concerns that population stratification may be driving the significant G × E effects observed in CDP. In FinnTwin12, principal components analyses of the population structure performed in Eigenstrat indicated a single dimension of ancestry. Second, although we are able to bring two datasets together to examine the role of life events and genotypic differences in intergenerational continuity in externalizing problems, there is imperfect correspondence in the measures, reporters, and study designs. For example, in the CDP sample, mothers reported on life events, and in the FinnTwin12 sample adolescents reported on life events. In the context of these differences, the parallel but non-significant trends observed in our replication sample are encouraging.
