highly significantP < 0.0001
Whereas the original data analysis reported a highly significant effect ( P < 0.0001), generalized mixed-effects models with a random effect for name indicates that the effect is not robust, neither with the data from the general public (b = −0.108 [log odd units], z = −1.712, P = 0.087) nor with the data from elected officials (b = −0.065, z = −1.432, P = 0.155). * While this does not mean that we should accept the null hypothesis, it shows that the study is underpowered.