Description: In the third round of Charted Territory’s “Illuminating America” series1, county-level divorce rates only explain 25% of the 2016 election result, the poorest explanatory variable to date.
Chart: Opinion of climate change happening (x-axis) plotted against Trump vote for the 2016 election. Each bubble is an individual county. All counties above 50% on the y-axis voted for Trump, counties below voted for Clinton. Bubble area is proportional to county population. Bubble color depicts whether the county is in a Red or Blue state.
Observation: With an r-squared value of only 0.05, only 25% of the county-level 2016 election can be explained by the county-level divorce rates (increasing divorce rate is weakly associated with an increasing Trump vote).
1Rank of each variable; 2Variable name; 3Percentage of counties that have data coverage; 4Regression type (linear; exponential; or power); 5Data weighted by county population; 6Correlation between variable and Trump vote; 7R-squared value; 8Percent of county-level results correctly predicted (value outside parentheses is the average success of predicting Republican and Democrat counties separately).
Methods and data: County-level divorce rates are taken from the National Center for Family & Marriage Research with the Center for Family and Demographic Research2. County-level 2016 election results were compiled by Charted Territory by visiting each state’s Secretary of State web page. The data were compiled and visualized using Microsoft Excel3.