The status-caste exchange thesis has been a theoretical workhorse for the study of racial intermarriage in the United States since its introduction in the 1940s, and has enjoyed a revival in recent decades. Some recent studies, however, challenge this view. We test the thesis with multinomial logit regression using data on black-white marriages in the US and Canada. We find modest support for the theory in the US but not in Canada. In the US, white women married to African American men are somewhat more likely to marry up on education than white women in same-race marriages, but the same pattern is not observed when the intermarriage involves Caribbean blacks and whites. These statistically significant tendencies, however, reflect rather modest differences in the proportion of couples in interracial marriages with different educational levels compared to those found among same-race couples.
