Intragenerational mobility and mortality in Oslo: Social selection versus social causation

We investigate the relative importance of the selection and causation hypotheses of social inequalities in mortality, and estimate upper and lower bounds for the gender-specific mobility effects. For all inhabitants of Oslo aged 50–69 years in 1990, we knew their social class in 1960 and 1980 and whether they died between 1990 and 1994. Analysing these data with diagonal reference models, we found those moving upwards in the social hierarchy to have lower mortality rates than their class of origin but higher mortality rates than their class of destination. A corresponding pattern was found for those moving downwards. Thus, social mobility may increase or constrict the social class mortality divide. We estimated the upper bound to the mobility effect to be an increase of 52% for males and 28% for females (situation of no causation) and the lower bound to be a decrease of 24% for males and 21% for females (situation of no selection). Because both selection and causation effects are expected to play a role and to work in opposite directions, the resulting effect of social mobility on the mortality divide may be rather small.