During the last couple of decades there has been an increase in mental illness among school-aged children and adolescents in Sweden. The Public Health Agency of Sweden (2018a) reports that the mental illness of adolescents in Sweden has increased since the mid 1980´s, with symptoms such as sleeping problems and depression, as well as psychosomatic problems such as headache and stomach ache. The decrease in the wellbeing of students can, according to The Public Health Agency of Sweden (2018b), partly be connected to school-related stress and partly to factors such as family and socioeconomic circumstances (Karvonen, Tokola, & Rimpel, 2017; Pollard & Lee, 2003). However, one central explanation brought forward was that the decrease in wellbeing is due to increasing demands on school results, more frequent testing and earlier grading (from Grade 6). According to the report, perceived stress related to schoolwork among 13-year olds has quadrupled from the previous measurement in 2010.

In Sweden, the assessment system has gone through major reforms since the beginning of the 1990s. In 2011 the requirements increased and today students need to have eight passing grades to be eligible for entering a vocational upper secondary educational program and 12 passing grades to be eligible to enter a theoretical upper secondary educational program. From 2011, the number of students not eligible to enter upper secondary education has increased. In 2016, 17.5% were not eligible to enter a theoretical program at upper secondary education due to failing grades (National Agency for Education, 2018). High-stakes summative assessments such as grades, seem to affect students´ achievement differently, and it has been found that low-achieving students when being graded at primary level have a worse grade development and graduate from upper secondary school to a lesser extent, compared to low-achieving students not being graded (Harlen & Deakin Crick, 2002; Klapp, Cliffordson & Gustafsson, 2014; Klapp, 2015, 2017). Summative assessments seem to have differential effects in that low-achieving students, in relation to high-achieving students, have difficulties in understanding grades, have a risk of developing negative learning strategies, need to work harder to get better results and higher grades, and have a higher risk of giving up (Harlen & Deakin Crick, 2002).

It has been suggested by research that the main sources of stressful phenomena for children and adolescents relate primarily to the school context and relationships. In the literature on well-being, researchers argue that wellbeing is a multidimensional construct defined by individual characteristics of an inherently positive nature (Pollard & Lee, 2003). The construct of wellbeing comprises five distinct domains: i.e. the physical, psychological, cognitive, social and economic domains. The social domain is about family and peer relations and support while the cognitive domain is about intellectual or school-related indicators of achievement. Indicators such as self-esteem and motivation reflect facets of the different domains of wellbeing and do not reflect wellbeing in its entirety.

Purpose

The overall purpose of the study is to investigate social and cognitive wellbeing for two birth cohorts of students born in 1998 and 2004. The assessment system changed profoundly for these two cohorts and by using the ETF data it is possible to investigate changes in social and cognitive wellbeing for these cohorts as well as reciprocal relations between educational achievement and wellbeing. Educational achievement are measured by National test result from 3^th to 9^th Grade (age 10-16) and Grade Point Average from 6^th to 9^th Grade (age 12-16). Another purpose is to investigate the differences between subgroups of students (gender, parents´ educational level) in the relation between wellbeing and academic achievement. Yet another purpose is to investigate these relations taking into account school level effects.

The subjects were 9175 students born 1998 and 5201 students born in 2004. Data was retrieved from the longitudinal project Evaluation Through Follow-up (Utvärdering Genom Uppföljning UGU) (Härnqvist, 2000) which is one of the largest research databases in Sweden and classified as a national infrastructure. It comprises background and questionnaire data for a national representative sample of every fifth cohort born between 1948 and 2004 along with register data. Information was collected by Statistics Sweden (SCB). It contains information about grades, standardized test results, national test results, school, class, class size and study choices etc. for almost all cohorts, at different time points. The data used are school subject grades (GPA) in the ninth Grade, when the subjects were 15-16 years old, information on gender, parents´ educational level. Questionnaire data from the 6th Grade was used. Questions such as “How content are you with the other pupils?”, “How content are you with school work?” with five response categories from very “poor” to “very good”. Further, “How do you feel in school? I have friends who I can be with in school?”, and “How do you feel in school? I am happy when I am in school” Five response categories ranging from “always” to “never”. Methods of analyses First, descriptive statistics were estimated. Second, confirmatory factor analyses (CFA) and structural equation models (SEM) will be estimated. The intra-class correlations (ICC) for the variables will be presented. To take account of possible clustering of students in schools (school level), the complex option offered by the Mplus program was used. As measures of model fit, the Root Mean Square Error of Approximation (RMSEA) and (the Standardized Root Mean Square Residual (SRMR) were used. The Comparative Fit Indices (CFI) measure was also used. This index should be as close to 1.0 as possible, and preferably above .95 (Bentler, 1990). The Tucker-Lewis Test (TLI) was used which is a measure similar to the CFI but which has a penalty for models with many parameters. The TLI should be as close to 1 as possible but values above .90 are considered acceptable (see for example Hu & Bentler, 1995). Tests of linearity and of the assumption of homoscedasticity of the residuals were made and no deviations were found. The analyses were conducted in the SPSS program, version 24 (2016) and in the Mplus program, version 5 (Muthén & Muthén, 1998-2015).

