Following Mare’s (2010) presidential address to the Population Association of America, a great deal of research has been conducted on the dynamics of intergenerational mobility and socioeconomic inequality. Due to this relatively short time span, however, comparatively little is known about the when, where, how, and why of intergenerational effects beyond parent-child ties. In this article, we examine educational attainment across three generations using the Panel Study of Income Dynamics. We find evidence that grandparental education is associated with grandchild education net of parental education, but that association is greatly reduced once spousal education is accounted for – suggesting spousal mediation as a key mechanism of the intergenerational transmission of educational advantage. We also find tentative evidence in favor of the augmentation hypothesis that grandparent and parent educational attainment interactively influence grandchild education. However, the augmentation hypothesis is not supported in spousal mediation models nor in cousin fixed effects models, suggesting that these results may be spurious. Contrary to prior work outside the U.S., we find no evidence of stronger grandparent effects when grandparents were still alive when the grandchild reached age 6. Supporting the claim that genetic transmission may partially explain these relationships, we find weak evidence that grandparent effects are stronger for biological grandparents than non-biological ones. Overall, this research suggests that grandparent’s educational attainment matters, but is mostly mediated through child’s spouse’s educational attainment. We suggest that three-generational transmission of educational advantage is a complex topic in need of further, careful study.