The comparative study of inequality of educational opportunity (IEO) lies at the heart of sociological stratification research. Within this tradition, scholars have long sought to separate empirically the allocation of from the dispersion in formal schooling. Robert Mare’s (1980, 1981) pioneering work on the educational transition model resolved this issue and advanced the state of practice in two regards. First, Mare presented the log odds ratio as a “pure” measure of allocation net of dispersion, making it possible to compare IEO across cohorts and countries. Second, he modeled educational attainment as the outcome of a sequence of decisions at transitions in the educational system, enabling scholars to study change in the impact of socioeconomic background over the course of the educational career.

However, despite its widespread popularity in comparative stratification research (Breen et al. 2009; Shavit and Blossfeld 1993; Shavit et al. 2007), the educational transition model has recently fallen into disfavor. Apart from issues caused by increasing selection on unmeasured characteristics in the transition model (Holm and Jæger 2011; Mare 1981, 1993), the magnitudes of the coefficients (log odds ratios) meant to capture allocation net of dispersion are not identified (Breen et al. Forthcoming; Cameron and Heckman 1998; Mare 2006). Consequently, the educational transition model is generally uninformative about differences in IEO between cohorts or countries and over transitions in the educational career.

In this paper, we devise a new approach that resolves the issues pertaining to the educational transition model and comparative research on IEO more generally. Our approach measures pure allocation net of dispersion in that we analyze the degree to which educational rank—i.e., one’s position in the educational distribution relative to all others in that distribution—is passed on over generations. Using data from the U.S. General Social Survey, we estimate the educational ranks of parents and offspring using an algorithm that allocates ranks conditional on information about the educational attainment of individuals in their respective generations. Given the resultant information on educational ranks, we employ methods developed in research on intergenerational income mobility (see, e.g., Mazumder 2011; Solon 1992) to study (1) trends in the educational allocation mechanism, (2) differences in the allocation mechanism according to groups such as race or gender, and (3) whether the allocation mechanism is stronger at the top or bottom of the educational rank distribution. Taken together, our approach resolves issues inherent in the educational transition framework by providing new tools for studying change in the mechanism of educational transmission, including identifying where in the educational distribution and for whom potential change is brought about.