Want these election updates emailed to you right when they’re published? Sign up here.
Every new poll seems to provide support for one of two impressions of the race: one in which Hillary Clinton is pulling away toward a historic landslide, and another in which Clinton holds a lead but Donald Trump remains on the fringes of contention.
On the whole, the data released over the past several days suggests that the race may have tightened just the slightest bit. But this seems to be the result of Trump having seen his image rebound some among Republican voters, rather than having taken any votes away from Clinton. In fact, Clinton’s standing in our national polling average — 46 percent — is the highest it’s been all year, including when she was in the midst of her convention bounce. But Trump’s at 40 percent, about 1 percentage point better than a week ago, and — believe it or not — also not far from his high on the year (Trump peaked at 41 percent in late September).
Both candidates, in other words, are slowly gaining votes from undecided voters and from third-party candidates. Emphasis on “slowly,” because there are still a lot of these voters up for grabs. About 15 percent of the electorate isn’t yet committed to Clinton or Trump, as compared to just 5 percent who weren’t committed to President Obama or Mitt Romney at this point in 2012. That’s one of the reasons why our models still give Trump an outside chance at victory. In theory, with Clinton at “only” 46 percent of the vote, he could beat her by winning almost all of the undecided and third-party voters. (In practice, there’s no particular indication that these voters have Trump as their second choice.)
These undecideds, however, aren’t distributed evenly across the various states. Florida and North Carolina have relatively few of them, for example, while New Hampshire and Colorado have more. This could affect each campaign’s strategy over the final few weeks: In states with few undecideds, it’s mostly a matter of turning out your vote; in states with more of them, voters may still be open to persuasion.
Let’s look at some data using the adjusted polling average in each state from the FiveThirtyEight polls-only forecast. We’ll look at the 16 states that we term “states to watch” — already a fairly broad group that includes states with very dissimilar races, such as Maine and Utah — plus a few others that could be extreme reaches for each candidate: New Mexico for Trump, and Texas, Alaska, Indiana and Missouri for Clinton.
I’ve sorted the states by the combined number of undecided and third-party voters (from fewest to most). I’m deliberately blurring the distinction between these groups because third-party voters often have a weak commitment to their candidates and can be picked off by one of the major-party candidates. The exception is when the third-party candidate has a viable chance of winning, which probably applies only to Evan McMullin in Utah at this point. Henceforth, I’ll use the term “undecided” to refer to this combined group of undecided and third-party voters.
|ADJUSTED POLLING AVERAGE|
|STATE||CLINTON||TRUMP||3RD PARTY OR UNDECIDED|
The fewest undecideds are in Florida, North Carolina, Georgia, Pennsylvania and Nevada. And that makes sense. These are states where both parties have their bases, with voters split heavily along racial, religious and educational lines. In these states, it’s mostly a competition to see whose base is a little larger and who can turn out more of their voters. Coincidentally or not, these states also have a lot of early voting, except for Pennsylvania. So you’re seeing a lot of campaign activity in most of these states, especially in Florida, North Carolina and Nevada. Campaigns generally think that by this late stage of the race, they can improve their margins more by focusing on turnout rather than persuasion — especially in early-voting states where the election has already begun.
The state with the most “undecided” voters is Utah, but most of them are actually McMullin voters. He has a real shot in the state (a 14 percent chance according to our polls-only model, and that’s probably too low), so he’ll probably hold on to the voters he has so far, or even pick up more. After that comes New Mexico, where Johnson has his largest residual share of support but has faded from contention, and Alaska. Trump isn’t a good fit for Alaska’s more libertarian-ish brand of conservatism, but it’s a real stretch for Clinton and, because it has just three electoral votes and is out of the way, isn’t likely to be the subject of any last-minute campaign activity. (Campaign nerds might remember Dick Cheney’s quixotic visit to Hawaii at the end of the 2004 campaign; Cheney and George W. Bush still lost the state by 9 percentage points.)
Among the more traditional swing states, Maine, Michigan, Colorado, Iowa and New Hampshire have more undecided voters than the others. You might notice that they have something in common: a lot of white voters, and particularly a lot of middle-class whites, which is one group that’s still relatively torn between the candidates.
These states are important because if there’s some sort of last-minute surge back toward Trump, he has more opportunity to make up ground in these states than in places like Pennsylvania, where more of the vote is locked in. There are Election Day scenarios where Clinton finds herself in unexpected trouble in one of her supposed firewall states such as Michigan or New Hampshire, but she squeaks by with a win in North Carolina or Florida because her turnout operation and early voting save the day. That might not be the way Clinton would draw things up, but it would still count as a win.
Overall, Clinton’s chances are 85 percent in our polls-only forecast and 83 percent in polls-plus, not meaningfully changed over the past few days.