There's a very good reason that political hacks and the hacks who love them like to insist that elections all come down to turnout: Elections all come down to turnout.
For a salient recent example, we can peek over the Atlantic at the results of the British referendum on leaving the European Union. Young Brits overwhelmingly preferred to stay in the E.U., but older ones wanted to leave — and those voters turned out more heavily. If more younger voters had gotten to the polls, the narrow result would probably have flipped to the other side. It all came down to turnout.
Younger voters voting less heavily is not unique to the U.K. We've noted before how age and turnout correlate here, as well. Younger voters — who are more likely to move frequently, have jobs that demand odd hours and may not be in the habit of voting — turn out out less than older voters, across the board.
But there are differences among other groups, too. And when we consider the new Washington Post/ABC News poll of the 2016 general-election campaign, Hillary Clinton's large lead looks a bit shakier when we consider who is firmly committed to getting to the polls in November.
We asked that question explicitly in the poll. Different demographic groups had different rates at which they said, yes, I'm absolutely certain to vote in the general election — and by comparing those figures to similar Post/ABC polling in the past two presidential elections, we can get a sense for what the electorate might look like.
By race. Seventy-three percent of white men said they were certain to vote, compared with 70 percent of whites overall. Among non-whites, the rates were much lower, with 55 percent saying the same. That's a much lower rate than in 2012 for non-white voters, which was itself down from 2008. Bear in mind, there are more non-white people eligible to vote now than there were eight years ago. As we've pointed out before, the electorate in 2014 was as diverse as it was in 2008.
Most alarming to Clinton supporters may be that only 44 percent of Hispanics said they were certain to vote in November. (Without data for that group from 2008 and 2012, it isn't shown on the chart above.)
By age. Here we can see the pattern above. Voters ages 30 to 39 have become much less likely to say they're certain to vote, while only half of those younger than 30 said in the most recent poll that they were certain to do so. Older voters are more likely to be committed to turning out in November.
By education and income. These two groups overlap to some extent (which is to say that income and education levels are correlated). Those who earn more and are more educated are much more likely to be committed to voting in November.
But let's dive into why this is a problem for Clinton, in case it wasn't already obvious: The groups that are less likely to say they're certain to vote are also groups among which Clinton does better.
Overall, slightly more Donald Trump supporters say they're certain to vote than are backers of Clinton. Seventy-six percent of Clinton backers say they're certainly or probably going to vote in November; 84 percent of Trump backers say the same.
Consider the race/gender split. We don’t yet know how white and non-white voters will cast their ballots, but we know whom they prefer in our most recent poll. If we compare those margins to the exit poll results from 2008 and 2012, the concern becomes apparent: same support from non-white voters, but less certainty of actually voting.
Clinton gets the same support from non-white voters in our current poll as Barack Obama got in 2008 and 2012. But at this point, those voters are much less likely to say they're certain to vote than they were four and eight years ago. Meanwhile, Trump does about as well with white men as did Mitt Romney four years ago — and white men are just as likely to say they're going to vote as they were then. White women are much less supportive of Trump than Romney, but they are also less likely to say they're certain to turn out to vote.
Part of this probably overlaps with the general dissatisfaction with both candidates. White women are less likely to support Trump than they were Romney and less certain to vote — perhaps because neither option was palatable. That may also be the case with younger voters. They prefer Clinton to Trump but heavily backed Bernie Sanders in the Democratic primaries. They, too, may be more indifferent to voting in November as a result.
It's worth noting that the figures above include data from people who aren't necessarily registered to vote. Just among those who are registered, though, the graph above doesn't change much. Registered voters are more likely to say they're certain to vote (again, in part, because voting tends to be a habitual act). The most noticeable change here is that the drop among non-white voters is much smaller than in past polling.
But then we loop back to those figures among Hispanic voters. The Democrats will and are putting a huge emphasis on registering and turning out Hispanic voters, a group that heavily prefers Clinton. In fact, the graphs above tell the story of the most recent elections: When the electorate is younger and more diverse, Democrats often do better.
This election, like every other, will come down to the voters who make it to the polls in November. There may be a more succinct way to say that, which we will leave as an exercise to the reader.