Do socio-economic community factors affect mortality in Norway? An analysis based on unique multilevel longitudinal data.

The intention is to find out whether community characteristics exert an effect on an adult person's cause-specific mortality, above and beyond that of the person's individual characteristics. The focus will be on socio-economic community factors, such as average income and education and unemployment rate. Such multilevel mortality models have never been estimated for Norway before. Municipalities will be the unit of aggregation in most of the project. The Norwegian register data, which are longitudinal bo th at the individual and community level, provide a rare opportunity to used advanced statistical methods to account for individual and community-level unobserved factors and selection via migration into municipalities. They also allow the researchers to assess the importance of earlier versus current exposure to good or bad neighbourhoods. Besides, because of the size and scope of the data, a detailed exploration of small groups and complex interactions is possible. Results from this project, largely in the form of journal articles, should therefore be an important contribution to the scholarly literature. In addition, they may have significant policy implications. Both the Norwegian principal investigator and the American collaborator have long experien ce with similar analyses within demography and epidemiology. Four sub-projects are specified.