California Center for Population Research

This study uses log-linear models and data from the National Longitudinal Survey of Youth (NLSY79) to compare the odds of educational homogamy in marital and cohabiting unions. Differences in the educational resemblance of married and cohabiting couples vary depending on the sample used. Cohabiting couples are less likely to be educationally homogamous than married couples using a sample of prevailing unions. Restricting the sample to newly formed unions, however, eliminates this difference. Furthermore, I find little support for the hypothesis that cohabiting couples who transition to marriage are more homogamous than cohabiting couples who separate, although these results vary by respondent’s sex. My results suggest that differences in educational homogamy by union type in prevailing unions are driven by the accumulation of the most homogamous marriages over time rather than differences in sorting into unions or selection out of cohabiting unions into marriage.