In one of the empirical analyses presented in Sect. 11.3 , it has been taken into account that Americans sort into couples both along the educational and racial traits. In that section, we presented some of the results by Naszodi and Mendonca (A new method for identifying what Cupid’s invisible hand is doing. Is it spreading color blindness while turning us more picky about spousal education? (2023b). https://arxiv.org/abs/2103.06991 ) in the context of a sensitivity analysis. In particular, we used their results to address the question whether the stylized U-shaped pattern in the degree of sorting along the educational trait is robust to controlling for sorting along race and the intergenerational changes in the racial composition of the marriageable individuals in the USA. To recall, Sect. 11.3 showed that both the GNM-method—designed to account for sorting along education and race jointly,—and the NM-method—designed to analyze sorting along the single trait of education—document the historical trend in educational homophily to have had a U-shape in the US over the twentieth century (see Figs. 11.8 b and 11.9 ).