Table 4.

The variable importance values for an L1-penalized logistic regression that regresses political ideology (conservative or liberal) on the full model of all available survey-based predictors and FC predictors as well as the estimated effects and standard errors for each important variable from a logistic regression model. The L1-penalized model was tuned using cross-validation in the glmnet package in R software. Only variables that had nonzero importance in the resulting LASSO model are reported and variables are reported in order from greatest to least importance.

Variable	Importance	Coefficient (SE)	P-value
Empathy	0.570	0.151 (0.07)	0.039
Retrieval	0.460	0.148 (0.09)	0.087
Reward	0.371	0.103 (0.07)	0.142
Mother’s conservatism	0.340	0.082 (0.03)	0.003
Father’s conservatism	0.154	0.046 (0.03)	0.081
Mother’s education	0.083	0.039 (0.03)	0.172
Father’s education	0.053	0.029 (0.02)	0.253