- Computer Science
We report state-level relationships between measures of well-being (e.g., IQ, income) and racial minority composition as predictors of this century's U.S. presidential election outcomes. In bivariate analyses, race only weakly predicted votes cast for democrats. Instead, large mutual suppression effects existed between racial composition and well-being. For example, when race appeared in the regression, the IQ sub-domain of well-being predicted votes cast for democrats in all elections since 2000. Likewise, when IQ (or any other well-being sub-domain) appeared in the regression, race strongly predicted votes cast for democrats. Suppression effects emerged because of negative correlations between well-being and minority composition, yet positive correlations between these variables and election outcomes. In sum, states with high well-being tended to favor democrats, as did states with larger minority populations.