RT @ryanjmccomb: "On Point Polling," which just released a South Carolina poll, is run by the team behind The Public Sentiment Institute (T…

Some less concise thoughts for those who want them: That's what early vote analysis is ultimately about: trying to figure out who is going to show up and vote in November. In some elections, that can be incredibly informative. In 2024, the electorate that turned out was simply different in its composition from the electorates we saw in 2020 and 2022, which was more of what pollsters were assuming. Most likely voter (LV) polls are currently pointing toward a fairly Democratic-friendly electorate, particularly compared to 2024. For those who remember, once pollsters started applying their likely voter screens around six weeks ago, the numbers began moving noticeably toward Democrats. And so far, the Early Vote data I've seen has been broadly consistent with what those likely voter polls are suggesting. Across the board, the people showing up to vote in person appear to be disproportionately Democratic. We first saw it in Virginia and Ohio, and now this week's early voting in Georgia and North Carolina has been particularly concerning for Republicans. Mail ballot return rates in New Jersey and Arizona tell a similar story. Nearly every measure of enthusiasm I've seen, both in surveys and in early voting, currently favors Democrats. At some point, the question becomes how much of this holds through Election Day. Do lower-propensity Trump voters eventually turn out in sufficient numbers to keep states and districts that Trump won by double-digits red? Republicans will need to recover substantially from what we're currently seeing in Georgia and North Carolina, or they could face some very difficult results in November. As I've said before, I think there's a real possibility that the floor falls out for Republicans this November. The 25th-percentile outcome for the GOP is an my mind an exceptionally ugly result for the party. It's possible that Republican voters simply aren't as motivated to turn out at the beginning of early voting, or even before Election Day, but will ultimately show up in sufficient numbers to produce a more competitive (albeit still Harris-won) national turnout picture. There are, of course, places where Republicans appear to be holding up better. Florida's mail ballot returns look relatively encouraging for the GOP (we'll get a better picture of in-person voting there next week), Michigan's mail vote hasn't been especially concerning for Republicans (although I do still think we are headed toward a Harris-won electorate). Democrats may also have some relative weaknesses in New York, Oregon, and California. And I haven't even gotten into poll nonresponse or how undecided voters might break. I don't want to become too focused on turnout simply because we're getting so much early vote data. Even in midterm elections, persuasion is at least half the battle. How do Paxton, Husted, Hinson, and Sullivan perform among voters who supported Trump in 2024 but now somewhat disapprove of his administration? How many Harris voters can Rogers and Collins persuade to cross over? We can look at polling, but no one quite knows, and those questions matter just as much as the turnout patterns we're observing. The early evidence is certainly encouraging for Democrats, although so far it isn't pointing hard at anything radically different than what we are seeing in likely voter screens. Republicans still have time to recover, and will recover at least some. Still, the picture is becoming increasingly consistent across different states and different types of data. But we're still trying to understand the composition of an electorate that hasn't finished voting, much less how those voters will ultimately cast their ballots. There's a great deal we can learn from early voting, but also a great deal we simply won't know until November.

wow @earlyvotedata @ZacharyDonnini @TonerousHyus @DataRepublican this is not good for us

Georgia early voting update (12:34 PM): Total voters have increased from 60,704 to 81,623 (+34.5%), excluding Other/Unknown. Black voters now account for 37.6%, down from 38.1% (−0.4 points). White voters increased from 57.9% to 58.5% (+0.5 points).