Plug, E. (2004). Estimating the effect of mother's schooling on children's schooling using a sample of adoptees American Economic Review, 94(1):358--368.

This paper examines the impact of parentalschooling on the child¿s schooling and usesadoptees to get rid of persistency effects causedby the parents¿ genes. The results indicate that,especially for mothers, inherited abilities andassortative mating play an important role in theintergenerational transmission of schooling. Infact, for adoptees I found no treatment effect forthe mother¿s schooling, conditional on her husband¿sschooling.It should be noted, however, that the WLSdata on adoptees and their parents do not possess the properties of a clean and well-de¿ nedexperiment, and that obtained results require acareful interpretation. There are two potentialdangers to an adoption experiment. First, adopteesand adoptive parents are different fromother children and their parents. This argumentsuggests that my maternal schooling estimatesmay be biased and suffer from omitted variables,but I have little indication of what thesemight be. The sensitivity analysis ruled out anumber of plausible candidates. Second, adopteesare not always randomly assigned to theiradoptive parents. This argument suggests that aportion of what is interpreted as the impact ofthe parent¿s schooling may in fact be genetic.With respect to paternal schooling estimatesthere is some merit to this view. However, withrespect to the estimated maternal effect it is not.Nonrandom assignment and corresponding upwardbias form no danger when interpreting theabsence of maternal schooling effects.In all, these results, in combination with theparallel finndings of Behrman and Rosenzweig(2002) using twins, support the idea that thepositive influence of mother¿s schooling on thatof her child disappears when heritable abilitiesand assortative mating are taken into account.