Skola, lärande och psykisk hälsa: påverkansfaktorer och prevention av skolmisslyckande - SLANT

Project summary

The life chances of children and young people are to a large extent determined by socio-economic characteristics of the family and society and by the knowledge, skills and personal characteristics developed during upbringing and schooling. While a majority successfully develop the competences required to meet the demands of adult life, some fail to do so, for example because of a lack of resources in the family, inadequate schooling, and mental or physical illness. In order to better understand the factors which influence children’s life chances, the proposed project has three main aims, namely to develop knowledge about determinants of outcomes of schooling; consequences of school failure and mental health problems; and how school failure may be prevented and compensated for.

The project is interdisciplinary, with participants specialized in different areas within education, public health, psychology, sociology and social work. The project work is divided into three strands and six sub-projects. The project is funded through a donation by Stenastiftelsen. The research project belongs to the research environment FUR: förutsättningar, utbildning och results/ Pre-requisites, Education and Results. 

Data

The empirical research is based on three longitudinal databases available at the Department of Education and Special Education, University of Gothenburg:

(1) The Gothenburg Educational Longitudinal Database (GOLD) which is a register-based population database comprising all students born 1972 and later, and which includes information from educational and other registers.

(2) The UGU (“Utvärdering Genom Uppföljning”) database which comprises 10 representative samples, each with around 10 000 participants born between 1948 and 2004, and which in most cases have questionnaire information and test data from grade 6 and also from later follow-up occasions. Since 2018 the UGU database is a national infrastructure with funding for development and maintenance from the Swedish Research Council.

(3) The RIDDLE (Reading Acquisition: Individual Differences, Development, and Enhancement) project which is a longitudinal prevention study comprising 364 children who were four years old at project start, and who currently are in grade 8.

Strand 1: Determinants of outcomes of schooling

The first strand of research includes two subprojects which investigate determinants of outcomes of schooling:

Subproject 1.1: School outcomes in a long-term perspective

The main aim of this subproject is to assess the relative importance of determinants of school outcomes across different groups of birth cohorts. The cohorts have followed different combinations of curricula for compulsory school (U-55/Lgr 62/Lgr 69/Lgr 80/Lpo 94/Lgr 11) and for upper secondary school (Läroverksstadga 1955/Lgy 70/Lpf 94/Gy 11) which differ with respect to guidelines for instruction and organizational aspects of schooling, such as systems of grading, differentiation of students, and organization of special needs education. The main question, which can be investigated for all UGU birth cohorts between 1948 and 2004, is the relative importance of cognitive ability, social background, cultural background and gender as determinants of school failure and general school achievement, and how this varies as a function of school characteristics and societal factors. The data allow a large number of cohort comparisons, focusing on curricular and organizational aspects, and on societal changes such as increasing economic inequity and school segregation. Both comprehensive school and upper secondary school will be investigated.

Subproject 1.2: Individual and societal factors behind school failure

The main aim of this subproject is to investigate different phases and factors in the development of school failure using a longitudinal approach for multiple cohorts. For UGU birth cohorts 1982, 1992, 1998 and 2004 information about birth month and socioeconomic background will be used to explain individual variation in cognitive abilities in grade 6. This model will be extended with results on national tests and will be used to predict achievement and school failure in grade 9.  For cohorts 1972 and 1977 a similar model will be built which also takes advantage of the fact that information about cognitive abilities is available in grade 3. Next, explanatory variables derived from the student questionnaires (e.g., self-concept/self-efficacy, achievement goal preferences, motivation, coping, self-reported mental health) and registers (e.g., school relocations, participation in special needs education, mental health problems) will be added to the model to investigate to what extent they affect the risk for school failure.

For upper secondary school a similar approach will be taken, starting with a basic model including grade point average (GPA) from compulsory school, along with cognitive abilities from grade 6 and background variables, predicting school failure in the form of dropout or low grades. The differentiation of students into different programs will be dealt with through a dummy variable approach and/or through fitting separate models for different programs or groups of programs. As for the compulsory school model, explanatory variables will be added in the next step, using the same sources of information. In combination, the results from the models for compulsory and upper secondary school will provide a good coverage of individual and social determinants of school failure.

Strand 2: Consequences of school failure and mental health problems

The second strand of research includes three subprojects which investigate consequences of school failure and mental health problems.

