Back to search

FINNUT-Forskning og innovasjon i utdanningssektoren

Understanding the causes and consequences of dropping out of school: The need for a life-course approach

Alternative title: Årsaker til, og konsekvenser av å droppe ut av skolen: et livsløpsperspektiv

Awarded: NOK 6.0 mill.

Project Number:


Application Type:

Project Period:

2015 - 2018


Partner countries:

We have studied risk factors and protective factors for dropout from upper secondary school (high school), as well as consequences of dropout for work participation and health, based on register data for all persons born in Norway 1967-1976 (N=626 928). Only 2.1% were not in work during 1993-2011 (age 17-44 years). People with education and health problems prior to 1993 accounted for 2.2%, but 36% of those without labour participation. Problem accumulation had a stronger negative impact than the sum of the individual problems. Men who had not completed high school at 21 years, had lower employment rate and labour earnings and higher disability rate at 34 years. Birth order, birth weight and chronic childhood disease were of major importance for employment, while cognitive skills and parents' education and income played a greater role in relation to income and disability. Using multi-state models, which show the dynamic nature of outcomes over time, we studied the effects of completing high school in a follow-up from 23 to 35 years. High school completion led to higher work participation, more education and less health-related absence (sickness absence, disability pension), for students in both general and vocational education. Vocational students came into work faster, but had more unemployment, sickness absence and disability. Differences between completers and non-completers were reduced during follow-up. Previous studies have often focused on single outcomes and applied cross-sectional design or short follow-up periods and may have overestimated the negative effects of high school dropout on subsequent work participation. People with low education have more sick leave than others. Differences in working conditions could be one reason. Among 3328 persons in the cohort who participated in the HUNT3 Study, part of the social gradient was due to heavy physical workload. Thus, interventions to reduce heavy physical work could reduce social differences in sick leave. The personality trait neuroticism and parents' social position influenced these relationships, more for women than for men. Moreover, approx. 14% of people who had not completed high school, had sick leave with musculoskeletal diagnoses in 2009 vs. 2% of those with at least 4 years of college or university. The difference was smaller for mental diagnoses. Psychosocial work factors (job demands, control and bullying), as measured with the HUNT3 questionnaire, could only explain a small part of these differences. Education is important for later work and working conditions. We have data on occupation, but not on working conditions for the whole cohort. Based on Statistics Norway's Living Conditions Surveys, we have developed Job Exposure Matrices (JEMs) for mechanical and psychosocial exposures, which are used to assign similar exposures to everyone in similar occupations. We use the JEMs to study the impact of these exposures on differences in musculoskeletal disorders between high school completers and non-completers. Self-reported education and sickness absence are used in research, but we do not know how valid these data are. We have compared self-reported data for nearly 50,000 women in the Norwegian Mother and Child Cohort Study (MoBa) with data in national registers. Self-reported educational level was slightly higher than notified in the in the Norwegian Education Database (NUDB), but nevertheless corresponded quite well. However, there was clearly higher sickness absence in the register data (48%) than self-reported in MoBa (31%), with greater difference among lower educated pregnant women. This may be because the questions were comprehensive and detailed, so some women skipped them or answered them incompletely, especially women with lower education. High school dropout can have unfavourable consequences for health and socioeconomic attainment. Among nearly 400 000 people in the cohort, drug-related mortality (454 cases) was clearly associated with dropout during 10 years of follow-up, with a five times higher risk among non-completers than completers. Unfavourable careers relating to work and other socioeconomic characteristics following high school dropout could be an important mechanism for this association. Within a school class, some students are almost a full year younger than others; in addition, some may have early or delayed school entry. Young relative age within grade is associated with adverse childhood outcomes, such as ADHD and learning problems. However, research into long-term effects is lacking. We found that those old for their grade had higher rates of high school dropout and lower IQ at age 18 (men), but were less likely to have diagnosed ADHD and alcohol/drug problems in adulthood. Boys and children from lower social background had more often delayed school entry and fare worse on all outcomes, including higher dropout, lower IQ, more ADHD diagnoses, alcohol/drug problems, suicide and drug death.

Få studier har hatt så langvarig oppfølging som vår studie. Vi fant at forskjeller mellom dem som fullførte og dem som ikke fullførte videregående skole, ble redusert over tid. Studier med et kortere tidsperspektiv har ikke kunnet dokumentere dette. Vi har benyttet nye statistiske metoder som viser komplekse sammenhenger og hvordan flere utfall samtidig fordeler seg og utvikler seg over tid. Dette bidrar til et mer fullstendig bilde av virkeligheten enn når man studerer ulike utfall hver for seg. Få studier har hatt et livsløpsperspektiv, med data fra bakgrunn og tidlige forhold i livet. Vi har kunnet studere betydningen av disse forholdene og ta hensyn til dem i analysene, slik at muligheten for at de konfunderer og gir feilaktige resultater reduseres. Vi får kunnskap om bakenforliggende årsaker, seleksjonsprosesser over generasjonene og mekanismer, og derved kunnskap om hvor forebyggende tiltak bør settes inn for å ha effekt.

Our main objective is to gain a causal understanding of the processes that lead to dropout of upper secondary school, and the subsequent effect it has on lifetime work trajectories. The development of risk across time and context are seen as crucial in un derstanding how risk develops. While individual factors, such as cognitive ability and health, are thought to causally influence risk of dropping out, we hypothesize that part of the association between individual factors and dropping out can be explained by childhood family background, which may increase the risk of dropping out of upper secondary school directly. Further, we hypothesize that the association between dropping out and later work outcomes is to a large extent explained by both individual fa ctors and childhood family background. We also wish to explore to what extent conditions of the workplace obtained impacts leaving the workforce; how family formation influences both transitions into, or out of, the workforce; and to what extent social an d economic context influences may weaken or strengthen the pathways between dropping out on later work outcomes. Our study is based on register data that allows us to follow all individuals in Norway from birth (between 1967 and 1976) up until 2010. We ha ve obtained data on a wide array of explanatory variables from register and health surveys, collected over repeated time-points across the lifecourse. We shall use new analytical methods based on causal inference. Our findings will inform us as to which m echanisms are leading to decreased work participation and dropping out. This will have implications for where interventions will have the greatest impact. If our findings suggest that most of the effect of dropping out on later outcomes is due to early fa mily conditions, then this would imply that interventions ought to be targeted at family and income policy, as opposed to placing most focus on school policies and improvements.

Funding scheme:

FINNUT-Forskning og innovasjon i utdanningssektoren