As also mentioned in the introductory and review sections of these webpages, the CAMSIS approach investigates structures of social association between the incumbents of occupational units, in terms of typical national patterns in marital partners' occupational combinations. However, we also expect that the same structures of social association between occupations should be revealed by any other data which links occupational units in terms of typical contacts between occupational unit holders within the population.
Certainly, previous comparisons have already revealed that occupational units linked by friendship patterns form the same structure as those linked by marriage patterns (eg Prandy and Lambert 2003 ). Secondly, patterns of inter-generational occupational mobility can also be conceptualised as data on the relative frequencies of social associations between occupations, and analysed with similar techniques.Thus, Rytina (2000) has analysed father-son mobility in the UK in this way, and suggested that inter-generational occupational associations do indeed reveal a structure of stratification dominated by a low dimension of hierarchy. (Of course, such an analysis engages directly with a much wider literature concerning sociological conceptualisations of 'class' and stratification, as discussed in the paper).
Below, we briefly describe a preliminary investigation into structures of intra-generational mobility using the same techniques again, this time where the binary association modelled concerns the transitions from 'start' to 'end' jobs within individual careers. Intra-generational mobility records have the attraction that they are widespread in retrospective and other longitudinal data resources, and are fairly unambiguous in their representation of socially connected occupations. On the other hand, they introduce a number of complex issues in the analytic treatment of longitudinal data. The results of an engagement with these issues is likely to be of considerable substantive interest, which we hope to investigate under a longer project. The preliminary investigations described below, however, only cover the most basic treatments of intra-generational data.
A presentation to the ESRC Research Seminar Series "Occupational Information: Occupational Careers", held at the LSE on the 10th May 2002, concerned the initial patterns of results found from an early investigation into the structures of association between occupations as revealed by intra-generational occupational transitions. This archive contains the slides used for that presentation, as well as an Excel file listing an example of the derived occupational unit scores which the paper discussed. (13.5.02: We are hoping to extend this research in the near future; an interim - and incomplete - paper reporting current findings is available on request to the authors, whilst this webpage should be updated as further results progress).
In summary, we found evidence to suggest that patterns of occupational associations in within-career job-transition records from two retrospective UK surveys (the BHPS and FWLS), appeared to form a low dimensional hierarchy that was closely correlated with that found by other CAMSIS and related investigations. There was no evidence of greater categorical clustering in the intra-generational orders than those derived from marital association. On the other hand, a slight difference between the marital association and intra-generational models was perceived to lie in the lesser influence of educational qualifications, and greater influence of careerist trajectories, in structuring the intra-generational scores. Less satisfactorily however, there was additionally some suggestion that the dimension scores may be conflated with time period structures, since newer occupational units tended to be scored towards one end of the scales, and older or declining occupational units towards the other.
At present however it seems likely that further work on the scale constructions will be needed to adequately confront the complexities of the intra-generational data. Our current results, based on male career transitions only, do not include duration weights which might be used to account for the differential contribution of the number of career transitions reported by different individuals. Also, it may be of considerable interest to deal with accounts of transitions into and out of employment titles, such as transitions to unemployment and labour market inactivity, whereas the current account only deals with job-to-job transitions.
It is intended that we will update these pages as and when further research is conducted.
