Background: Education fundamentally shapes an individual's life chances and social status. Duration and level of education are associated with people's incomes, socio-economic status, living standard, lifestyle and the respect and esteem they enjoy. Failure to fulfil educational potential may have long-term consequences for later occupational and social life, as well as creating problems for society at large. Purpose: The aim of this prospective follow-up study was to investigate which health, lifestyle, family and school factors in adolescence had the strongest impact on educational level attained in adulthood. Sample: The original study population included all 9th-grade school pupils aged 16 years attending secondary school in spring 1983 in Tampere, Finland (n=2269). Subjects were followed up at 32 years of age in 1999 (n=1471). Design and methods: In 1983, 2194 pupils (96.7 percent) with a mean age of 15.9 years (SD 0.3 years) completed questionnaires during school hours. The questionnaire included questions about psychological and somatic health, lifestyle, family background, school factors and life events. In 1999, postal questionnaires were sent to the same study population when they were 32 years old. The follow-up questionnaire included basically the same questions about health and lifestyle as in adolescence and age-appropriate versions of questions about social relations and life events. Education was measured by the highest level of attained education. The follow-up cohort totalled 2091 persons (92 percent), and the response rate was 70.3 percent (n=1471). The univariate associations between independent variables in adolescence and attained educational level in adulthood were analysed using the X[superscript 2] test, and Student's t-test where appropriate. Multivariate logistic regression analyses using forward stepwise procedure were used to determine which factors in adolescence had the strongest impact on adult educational level. Results: Several health, lifestyle, family and school factors in adolescence were univariately associated with educational attainment in adulthood. Multivariate analyses showed that, in both genders, low school achievement and manual class of origin were the strongest predictors of low educational level in adulthood. Additional significant predictors among females were poor perceived health status, spending less leisure-time on hobbies and more on dating, and among males, poor relationships with teachers and heavy drinking. Conclusions: Mid-adolescence is an important period for future educational life trajectory. Based on our study results, more attention should be paid to adolescents with school difficulties, unhealthy lifestyle and poor perceived health. It is also important to promote educational equality in relation to family background.