Loglinear analysis is used to assess the degree of educational homogamy in 65 countries. Differences in educational homogamy among these countries are then explained in terms of level of economic development, degree of political democracy, the dominant religion, and the technological background of developing countries. An inverted U-curve relationship is found between level of economic development and educational homogamy. Furthermore, cultural characteristics are found to be important explanatory variables. Grouping countries into "families of nations" according to dominant religion and technological background helps explain the differences among countries. Catholic, Muslim, Confucian, and mixed Catholic/Protestant countries show significantly more educational homogamy than do Protestant countries, and industrializing societies with a horticultural background show significantly less educational homogamy than do industrializing societies with an agrarian background.