Journalists are hungry at this point for any information that might give them some special insight into the eventual outcome in November, and early voting seems particularly valuable because, unlike a poll, early voting reflects, obviously, actual voting. What’s much less clear is what we can conclude from early turnout at this juncture. As Michael McDonald warns routinely in his blog posts, these are early indicators, and may reflect changes in campaign mobilization, changes in state and local practices, along with voter interest and enthusiasm.
We have, for example, no research that I know of that correlates absentee ballot requests, returns, and eventual vote totals in a state. Comparisons between 2016, 2012, and 2008 seldom take into account in a systematic way spending rates and mobilization efforts by campaigns. Speculation about these things may make for good copy, but I don’t spend a lot of time poring over these numbers.
I am quite willing to speak to reporters, however, on the topic of “voter regret,” and whether early voting is “bad” because of the dynamic nature of “this year’s campaign.” These questions come in predictably just about this time, when the early vote totals really start to pile up, and reporters are on full alert, filing multiple stories about the campaign.
Here’s the basic math: only a relatively small percentage of total votes have been cast at this point (two weeks out from Election Day). And a relatively small percentage of registered voters (and an even smaller number of likely voters) are truly undecided at this late stage. This translates into a tiny theoretical percentage of voters who will cast an early vote and who are likely to change their vote, ignoring in that calculation lots of research indicating that most early voters are decided voters who are not going to be swayed by information provided over the next two weeks.
This sounds like a big number, but we expect over 130 million ballots to be cast in 2016. The early vote totals to date are about 5% of the total expected. Three days ago (October 21), it was less than 4%. Five days ago, the total was under 2%. That bears repeating: 20 days before the election, less than 2% of ballots were cast early.
Now, let’s see how many voters are really likely to be undecided at this stage of the campaign. According to Lynn Vavreck, a YouGov conducted three weeks ago found just 8% of registered voters were undecided. These undecided “… are less interested in politics and the news, less partisan, and less likely to hold opinions on issues dominating campaign discussions. Essentially, they think less about politics.” A HuffPollster study conducted a week ago found a somewhat higher percentage of undecided, 14%, yet Natalie Jackson found little space for movement within this segment.
Now we can put everything together. Who are these “early” early voters? Are they voters who will change their minds? All the evidence we have accumulated from past elections says “no.” Few early voters report that they regretted their choice, and most early voters report that they made up their minds relatively early. They tended to be more ideological and partisan than the average voter. Now, we have never had a study that samples early voters on the fly, after they cast their ballot, and follow up later on. Instead, we have to rely on post-election reports, and there may be some rationalization going on.
Nonetheless, as much as it may disappoint reporters, for the vast majority of citizens, the campaign is essentially over. It’s possible that there may be a late breaking scandal, or late breaking information, or a late breaking event that could move millions of citizens to reevaluate these two candidates who have been in the klieg lights for nine months. Every reporter is dying to break this year’s October surprise. But those surprises almost never occur, and given how strongly voters feel about these two candidates, there is very little that will move voters at this point.
Election “day” has begun! Enjoy the ride.