English

This thesis comprises three independent but closely related empirical studies dealing with different aspects of health inequalities at different stages of the life cycle. First, the thesis focuses on children aged 9-12 and explores whether maternal employment is related to the probability of being overweight among these children. Using an instrumental variable strategy, the results indicate that children of full-time employed mothers are more likely to be overweight. This can likely be explained by unhealthy dietary and activity habits among children of full-time employed mothers. Second, this thesis focuses on adolescents by investigating educational disparities in smoking and whether formal education causally affects smoking decisions. Drawing on smoking biographies, the results suggest that educational differences are already apparent at smoking initiation. Whether an individual ever smokes is predominantly determined at an age before compulsory schooling is completed. The results are incompatible with the widespread finding that formal education has a strong causal effect on smoking behavior. Rather, it is more likely that (unobserved) factors determining both the selection into smoking and education as well as resulting peer effects, are responsible for educational differences in smoking. Further analyses lend support for this selection hypothesis. Third, this thesis considers working individuals and examines whether occupational demands mediate the relationship between education and health (behavior). The results indicate that occupational demands partially mediate the relationship between education on the one hand and health status, BMI and smoking behavior on the other hand. Education coefficients on BMI and smoking status significantly decrease up to 21 % and 27 % when the different occupational demands are included. Especially social demands seem to play a major role in mediating the relationship between education and health behavior. Existing inequalities in working conditions appear to matter for educational differences in health.