1. Sex and age standardisation of rates was performed with the direct method using the population ofNorway at the end of 1985.15 Indirect standardisation was used for rates of subjects participating in the interview survey, using the rates of respondents to the present postal survey. To evaluate the effect of several factors on a dichotomous outcome variable, the method of stepwise multiple logistic regression analysis was applied.'6 For outcome variables with more than two categories the method of backwards stepwise selection of log linear models was applied.17 An all significant effects model was defined as a model in which only terms were retained that had a significant effect on the outcome at the 0-01 level.18 The nominal significance level of 0-01 was chosen to adjust for multiple significance testing
2. Obstructive lurg disease in an urban population;Gulsvik, A.,1979
3. A study of disease in migrants and their siblings: Development of sibling rosters;Magnus, K.; Hougen, A.; Andersen, A.; Pedersen, E.;J Chronic Dis,1970
4. Population and housing census 1980. Vol 1. Housing statistics;of Statistics of Norway, Central Bureau