BACKGROUND In historical perspective, the transition from pre-industrial to modern societies is associated with increasing social status heterogamy. As individual’s acquired characteristics became more important for partner selection than inherited class status, the importance of status homogamy declined and marrying outside one’s own social group became more frequent.

OBJECTIVE We investigate educational heterogamy in a university town at the eastern border of the Hajnal line at the end of the 19th century. We ask whether marriage of unequally educated partners is related to dissimilarity in other characteristics of the partners. Ethnic background, origin (place of birth) and age difference between the spouses are considered as characteristics that may associate with sorting into educationally heterogamous unions.

METHODS The analysis uses data from the 1897 census in Tartu. Using logistic regression models, we estimate how age difference, origin heterogamy, and ethnic heterogamy of the spouses is associated with the probability of educational heterogamy.

RESULTS The results indicate a positive relationship between educational heterogamy and marrying outside own ethnic or origin group, but no effect for spousal age difference.

CONCLUSIONS Our study provides new evidence about marriage markets during modernisation, more specifically about the role of education. We show that educationally heterogamous unions in Tartu were also more often heterogamous in terms of partners’ background characteristics. This suggests that education may have motivated intermarriage by ethnicity and origin.

CONTRIBUTION Previous literature on this period has focused on social homogamy based on occupational information, while research on educational assortative mating mostly exists for the second half of the 20th century and later. We contribute by studying the importance of education in marital selection in the early phase of the educational expansion and economic modernisation.