In the meta-analysis of the 27 twin correlations in Table 1, the base model for Neuroticism with 5 parameters (h m, h f, d m, d f, and one for allowing the opposite-sex twin correlation to differ from its expectation under the hypothesis of no qualitative sex differences) did not show a better fit than one where the opposite-sex twin correlation was equated to its expected value (total N = 29,496 pairs). The base model χ2 was 88.33, and the restricted model χ2 was 88.89, a non-significant change with 1 degree of freedom. Next, this restricted model with qualitatively the same additive and non-additive genetic effects for males and females was compared with a model that specified that the proportions additive and non-additive genetic variance were equal across sexes. This model had a χ2 statistic of 91.63, a non-significant increase of the χ2 statistic by 2.74 for 2 degrees of freedom. Next, it was tested whether the non-additive genetic effects could be dropped from the model. The χ2 statistic increased to 170.39, which is highly significant. Thus, for Neuroticism, both