Subproject 2.1: Mental health consequences of school failure

The main aim of this project is to investigate the extent to which school failure has consequences for mental health. One part of this subproject focuses on mental health outcomes in compulsory school and upper secondary school using self-reported information collected in UGU (grades 9 and 12 for the 1998 cohort; grades 6 and 9 the 2004 cohort). Given the longitudinal nature of these data, it is possible to investigate changes in self-reported mental health, and relate these to changes in self-perceived ability to cope with academic demands, as well as to actual levels of school achievement as indicated by number of failed grades and drop out from upper secondary school. These data provide a relatively strong basis for making credible causal inferences concerning the impact of school failure on mental health. Because this approach may be vulnerable to bias in self-reports and to bias due to non-response, recent advances in survey methodology will be applied in order to minimize these risks.

A second part of this subproject takes advantage of the population data available in the GOLD database in combination with the information on mental health problems available in the registers CDR, NPR and PDR managed by Socialstyrelsen. The NPR includes all in-patient (hospital) care in Sweden from 1987. Starting in 2001 also outpatient doctor visits (day surgery and psychiatric care) from both private and public caregivers are registered. Visits in primary care are not included in the NPR.

The PDR was the latest register to be established and it includes complete information from 2005, also including prescriptions from primary care. This implies that the 1990 birth cohort can be followed from age 15 to age 30 and that the 2000 birth cohort can be followed to age 20. We will therefore focus on the birth cohorts 1990-2000. The GOLD database does not include any self-reported information, except for those persons who also participate in UGU, but it provides rich information about social background, educational achievement and choices, and labor market outcomes. These data are therefore well suited to investigate consequences of school failure on mental health.

Subproject 2.2: Relations between mental health problems in early life and mental health, education and social conditions in young adulthood

The main aim is to study the relations between mental health problems in early life and mental health, education and social conditions up to age 24, including possible changes in socio-economic inequalities in mental health. A related purpose is to examine the validity of self-reported data for invariant comparisons of mental health among adolescents and for prediction of mental health later in life. The research will be based on data from the two most recent UGU cohorts, which both included a measure of mental health (The PsychoSomatic Problems scale, PSP, Hagquist, 2008): for UGU 1998 in grade 9 and grade 12, and for UGU 2004 in grade 6 and grade 9. Data from the registers CDR, NPR and PDR will also be included, allowing for analyses of the predictive power of self-reported mental health.

These data will be used to investigate how self-reported mental health problems in grades 6, 9 and 12 are associated with mental health in succeeding years, educational outcomes, and child-parent communication and school experience; and to what extent these relations are moderated by gender and parental education. It will also be investigated if prescription of psychotropic drugs during adolescence affects the risk of mental health problems (e.g. depression and anxiety) and are associated with negative social consequences later in life.

The studies of validity of self-reported mental health data will investigate measurement invariance across cohorts, genders, and data collection formats (paper and pencil or web); concurrent validity through comparisons with register data; predictive validity in relation to mental health problems later in life; and predictor concordance in term of similarities and dissimilarities between short term and long term determinants of mental health.

Subproject 2.3: Consequences of school failure for labor market outcomes, and compensatory effects of adult education

One aim of this subproject is to investigate consequences of school failure on labor market outcomes in the form of employment and wages using the GOLD population data. The analyses will include persons born between 1972 and 1990, which makes it possible to follow them up to at least 30 years of age. It also makes it possible to investigate consequences of the changes in upper secondary school due to the Lpf 94 curriculum, such as making all programs three years long and giving access to tertiary education. The analyses will have a special focus on comparisons between immigrants and non-immigrants.

Another aim of the subproject is to investigate to what extent and for which groups of person municipal adult education does have a compensatory function through providing the skills and certifications needed for full establishment on the labor market. A longitudinal approach will be used to investigate educational careers of immigrants and non-immigrants after compulsory school, focusing on dropout from upper secondary school, and reentry into education via the adult education system.

Strand 3: Prevention of school failure

The third strand of research includes one subproject which investigates prevention of school failure.

Subproject 3.1: Prevention of reading difficulties through a pre-school training program

The main aim of this project is to extend data collection and analyses of data from the ongoing longitudinal intervention project RIDDLE. Analyses of main and interactive effects of the phonological training of four to six-year-old pre-school children have demonstrated that there was a general positive effect of the training, and that in particular children with lower levels of phonological and cognitive skills benefitted from the program. It thus had a compensatory effect on phonological skills which was also seen in early reading skills in grade 1. Follow-up data from later grades have been collected but it has not yet been possible to analyze these data and report the findings. Further follow-up data will be collected in grade 8 in 2020/2021 and also two years later in 2022/2023. The data from age four and onwards will be approached with techniques for longitudinal data analysis, investigating general and differential effects of the prevention program on development of literacy skills.

Publications

TBA