Key points are not available for this paper at this time.
A general "logistic-multiplicative" model is developed which incorporates variables besides origin category to predict destination category in occupational mobility tables. The additional variables can be continuous or categorical. The (partial) bivariate relationship between origin and destination can be modeled with any of the existing multiplicative mobility models. In an empirical example, two main results emerge: interactions off the main diagonal become insignificant with the introduction of education and race variables; main diagonal effects for one origin category become insignificant while those for other categories do not, suggesting that different mechanisms of status transmission are at work in different strata.
John Allen Logan (Thu,) studied this question.