There is mixed empirical evidence for the modernization theoretical prediction that inequality of educational opportunity is reduced through societal developments generated by industrialization. This paper conducts a multiverse analysis with 768 distinct multilevel regression models in order to comprehensively test this prediction and to investigate how sensitive inference about the cross-level interaction effect is to variation in analytic choice. The alternative analysis scenarios or “universes” differ in the criteria used for the exclusion of observations, the operationalization and adjustment of micro- and macro-level variables, and the choice of regression model formula. The models are fit to data from 26 European countries that cover from four to six ten-year birth cohorts spanning a period from the 1930s to the 1980s. In aggregate, the multiverse of results is consistent with an industrialization-related decrease in the effects of social origins on the expected duration of schooling and on the completion of upper secondary education. However, the evidence for declining social origin differentials in the odds of completing tertiary education is considerably weaker, suggesting that the equalization is concentrated at the lower levels of education. It was also found that for all three educational outcomes there are reasonable analytic choice combinations that lead to an inability to observe the predicted equalization.