Social class assignment and mortality in Sweden.

The earlier practice of assigning all members of a family to the same social class as that of the household head, typically the father, has in recent years been replaced by either basing individual class position on one's own occupation or of one of the family members, not necessarily the father. These various practices have been extensively scrutinised for more than 20 years. The validity of the approaches has chiefly been tested by checking how well they account for the variation in some criteria, mostly class identification, political attitudes and voting behaviour. Here it is shown, using census data from Sweden, that mortality-rate differences between social classes covering the period 1991-1997 are greater for both men and women when both spouses are assigned to the same social class on the basis of the dominance approach, where the labour market position of either spouse may determine the social class of the family. It is suggested that the common observation that class differences are smaller among women than among men may, at least to some extent, be the result of establishing a woman's class position on the basis of her own occupation rather than the labour market position of her spouse.