Tuesday was National Voter Registration Day, one of the myriad awareness days that pepper the American calendar. This one, though, contains the seeds of its own destruction: If you hate awareness days and want to make them illegal, you’ll need to vote for candidates who support that idea — and guess what has to happen before you can vote?
A few weeks ago, a news site produced a map comparing turnout in the 2016 election to population across the country. (If you know where this was, please let me know; I wasn’t able to track it down.) It looked something like this, showing places where the number of votes cast in a county during the presidential election was equal to at least half of the county’s population.
That’s a somewhat unfair metric, though, because little tiny babies can’t vote (legally). Comparing the population older than 18 (according to Census Bureau data) to county presidential voting gives us a map that covers a bit more of the country.
The aforementioned map went further, though, also showing how those counties voted. For our map of counties where turnout was higher than 50 percent of the county’s full population, that map would look like this.
(The darker the red or blue, the more overwhelmingly the county supported Donald Trump or Hillary Clinton, respectively.)
As with any political map that displays geography, this map can be misleading. A lot of the counties that show up are spacious rural counties with low populations while urban counties with large populations disappear as dots. But these maps are not uninformative.
For example: We also created a version of the map above focused on turnout in the 2012 election. As with the 2016 iteration, much of the upper-Midwest sees heavier voting. That’s typical. Wisconsin and Minnesota often jockey for the highest turnout statewide in an election.
But notice the difference in the results in those counties from 2012 to 2016. In 2012, a lot more blue and a lot lighter reds as Barack Obama fared much better against Mitt Romney than Clinton did against Trump. This makes sense, given that Trump won Michigan, Iowa and Wisconsin (though not Minnesota, by a narrow margin).
A lot of these counties are big, empty rural counties. In Minnesota and Wisconsin, Clinton won more votes in counties that had vote totals over 50 percent of the population. In Michigan, Trump did.
There’s an interesting overlapping factor. Many of the counties with the heaviest Hispanic populations were included among those where vote totals were less than 50 percent of the over-age-18 population. (On the maps that follow, brighter colors indicate higher population densities.) Why? In part because many of those Hispanic residents aren’t citizens and can’t vote.
Many counties with high densities of black residents also had lower turnout.
That’s in part because white people nationally tend to vote more consistently. Black turnout increased in 2008 and 2012, but dropped again in 2016.
Among those counties where the number of votes was more than 50 percent of the total population? A lot of heavily white places: New England, Minnesota, Iowa.
We can look at this another way. The graph below combines a number of factors: white population density in a county, turnout as a function of population, overall population in a county and vote results.
Larger counties voted more heavily Democratic and are less white. A lot of very white counties with small populations preferred Trump. Notice that there are some high-population counties that voted more heavily Democratic. Those are cities that are less likely to show up on that initial map.
We’ll also note that 25 of North Dakota’s 53 counties had vote totals in 2016 that were 50 percent of the population or more. Only seven had fewer votes than represent 50 percent of the over-18 population.
North Dakota also has no voter registration.