This Tuesday is the big day.
If you think I mean National Saxophone Day, please remember to put your sax down on November 6 and go vote! Otherwise, voter turnout for the election will be nothing to toot your horn over.
As it is, the percentage of people who turn up at the polls to vote is much lower in the U.S. than in many other democracies around the world.
Or is it? Surprisingly, a lot depends on how you define the variable you use to measure voter turnout.
The Response: VAP vs VEP
Voter turnout is often calculated as the percentage of persons who voted out of the total number of people of voting age (> 18), called "VAP" for short.
What’s wrong with VAP? Age isn’t the only factor that defines voter eligibility in the U.S. Lack of citizenship, a felony record, and other factors can also make a person ineligible to vote.
If you adjust VAP to exclude people who are not eligible to vote for reasons other than age, the turnout is higher than one might think. Take a look at the voter turnout by state as defined by VAP and VEP for the 2008 presidential election.
The clustered bar chart created in Minitab Statistical Software shows that for some states, the voting age population comprises almost entirely eligible voters.
For example, in Maine and Vermont (the only states that allow felons unrestricted voting rights, by the way), VAP is very close to VEP. So the turnout for the 2008 presidential election was about 66% in Vermont and about 70% in Maine, regardless of which metric you use.
But in other states, it’s a different story. Compare VAP and VEP turnout for California—an extremely populous state. Based only on age eligibility, voter turnout in the state is less than 50%. But once you exclude other types of ineligible voters, the actual turnout rate is about 62%. In fact, for most states the turnout rate based on VEP is markedly higher than VAP (check out New York, Florida, and Nevada).
So although conventional wisdom says that voter apathy in U.S. elections has increased over the years, and that turnout has declined, some statisticians argue that once you use VEP instead of VAP as your measure, the turnout rate in the U.S. presidential election is pretty much on par with what it’s been historically since WWI—about 60%.
The moral of the story: it’s not only critical to measure your response variable accurately, you need to define it carefully as well. The difference could make or break your analysis.
At any rate, we certainly could do better with our voter turnout, even as defined by VEP.
Can you imagine a national voter turnout of 80%? As shown on the bar chart, that was the percentage achieved by the hardy, civic-minded citizens in my native state of Minnesota in the last presidential election.
“Not too bad,” a Minnesotan might say.
So on this election day, let’s all follow the lead of Minnesotans. Not just by voting, but by asking others, in a friendly tone, as a true Minnesotan might ask it:
“So, are you going to vote, then?”