Abstract
Sociological researches on educational opportunity generally suggest that the class inequality in educational attainment has not been diminished for a long term. Although they carefully separate the changes in class inequality of education from those caused by the marginal expansion of educational categories, they fail to realize the relationship between class inequality and meritocracy. This paper proposed a cumulative logit model (it can be called “a threshold model”) to test the stability of educational inequality under the working of meritocratic selection, and applied it to the SSM survey data to examine the Japanese case. The model assumed a continuum distribution of educational advantage for each class, and also assumed several thresholds (cut-points) on the continuum such that the educational categories corresponded to the intervals defined by these threshold values. With this framework of analysis, we found that the educational change both for prewar and postwar periods resulted from lowering of the threshold values only. This means that the advantage distribution of classes have utterly unchanged and the meritocratic selection, not class biased selection, has applied to each class equally. This structure of educational inequality is the same for gender differentiation except for the threshold values. Further, the model forecasts a gradual decline in class inequality of education at the absolute level, but the sustained inequality at the relative level as before. The model's good fit to data suggests inadequacy of a single factor approach on this topic, because the advantage distribution is stable over a long period of time regardless of fluctuations in such a factor.
