ABSTRACT
There has been a revolution in the study of social mobility: the once dominant Blau-Duncan paradigm has been overthrown by log-linear modelling. The techniques were first developed by Goodman (1978) in statistics and applied to mobility problems by Goodman, Hauser, Featherman, Duncan, Goldthorpe and many others. The log-linear revolution was a noble experiment, and at first seemed to offer a bright new future beyond Blau and Duncan. Goldthorpe and his colleagues elaborated a particular version of the paradigm, combining log-linear methods, cross-cultural data and a particular schema for measuring class position (Goldthorpe, 1971, 1980a, 1985a; Erikson, Goldthorpe and Portocarero, 1983; Erikson and Goldthorpe, 1985). They have done impressive work with it; that is freely conceded even in this critical appraisal. Their Herculean labours in data preparation are ambitious in scope and commendable in quality. Their analytic methods are modern, employed with agreeable technical virtuosity and presented with commendable clarity. There is much good here.
