highly significantp<0.0001
We repeated this 2500 times to build a null distribution of Δβ, and calculated the proportion with equal or stronger reductions for a one-tailed p-value, which was highly significant (p<0.0001).
We repeated this 2500 times to build a null distribution of Δβ, and calculated the proportion with equal or stronger reductions for a one-tailed p-value, which was highly significant (p<0.0001).
Mediated moderation was measured by the change in absolute incentive*THP effect when THP*CNV was included in the mixed-effects model (β=0.0214, t(1,16,472)=3.7298, p=0.0002; change in beta-coefficient: Δβ=–0.0008), and permutation testing (permuting the CNV as above) found a weakly significant effect (p=0.0132).