Thus, there were three risk behavior factors that could be used for the mediation analyses, all extracted from the UKB. There was only one measure of self‐control included in the UKB, namely, risk‐taking proneness (yes/no). Results of the analysis using this measure as a mediator between CADM2 and the three risk‐taking behavior factors are presented in Figure 1 (with p‐values rather than regression weights as MAGMA does not provide such estimates). Path a for the association between CADM2 and risk‐taking proneness controlling for sex, age, and PCs was tested earlier and found to be significant (see Table 2). Paths c1–c3 for the associations between CADM2 and the outcomes (risk behavior factors) were reported in Table 3. Paths b1–b3 between risk‐taking proneness and the risk behavior factors were all significant (tobacco [ab]use factor OR = 1.27, p < .001; lifetime smoking factor, OR = 1.27, p < .001; and alcohol abuse factor OR = 1.21, p < .001). In step c′, the associations between CADM2 and lifetime smoking and risky alcohol use factors were attenuated when including the mediator (p =
