First data on how Texas's in-person early voting electorate shifted from Day 1 to Day 2. No major counties yet, so don't read too much into it, but shifts have been modest. The two largest counties in this batch, by far, moved a couple of points toward Republicans.

The first of many daily updates to this graphic, tracking the estimated electorate in four key states. I expect all four to finish with more Republican electorates than we're seeing today, but Democrats are off to a strong start in early voting across the board.

Solid Day 1 early voting numbers for Talarico. Our model estimates Black voters made up 13.4% of the IPEV electorate, up from 11.6% in 2024, with an overall electorate of Trump+3 in a state he won by 14. Still early, though, and Day 1 has typically been Democrats’ strongest.

But I'd absolutely do it again. We need more polling, and I love seeing young people passionate enough about elections to spend their own time actually doing the work. That kind of initiative deserves encouragement.

When Wedgewood first appeared, I reached out to their head pollster to try to learn more about them. We spoke on the phone for about 15 minutes. He told me he was a college student at Boston University who had previously worked with SoCal Strategies and had paid for some inexpensive polls out of pocket. He also mentioned that the other member of the team was involved in local Democratic politics in Ohio. They struck me as nice kids with at least some understanding of polling, but obviously no professional experience. With that in mind, we assigned them a D grade and included their polling data on our site. A few years ago, I was a college student following politics as a hobby myself, so I try to give people in that position the benefit of the doubt. It's disappointing to see how this has unfolded, because it's going to be a hell of a lot harder for the next college kid who's actually well-intentioned and super into politics to break into the space after this.

We'll have more tonight, but Texas early voting opened bluer than in 2024 across key counties, here's a taste below. Day 1 shifts can just reflect changes in when people vote. We need more data, particularly by voting propensity, to have more confidence in the dynamics at play.

Democrats’ early vote lead in Georgia isn’t just about turning out their high-propensity voters faster. 2026 voters who sat out 2024 are overwhelmingly favorable for Democrats, and have become even more so over the past week, even as Republicans have gained some ground overall.

Florida Early Voting | Day 1 🔴 2024: R+25.0 (377,035 votes) 🔴 2026: R+0.4 (171,015 votes) 🔵 24.6-point shift toward Democrats Republicans still have plenty of time to make up ground, and the first day of in-person early voting has generally been their weakest.

Voters in Kansas were overwhelmingly Democratic on the first day of early voting, especially low propensity ones (KS EV will certainly get much redder over time). And while NC's electorate is becoming less Democratic overall, its low-propensity voters remain just as blue.

