Welcome to the Daily Kos Elections early voting roundup, which appears every weekday until Election Day. Click here to find out if and when early voting is available in your state.
Core Democratic voters, rattled early in the week by poll numbers that seemed to be sagging, got a big shot in the arm on Tuesday night with the unveiling of a new Florida poll on MSNBC’s The Last Word with Lawrence O’Donnell. The poll, conducted by the data analysts at TargetSmart on behalf of the College of William and Mary, found Hillary Clinton with a considerable 8-point edge over Donald Trump (48-40) in the absolutely pivotal Sunshine State.
While we can definitely debate whether those numbers are too good to be true (and, to be sure, only one of the last 50 polls in Florida has looked this positive for the Democratic nominee), there was a piece of data buried in that poll that immediately drew the attention of the online political community. Indeed, it was such a wild data point that it actually was, briefly, a trending topic on Twitter.
That data point also, if it is anywhere close to the fairway, might challenge a lot of the assumptions we have been making about the early vote thus far.
The “holy shit” statistic inside the toplines is probably already known to many people reading this article. It was, of course, the reveal that, among those who have already voted, 28 percent of Republicans have cast ballots for Hillary Clinton.
Now, in fairness, this seems flat crazy. Exit polling in Florida in 2012 saw Republicans voting 92 percent for Mitt Romney, so the idea that the Democrat is suddenly nabbing more than one-quarter of the GOP vote, no matter how odious or controversial the GOP nominee, just seems like a bridge too far.
But, if for nothing else than the intellectual exercise, let’s assume that the TargetSmart poll is not entirely amiss. Much has been made in the past week or so, in this regular feature, about how the early vote in Florida is not behaving in quite the same way as it did in 2012. Whereas the Democratic Party enjoyed a pretty vast advantage among in-person early voters in 2012 (a 10-point margin!), this year it’s considerably more muted (as of Tuesday’s tallies, sitting at 2.4 percent and falling). Now, while that was offset a little bit by better Democratic performance in the mail-in absentee balloting, nonetheless what was a 3-point Democratic edge prior to Election Day in Florida in 2012 is now a 0.4 percent Republican edge.
In theory, that would seem to predict (given that Obama won the state by about 0.9 percent of the vote) a narrow Republican win in the state, assuming the Election Day vote skews Republican as it has in the past several cycles.
However, the key word there is “assume.” And that, in short, is our issue here. What has driven most of the tea-leaf interpretation of early voting thus far is the built-in assumptions we have about voting behavior. Now, these are most definitely assumptions that are informed. We’re not pulling things out of nowhere here—as observers of elections, we know what has happened in the past. But the potential for error is in the faith we have that what has happened in the past is somehow ordained to happen this time. And that TargetSmart poll, whether you buy stock in it or not, is instructive because it shakes our assumptions.
Let’s get specific. When we look at early vote, we are generally looking at four factors: gender, race, partisan identification, and region (especially when the first three factors are not readily apparent). So, in Florida, what we have looked at more than anything else is race and partisan identification. And, on that latter point, we have noted that the in-person and mail-in voting universes for 2016 are more Republican than 2012 or 2008. But, if TargetSmart is right and a significant percentage of Republicans are voting for Hillary Clinton, then that D/R split is no longer a cause for deep alarm, if only because Republicans will need significantly more of their cohort to vote than Democrats, in order to offset that asymmetry in crossover voters.
But, if TargetSmart is incorrect, and Democratic and Republican voters behave as they typically do, then Clinton looks like she’d be in a fair amount of trouble. But even then, maybe not … because there is the other x-factor all this, which is the voting tendencies of independent voters. Many, like Florida Democratic analyst Steve Schale, have noted that a lot of younger Latino voters (a key Hillary constituency) are registered as independents rather than Democrats. And polls have a difficult time, as many of written about this cycle, defining what it means to be an “independent voter.”
What they’ve found is that those who register as independent (or whatever the term is for that voter in your state—it could be something like “No Party Affiliation” or “Decline to State”) tend to be more supportive of Democratic candidates than those who self-identify to pollsters as independent. The reasoning, most likely, is that a lot of registered Republicans who fall into the tea party camp call themselves independent, even if their registration is still formally Republican, out of some form of disgust with the GOP establishment. (It also explains why, in exit polling, independents wind up being a considerably larger percentage than their registration affiliation would suggest.)
In summary, we pore over the tea leaves of early voting for simple reasons: It’s close to Election Day, quality polling is at a comparative minimum these days, and we are eager for information. But tremendous caution should be applied in drawing any conclusions. While early voting does give us some important data to play with and examine, there are so many variables and cross-currents here that the bottom line is hard conclusions are elusive. Sure, sometimes early voting tells us a very clear tale (the complete implosion of the Democrats in Nevada in 2014 comes to mind), but sometimes, it doesn’t (see Iowa, 2014).
TODAY’S NUMBERS
- Nevada’s numbers, as they have throughout the early voting period, have mirrored 2012. Republicans have begun to chip away at the early Democratic edge in ballots in crucial Washoe County (Reno), but the Democrats, after a soft Monday, established a more substantial ballot edge in Clark County (Las Vegas) that more than offset Washoe’s narrow GOP win.
- Louisiana concluded its early voting, and there is another data point to raise concerns about the comparable dearth of African-American votes in 2016 during early voting. In 2012, the racial breakdown of Louisiana’s early vote was 64 percent white and 33 percent black. This time around, it was 70 percent white and 27 percent black. Now, in fairness, as many have noted, it is probably a little unfair to suggest that any Democrat (hell, any candidate) was going to match President Obama in inspiring African-American voter participation. But it’s another data point to the theory that the most loyal Democratic voting constituency isn’t turning out in numbers the Democrats might need. Now, of course, it’s not going to matter in Louisiana (where Obama lost by 17 points). But in key states, that dip could easily matter.
- In Colorado, the state is past a million votes cast, and Democrats have a more modest, but still clear, edge in ballots returned. The margin today stands at 2.6 percent, which is close to a reversal of 2012, when it was the Republicans that had the slight edge (1.8 percent).