The headline: In the latest from the “Illuminating America” series from Charted Territory, counties that voted for Hillary Clinton—Democrat counties—have a population density almost ten times higher than those that vote for Donald Trump. Regularly being exposed to a greater number—and diversity?—of people likely influences our values and voting preferences. However, “Population Density” is a surprisingly poor variable for explaining the 2016 election result: population density accounts for “only” 49.8% of the variability, scoring worse than “Robbery” and “Chlamydia” as a predictive variable.
The details: Democrat counties have a population density of 3,266 people per square mile (ppl/mi2) compared with Republican counties with a density of 366 ppl/mi2 (confidence of 100%). The correlation is a surprisingly low 49.8% (R2=0.248). This relatively low correlation could be due to variability of county size across the United States and other issues. The correlation can be increased by calculating population density as “largest city in the county divided by county area” resulting in a correlation of 55.1%.
The data: Population density (2015) data were sourced from ESRI1. 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 Excel2.
Hello, what variable are you using to determine dot size?