as predictors of emotional problems
Why do we observe this recent increase in mental health problems? How can this be explained? In light of Lazarus and Folkman's (1984) transactional model of stress the causes of mental health problems can be placed in two main groups: resources and demands. These can each be further divided in two: internal and external. To maintain emotional health, resources and demands have to be perceived as balanced by the individual. Thus, to reduce emotional health problems, we have to consider how we can enhance the adolescents' perceptions of internal and external resources, and reduce the perception of internal and external demands. Considerable research has investigated how to enhance adolescents' internal and external resources, especially in a school context. The first meta-review over research on mental health promotion and problem prevention in schools was conducted by the European Union Dataprev project (Weare and Nind, 2011). This project included 52 systematic reviews and meta-analyses of mental health in schools and the results show that interventions have a wide variety of beneficial effects on children, families and communities on a range of social, emotional and educational outcomes. A conservative estimate from the London School of Economics and Public Health England shows that for every pound Britain invests in social and emotional learning (SEL) school programs to prevent depression among children and young people, the UK saves five times the investment cost in just two years (McDaid et al., 2017). These findings were supported by a meta-analysis of 81 studies compromising 31 974 school students (Werner-Seidler et al., 2017), and another 1 3 Girls suffer: the prevalence and predicting factors of emotional… meta-analysis of 82 school-based social and emotional learning programs (Taylor et al., 2017). Although the effect sizes are small, these studies provide solid evidence that mental health can be strengthened, and emotional problems can be prevented, through school measures. The best-tested school programs are based on social and emotional learning, cognitive-behavioral, or multifamily therapy (MFT). A metaanalysis of 146 RCT-studies found that it is not sufficient to disseminate information only, e.g. through teaching materials or with educational or psycho-educational measures (Stockings et al., 2016). Research suggests that the measures should also contain a psychological component with an element of practice, such as training in cognitive techniques. In addition, Mendelson and Eaton's (2018) meta-analysis indicates that internet-or computer-based prevention strategies and mindfulness-based interventions are promising areas for further development and need further research.
On the other hand, there are a number of increased internal and external demands that twentifirst century adolescents experience that potentially can have negative consequences for their emotional well-being by contributing to the increasing prevalence of mental health problems (Bor et al., 2014;Seiffge-Krenke et al., 2009). For adolescents living in a world characterized by rapid changes and technological advancement, it is important to understand that different trends and factors can be related to their emotional well-being and ill-being. This does not mean that the fundamental elements of mental health have changed. However, as a result of various contexts and the environment that children face today, different trends arise, and accordingly various approaches exist to mitigate the negative effects. An example is the Programme for International Student Assessment (PISA), that was introduced for the first time in 2000. This was aimed to provide a "global" comparison across nations to determine the relative quality of each country`s educational system (OECD, 2001). This cross-national comparison led to considerable debate, especially among the countries with the lowest scores, and facilitated reforms in educational policy and practice, and enhanced expectations, demands and competition among schools and students (Cosma et al., 2020;Klinger et al., 2015). In addition, several theorists argue that too much responsibility and accountability has been placed on the internal and individual. Research has found a rise in internal perfectionism expectations and demands connected to both school and appearance that may contribute to increases in stress and subsequent mental health problems (Cosma et al., 2020;Curran and Hill, 2019). This finding is in line with the individualization thesis that is based on the idea that collective identities associated with class, family and gender are weakened in favor of individual aspirations that have become more important (Beck, 1992). This notion is relevant for the increase in emotional health problems in several ways. First, it is up to each individual to create their own future, find out "who one is", set goals, and reach them (Giddens, 1991;Illeris et al., 2009). Consequently, the responsibility for success and for one`s own destiny also increases, especially within the school system. Second, the individualization thesis is relevant in this context because mental health problems themselves may be understood to be more individualized than before. While conditions such as neurasthenia, melancholy, sadness and nervousness were perceived as a consequence of culture a hundred years ago, it is claimed that similar conditions today are to a greater extent interpreted as a weakness of the individual.
Overall, our analyses of the two pressure areas: school and appearance, indicate that girls are more vulnerable than the boys. In both of these areas, the pressure is more closely associated with emotional problems among girls than boys. Thus, this indicates that school stress, pressure related to appearance, and coping beliefs is potentially a greater health risk for girls than for boys. One possible explanation is that girls, to a higher degree than boys, reflect their intrinsic value on how they succeed in school (Esnaola et al., 2018;Låftman and Modin, 2012). There is also reason to believe that the perception of personal appearance is more important for the girls` self-image than for boys (Alm and Låftman, 2018;Esnaola et al., 2018). Although many argue that the ideals for body and appearance also have changes for boys (Bassett-Gunter et al., 2017;Karazsia et al., 2017), these findings indicate that looking good is still more important for girls compared with boys. Thus, it may be that girls actually experience more pressure in these areas, and also interpret the pressure as more threatening to their own identity. Because external stressors are perceived and interpreted subjectively, based on personal values, the social roles we hold, and from past experiences, the same amount of actual demands and expectations can be experienced without problems for some and pose a health risk for to others (Avison, 2010;Thoits, 2010). At the same time, the potential for experiencing a situation with a lot of pressure such as stress may be explained by gender differences in the amount of resources available in the situation and to differences in the coping strategies available to girls and boys. However, other research indicates that the suicidal rates are far higher among men compared with women (Oliffe et al., 2016;Rasmussen et al., 2017). This may indicate that stress and emotional problems among adolescent boys are underreported. Thus, we need more research to investigate long-term effects of perceptions of demands and resources across genders through childhood, adolescence and adulthood.

