This study examined the influence of educational achievement and occupational position on changes in risk behavior. Study population were 3753 men and women aged 25-64 years who were sampled by the first MONICA Augsburg Survey (Monitoring trends and determinants in cardiovascular disease). The subjects were sampled in 1984-85, were followed up for three years, and were reexamined in 1987-88. The baseline findings showed for both men and women a statistically significant inverse association between current cigarette smoking and educational level. During the follow-up period the differences between highest and lowest educational levels increased significantly among men. In 1987-88 only 21% of the best educated men were smokers compared to 38% of those with the lowest educational level. Men with low educational levels also drank more alcohol than better educated men, whereas among women those with lowest educational level drank less alcohol than the better educated. Statistically independent of education and age it was found that male civil servants and farmers had the lowest proportion of smokers in cross-sectional as well as longitudinal analyses. Among women, smoking was and remained most prevalent in simple white-collar occupations. In general, the findings indicate that the type of occupation and the actual working conditions have effects on life-style related risk factors which are in part independent of a social gradient. The results also suggest that the tendency to change unhealthy behavior is less pronounced in 'high risk' groups.