Abstract

A new mobility ratio has been proposed as a way of dealing with the confounding between main and interaction terms which is inherent in the old mobility ratio. It is shown that the new ratio does not in fact succeed in this aim, that values of the ratio are highly dependent on the model fitted to account for interaction, and that criticisms of Blau and Duncan's conclusions about the mobility of American society, which derive from application of the new ratio, cannot be sustained. The old mobility ratio, unlike the new, always has a determinate value, and its alleged defects are formally equivalent to properties which are not usually regarded as defects in a multivariate analysis of discriminance.

This content is only available as a PDF.
You do not currently have access to this article.