What makes a wave? To me, the biggest unanswered question about the midterms is what happens to Republican turnout over the next 15 days, particularly among voters who have participated consistently in past elections. At this point, the early vote data increasingly suggest that voters who sat out 2024 but are turning out in 2026 are disproportionately Democratic. That's certainly encouraging for Democrats, but we're talking about a relatively small share of the eventual electorate. In a midterm, when overall turnout is substantially lower than in a presidential election, the much bigger question is which voters who participated in 2024 will stay home in 2026. And that's something I think will have a huge part in deciding an ordinary Democratic victory and a blue tsunami. To understand why, it's worth looking at "3 of 3" voters: people who voted in each of the previous three elections. These are the most reliable voters in the electorate, but even their turnout can vary substantially between parties in midterm years. In fact, one of the interesting parts of the 2022 midterms was how differently these voters behaved across states (as seen in the graphic). Democrats held up pretty well in places like New Hampshire, Pennsylvania, Colorado, Kansas, and Maine, while Republicans enjoyed enormous turnout advantages in New York and Florida. Those differences among otherwise highly reliable voters were an important part of the very different electoral outcomes across states, even against the backdrop of the same national political environment under the Biden presidency. That's the dynamic I'm watching most closely right now: Republican 3 of 3 voters are lagging considerably behind their Democratic counterparts in early turnout across a number of states, most notably Georgia and North Carolina. But what that means for November is still far from clear. I see two broad possibilities. The first is that Republicans' early turnout problems reflect something more fundamental. Lukewarm feelings toward the Trump administration, combined with the absence of Trump himself from the ballot, could be translate into a big enthusiasm gap. If those problems extend to voters who normally participate in every election, Republicans could easily end up facing turnout deficits of 6-10 points among their most reliable voters in some key states, as we saw in places like New York and Florida in 2022. That's the kind of turnout dynamic that can turn Trump's low approval and a favorable national environment into a total tsunami. The second possibility is that Republicans are simply voting later. As I've said in past posts, Republicans have a lot of time to recover. GOP voters may be less enthusiastic and less inclined to participate early, but still broadly show up. Under this scenario, the substantial partisan gaps we're seeing in early voting could narrow considerably, looking much more like AZ and PA from 2022 than NY and FL. Of course, the eventual outcome could fall somewhere between these scenarios. And, it could differ enormously by state—that was the big story in 2022. Democrats were able to maintain strong turnout among their reliable voters in Kansas, Colorado, Pennsylvania, New Hampshire, and Maine, winning a number of important races even as their party got totally squished in California, Florida, and New York. There was no single national turnout story that adequately explained what happened everywhere. I wouldn't be surprised if 2026 produces similarly large geographic differences. And while the early evidence raises legitimate concerns for Republicans, I don't think we have nearly enough information yet to know how persistent those gaps will be. One important note is that this phenomenon is largely unique to midterms. In presidential elections, nearly all 3 of 3 voters turn out regardless of party. In 2024, the partisan gap was less than two percentage points in every state. Wyoming had the largest gap at 1.7 points, followed by Kentucky at 1.5. New York, Florida, and Nevada were all at 0.9 points, Iowa was at 0.8, and the remaining states ranged from 0.0 to 0.6. There isn't much room for high propensity partisan turnout differences in presidential years; nearly every habitual voter votes. But in midterms, when even reliable voters are more likely to stay home, those differences can become much more consequential. And that's why I'm paying so much attention to lagging Republican 3 of 3 turnout in states like Georgia and North Carolina. As we saw in Florida and New York in 2022, an early partisan turnout advantage doesn't necessarily disappear as voting continues, even if it narrows. Whether Republicans can eventually mobilize their habitual voters at acceptable rates is an open question, and the historical precedent for this is very mixed. And finally, an important caveat that I try to emphasize whenever discussing turnout: persuasion remains an enormous part of the equation, and often the most important one. If you told me today that the 2026 national electorate would have voted Harris+2 in 2024, I wouldn't have the faintest clue who would win the Senate. Even perfect knowledge of the eventual electorate wouldn't necessarily tell us the election result. But turnout can still absolutely help explain the difference between a close election and a wave. And over the next two weeks, I think one of the most consequential questions will be whether Republicans with a history of consistently voting end up coming out in large enough numbers.

If I sound like a broken record, it’s because little has changed. Democrats remain narrow favorites to win the Senate, and Republicans would need a historic upset to hold the U.S. House in our forecast. October polling has been surprisingly sparse.

Georgia Early Voting Update There is no party registration in Georgia, so models like VoteHub's early vote tracker rely heavily on past primary voting history and racial data, as well as home precinct. We can track the racial breakdown of the early vote daily using official data from the Georgia Secretary of State. Democrats are off to a strong start in Georgia early voting period. Black voters are banking their votes at high rates and much earlier than usual, and lower-propensity voters are disproportionately more Black than at this point in 2024. Black youth turnout is substantially outpacing White youth turnout relative to 2024, and among White voters, turnout in the Atlanta area is holding up better than in rural parts of the state. As it increasingly looks like lower-propensity voters who turn out in 2026 will be disproportionately Democratic, but will still make up a relatively small share of the electorate, the big question is whether Republican-leaning high-propensity voters end up turning out at the same rate as Democratic high-propensity voters (more on this from me, including a look back at 2022, later today). I expect Republican turnout to improve considerably this week, pushing the purple line up and the green line down. The key question is how quickly and how far those lines move. Republicans of course are hoping for improved lower propensity turnout, but also need to work through their reliable, three-of-three voters during the early voting period to avoid entering Election Day with an overwhelming deficit. There is still plenty of time for that to happen, but numbers after the first week suggest Republicans have considerable ground to make up.

