In an article published in the American Economic Review, Dorothe Bonjour et al. (2003) used a dataset on female monozygotic twins and showed that the within-twin estimated return to one year of education was 7.7 percent and statistically significant at the 5 percent level. This comment illustrates that the point estimate conclusion is driven by one twin pair, which is an outlier in the dataset. If one eliminates this twin pair, then the estimated return to education is 5.1 percent and statistically significant at the 10 percent level only. (JEL I21, J16, J24, J31)