Limitations and future directions
This study has some limitations that should be acknowledged. First, all data were based on students` self-reports. Future studies should employ more objective measures of the social environment (e.g. behavioral observation), triangulate subjective perceptions (e.g. different informants such as friends, parents, teachers, coaches etc.), and make use of both quantitative and qualitative methods. Second, the data from this study are limited to the school context. Future studies should test the associations between environmental factors from several micro-systems, subjective needs and adolescents` emotional health via longitudinal and experimental designs. Third, this study has only considered individual-level variables, without considered the effect of the differences between social classes and schools on perceived ecology with relation to emotional health among individuals. Several studies have revealed that differences in the classroom or school environment exist across school level (e.g. Danielsen et al., 2010;Wang et al., 2018). Future studies may consider using a multi-level modeling approach to study the impact of class-, school-and societallevel factors.
Despite these limitations, the results in our study may have important clinical implications for current interventions aimed at preventing and reducing mental health problems in school. Our findings suggest that targeting perceptions of self-concept and coping beliefs will reduce feelings of pressure, stress and emotional problems. Integrative interventions should focus on enhancing adolescents` self-concept and coping beliefs as early as possible to disrupt any potential detrimental cascading effect on mental health. This is in line with the findings of intervention studies that reveal that coping beliefs seems to be an important mechanism to improve mental health outcomes (Kendall et al., 2016;Ohannessian et al., 2019).
A comprehensive understanding of pressures and stress among young people today should include knowledge of who is the agent behind the significant amount of pressure young people experience in everyday life. Adolescents can be both passive recipients of experienced pressure and active agents that exercise resistance. Knowledge about the various actors is not least important. Thus, teachers, counsellors and schools should aim to find good methods and structures that reduce the amount of pressure or make youth more robust against the pressure they experience in everyday life. To what extent do adolescents themselves contribute to the pressure they experience, and to what extent are they actors in reducing the pressure? Is this the solution to teaching each one to prioritize better in their own lives? To what extent and in what arenas do adults, such as parents, teachers, coaches, school authorities and commercial advertisers have a responsibility? And to what extent does pressure from the adult community work together with peer pressure-both analog and digital-to increase the amount of pressure and the vulnerability of that pressure? These are important issues for the social debate and should be explored further in future research.
Several replication studies are needed that can confirm, disprove, or nuance findings from earlier studies. Studies with multiple data sources are needed, such as interviews with parents, teachers, counsellors and school leaders, which can be combined with self-reported data. Future research should also examine a longer-term follow-up of comprehensive data on prevalence and outcomes, encompass more high quality rigorous randomized control trial designs with comparable control groups, and also study the potential effectiveness and feasibility of teachers delivering these program sessions and training. The aspect of investigating teacher delivery of efforts to improve students' mental health is important. As several studies have pointed out, teachers can play a vital role in delivering such prevention programs. Future studies should also investigate the practicality of implementation in terms of costs, additional burden on teachers, and the necessary support that teachers need to deliver this training and programs. In order to prevent an increase in teachers' workloads, which can reduce a teacher's emotional well-being and in turn that of students, it will be important think about ways in which social and emotional learning can be incorporated into the curriculum and existing dayto-day learning activities.

Conclusion
Although this study has some limitations, the results in the present study underscore the need for research on self-concept, coping beliefs and environmental pressure when it comes to understanding adolescents' coping and emotional problems. In sum, our findings can be illuminated in light of other research that has suggested that girls perceive more pressure than boys and at the same time are more vulnerable to the effects of this perception (Bakken et al., 2018;Hankin et al., 2008). Thus, girls perceive their external and internal demands as higher than boys but their external and internal resources are perceived to be lower. In light of the transactional theory of stress (Lazarus and Folkman, 1984) and the theory of learned helplessness (Abramson et al., 1978), these unfortunate connections may explain why girls are developing higher levels of emotional problems. These findings are of great significance for practical implications and applications. Teachers, counsellors and the school as a whole should increase their effort to facilitate the development of psychological resilience as this seems to be both a decisive protective-and risk-factor when it comes to adolescents' sense of stress and emotional well-being, especially among girls.
Funding Open access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital -Trondheim University Hospital). The authors received no financial support for the research, authorship, and/or publication of this article.

Conflict of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Informed consent Informed consent was obtained from all participants, Additionally, parents were sent an information letter with the option to opt their child out of the study. The study was approved by the Norwegian Data Inspectorate Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.