INTRODUCTION
The analysis of social mobility is probably the field of sociology that has been most affected by methodological and theoretical developments in the period since the 1950s (see, e.g., Sørensen, 1986). In classical mobility studies Rogoff (1953), Glass (1954), Svalastoga (1959), and Carlsson (1958) studied social inequality on the basis of mobility tables. These tables were obtained from cross-sectional samples of men who were asked about their current position and about their father's position when they grew up. However, as respondents had different ages at the time of the interview and therefore had different historical experiences, these efforts did not provide very interpretable results (Sørensen, 1986).
Mobility tables were sometimes disaggregated by birth cohorts (Featherman & Hauser, 1978; Goldthorpe, 1980), but this does not alleviate the problem because careers take place in the labor force. This means the relevant period is the amount of time spent in the labor force and not the age of the people (Blossfeld, 1986). Members of the same birth cohort are at different career stages because they spent different times in the educational system. Thus, positions reflected by mobility tables of birth cohorts are observations of locations of people at different stages in their career (Sørensen, 1986).
The typical mobility table not only ignores differences in the historical location of cohorts and the amount of labor force experience of people, but also neglects the fact that job mobility is strongly influenced by education and other background variables of the individual.