Background: From a historical perspective, the transition from a pre-industrial to a modern society is associated with increasing social status heterogamy. As individuals’ 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 end of the 19th century. We ask whether marriage of unequally educated partners is related to dissimilarity in the partners’ other characteristics. 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 modelling, we estimate how age difference, origin heterogamy, and ethnic heterogamy of the spouses associate with educational heterogamy.

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

Conclusions: Our study provides new evidence about marriage markets during modernisation, specifically regarding the role of education. We show that educationally heterogamous unions in Tartu were often also more heterogamous in terms of partners’ background characteristics. This suggests that the partner’s 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 educational expansion and economic modernisation.