New data on identical female and male twins are used to estimate the impact of increasing parental schooling on child schooling that incorporates the existence of unmeasured heritable traits and marital sorting. These data yield cross sectional estimates that are consistent with previous studies of the impact of parental schooling on child schooling attainment. However, when twinning is exploited to estimate intergenerational schooling effects, the results are strikingly different. Controlling for women's earnings and childrearing endowments and husband's endowments and schooling leads to a marginally negative rather than a significantly positive coefficient for mother's schooling in the determination of child schooling.