Initial steps have been conducted and further steps will be conducted in the modeling process. First, descriptive statistics were computed and measurement characteristics of the scales were investigated. The internal consistency of the scales for two hypothesized factors were good, CA = .86 for Socwell and CA = .75 for Cogwell. Then, a covariance model will be estimated with all the included variables (Socwell, Cogwell), the background characteristic variables, and the dependent variables GPA 9 and GPA12. Next, SEM models will be estimated with relations between the factors Socwell and Cogwell and GPA in different Grades and with covariance between the factors. In the final step, the manifest background variables will be included in the SEM model and direct and indirect effects of the wellbeing factors on achievement will be computed.

Bentler, P.M. (1990). Comparative indexes in structural models. Psychological Bulletin, 107, 238-246. Harlen, W., & Deakin Crick, R. (2002). A systematic review of the impact of summative assessment and tests on students’ motivation for learning (EPPI-Centre Review, version 1.1*). In: Research evidence in educational library. Issue 1. London: EPPI-Centre, Social Science Research Unit, Institute of Education. Härnqvist, K. (2000). Evaluation through follow-up. A longitudinal program for studying education and career development. In C.-G. Janson (Ed.), Seven Swedish longitudinal studies in behavioral science. Forskningsrådsnämnden: Stockholm. Karvonen, S., Tokola, K., & Rimpel, A. (2017). Well-being and academic achievement: differences between schools from 2002 to 2010 in the Helsinki Metropolitan Area. Journal of School Health, 88, 821-829. Klapp, A. (2017). Does academic and social self-concept and motivation explain the effect of grading on students´ achievement? European Journal of Psychology of Education, 33(2), 355-376. Klapp, A. (2015). Does grading affect Educational attainment? A longitudinal study. Assessment in Education: Principles, Policy and Practice, 22(3), 302-323. Klapp, A., Cliffordson, C., & Gustafsson, J-E. (2014). The effect of being graded on later achievement: evidence from 13-year olds in Swedish compulsory school. Educational Psychology: An International Journal of Experimental Educational Psychology, 36(10), 1771-1789. Muthén, B. O., & Muthén, L. K. (1998–2015). Mplus user’s guide. Seventh edition. Los Angeles, CA: Muthén & Muthén. National Agency for Education (2018). Snabbfakta. Retrieved 2018-03-28 at https://www.skolverket. se/statistik-och-utvardering/statistik-i-tabeller/snabbfakta-1.120821. Pollard, E.L., & Lee, P.D. (2003). Child well-being: A systematic review of the literature. Social Indicators Research, 61(1), 59-78. The Public Health Agency of Sweden (2014). Health behavior in school-aged children. Solna: The public Health Agency of Sweden. The Public Health Agency of Sweden (2018a). Allt fler unga uppger stress. Retrieved 2019-01-29 from https://www.folkhalsomyndigheten.se/nyheter-och-press/nyhetsarkiv/2018/december/allt-fler-unga-uppger-stress/ The Public Health Agency of Sweden (2018b). Skolprestationer, skolstress och psykisk ohälsa bland tonåringar. Retrieved from https://www.folkhalsomyndigheten.se/publicerat-material/publikationsarkiv/s/skolprestationer-skolstress-och-psykisk-ohalsa-bland-tonaringar/