We repeated the same regression models using different contextual
characteristics such as unemployment, poverty level and a composite
measure of education poverty and unemployment obtained using Principal
Component Analysis (PCA). Table 2 reports the results of these
additional models. All models show the significant presence of a
negative interaction effect between PGS and contextual measures of
neighborhood deprivation indicating that genetic predisposition for high
education is mitigated by contextual characteristics.
TABLE 2 HERE