Early vote data across the county, but especially in Nevada, increasingly suggest that concerns about USPS handling of mail ballots are changing how some Democrats vote, complicating comparisons with 2024. Per @JohnRSamuelsen’s comparisons, Democrats appear to be shifting toward early in-person voting and dropping off their mail ballots, while Republicans are becoming more comfortable voting by mail. That would help explain why Clark County’s initial mail returns looked so much better for Republicans: the Democratic registration advantage fell from 18.6 points in the comparable 2024 snapshot to about 7.5 points in 2026. But as Saturday’s likely drop-offs entered the totals, the picture changed sharply. Two later batches were D+31 and D+38.8. Democrats’ weaker initial mail numbers seem to partly reflect voters choosing to bypass USPS, rather than slacking in initial turnout. This pattern has held somewhat across the country, as in-person early voting has been very blue while mail ballot returns have been more lackluster for Democrats than expected. There are certainly other possible explanations contributing to these dynamics, including the continued unwinding of COVID-era voting patterns, when absentee ballots were generally more Democratic. But the sharp Democratic shift in Nevada’s mail returns that appear to include substantial numbers of drop-offs rather than ballots sent through USPS, is strong evidence that more may be happening here. Anyway, I do think there's solid data to support that concerns about postal delivery, amplified by media and political leaders, are encouraging Democrats to deliver their ballots themselves or vote in person. These concerns include both changes already in effect, such as transportation changes that can delay postmarks, and restrictions that were ultimately blocked. On Sept. 14, the U.S. Supreme Court rejected the Trump administration’s request to allow new USPS mail-in voting restrictions to take effect for the November 2026 midterms. Those restrictions remain blocked, but the concerns they generated may still be influencing how Democrats choose to vote.

Georgia ended a Democratic week of early voting with by far the lowest White share of any non-Sunday day in four years. The electorate should become more Republican as voting continues, but the size of that shift is uncertain. The good news for Republicans is there's still time.

Democrats are off to a strong start in North Carolina early voting, with low-propensity voters showing an even stronger Democratic lean than the electorate as a whole. At this point, the early data strongly suggests that North Carolina's 2026 electorate will be more Democratic than it was in 2024. For Republicans, the question is how much they can limit that shift. Their best opportunity is to turn out high-propensity voters at a rate comparable to Democrats, gradually narrowing the gap as early voting continues. The good news for Republicans is that there is still plenty of time to do so. Democrats, meanwhile, are hoping that Republicans' difficulties turning out low-propensity voters extend to medium- and high-propensity voters as well. That is the kind of turnout imbalance that can produce unusually lopsided midterm electorates, as we saw in the opposite direction in Nevada and Florida in 2022, when Republican turnout substantially outpaced Democratic turnout among high-propensity people who voted in 2016, 2018, and 2020. EV turnout is particularly important in North Carolina, where most ballots are cast before Election Day. Republicans need to work through their reliable, three-of-three voters during the early voting period to avoid entering Election Day with an overwhelming deficit. There is still plenty of time for that to happen, but the early numbers suggest Republicans have considerable ground to make up.

Early voting among lower-propensity voters looks encouraging for Democrats in NC and ME, with racial and geographic trends suggesting a similar pattern in GA. FL is the exception, where the electorate could be even more Republican by registration than in 2024.

North Carolina (Statewide) Day 1 Early Voting 2024: D+1.7 by registration Day 1 Early Voting 2026: D+27.4 by registration