Using a compilation of county-level data from the U.S. census, the Institute for Health Metrics and Evaluation (IHME), the Guardian, and Townhall.com assembled by Ella Foster Molina and Ben Warren, my team and I sought to examine whether declining well-being at a community level predicted greater growth in support for Republican presidential candidates from 2012 to 2016. Drawing on a growing literature implicating diseases and deaths of despair as a symptom of and contributor to meaningful shifts in American society, we visualized the relationship between indicators of community-level declines in well-being (county-level changes in drug overdose deaths, suicides, and deaths related to alcohol) and increased support for Donald Trump in 2016 over Mitt Romney in 2012 with a three-dimensional Plotly scatterplot (see the link below for the original plot).
As is evident in the original 3D plot, changes in deaths related to drugs and changes in suicides both exhibit visible positive relationships with increased support for Donald Trump in 2016 over Mitt Romney in 2012. A linear model including each of these explanatory variables (with changes in drug overdose deaths being logarithmically transformed) along with population change and the logarithmic transformation of changes in violent crime rates suggests that changes in deaths due to drug abuse and changes in suicides did, in fact, significantly predict increased support for Donald Trump in 2016 over Mitt Romney in 2012 (p < .001 for each).