Ronald L. Breiger raised an intriguing set of issues for mobility analysts by suggesting that attention must be paid not only to the pattern of mobility between social classes but also to the manner in which occupations are to be mapped into social classes. This article examines two related models-collapsibility and internal homogeneity-for analyzing occupational mobility tables according to this scheme, by reformulating them as restricted latent class models in which the unobserved class variables are made explicit. The parametric representation of both models provides additional insight into how they differ. In particular, it clarifies the different ways in which they operationalize the criterion of "common mobility chances" for aggregating occupations into classes. The latent class model also permits the formulation of a new model for relationally defined social classes, a model that recognizes that the correspondence between class boundaries and occupational boundaries may be imperfect and that "marginal" occupations may be mapped into two or more classes. A model of this sort provides a good fit to the classic 8 x 8 British mobility table; this model appears to the both conceptually and statistically simpler than the internal homogeneity model.
No takes yet. Share an insight, caveat, or question.
Peter V. Marsden (1985) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: