What polling says about key 2024 Senate races

When it comes to the 2024 election, the contest between Vice President Kamala Harris and former President Donald Trump naturally sucks up a lot of oxygen. But plenty of other important races are on the ballot this November, including some closely fought elections that will determine control of the U.S. Senate — and thus influence whether one party could hold all the levers of power in Washington. With Democrats holding a 51-to-49 edge in the Senate (counting the four independents who caucus with them), the contest for Congress's upper chamber looks to be on a knife's edge.

With that in mind, 538 is proud to announce the release of its polling averages for Senate races! As of today, 16 of the 34 Senate elections on the ballot this November — including the handful of races that will most likely decide control of the Senate — have met 538's criteria for an average, which include having at least five different polls conducted by at least three unique pollsters in the 300 days before the general election. (Check out the methodology section below for more details on how our aggregation model selects and weights polls to produce these averages.)

Let's first look at what our averages say about the races that are most vital to the basic Senate math: Democrats' three most endangered seats in Montana, Ohio and West Virginia. These three contests undoubtedly constitute the GOP's top Senate targets in 2024, given that Trump carried each state in 2016 and 2020 and looks very likely to do so again this November. Moreover, vote outcomes in presidential and Senate races increasingly mirror each other: Just one state that held a Senate race in 2016 and 2020 voted for different parties for president and Senate (Maine in 2020). And the West Virginia seat is already an all-but-certain Republican pickup, thanks to the retirement of Democratic-turned-independent Sen. Joe Manchin — likely the only non-Republican who had any chance of winning the Mountain State race.

Montana and Ohio, then, are the leading races to watch, and Republicans have reasons to be bullish on their chances of flipping Montana, where GOP nominee Tim Sheehy is challenging three-term Democratic Sen. Jon Tester. 538's polling average has Sheehy ahead by about 2 percentage points, 47 percent to 45 percent. Should Republicans capture West Virginia and Montana while no other seats change hands, it would give the GOP an outright 51-to-49 majority.

Now, we're certainly not writing off Tester, who has won three races by an average margin of fewer than 3 points. This includes an almost 4-point win in 2012, his one victory in a presidential cycle, when he narrowly trailed in the polls. However, that year, Republican presidential nominee Mitt Romney carried the Big Sky State by about 14 points, whereas Trump won it by 20 and 16 points in 2016 and 2020, respectively. 538 doesn't yet have enough polling for a Montana presidential average, but Trump has led Harris by 15 or more points in most surveys, which suggests his margin of victory might once again be larger than Romney's. That means Tester may need even more voters to split their tickets and back him along with the GOP presidential nominee than he did in 2012.

The news for Democrats is somewhat better in Ohio, where Democratic Sen. Sherrod Brown is also seeking his fourth term. Per 538's average, Brown leads Republican nominee Bernie Moreno by around 3 points, 47 percent to 44 percent.

Before it took a sharp right turn in 2016, Ohio had long been a swing state, and Brown had impressively flipped this seat in 2006 with a 12-point margin — defeating current Ohio Gov. and then-Sen. Mike DeWine — and held onto it by 6 points in 2012 while President Barack Obama carried the state by only 3 points. Once we entered the Trump era, Brown showed his survival skills in 2018 when he won reelection by almost 7 points, although he was undoubtedly aided by that cycle's "blue-wave" electoral environment. This year, Trump leads Harris in Ohio by around 10 points, so while Brown likely won't need nearly as many split-ticket voters as Tester, he will still need a fair number of voters who back Trump to vote for him, too.

While those three seats offer the lowest-hanging fruit for Republicans to pick, the GOP also hopes to make inroads in five other seats currently held by Democrats (or independents who caucus with them). Two of these seats are open following retirements — Arizona and Michigan — while the other three have Democratic incumbents defending them — Nevada, Pennsylvania and Wisconsin. Time will tell, but Democrats currently lead in all five Senate contests, and are considerably outperforming Harris's numbers in each state.

In Arizona, Democratic Rep. Ruben Gallego holds about a 7-point lead over Republican Kari Lake. The intensely pro-Trump GOP nominee has underwhelmed so far: After narrowly losing the state's 2022 gubernatorial election — and refusing to acknowledge her defeat — Lake became the odds-on favorite to lead Republicans in this Senate race. But Lake had a lukewarm victory in the GOP primary, and Gallego has outraised her around three-to-one in the money race.

Democratic Rep. Elissa Slotkin's fundraising has been even more dominant in Michigan's Senate race: She's raised more than four times as much money as Republican Mike Rogers, a former House member himself. Overall, Slotkin enjoys around a 6-point edge in 538's average. Rogers will hope that outside spending from conservative groups can make up some of that gap. However, of all spending on the general election by outside groups, the combined amount of pro-Rogers and anti-Slotkin spending is about even with spending for Slotkin's camp (around $25 million for both sides).

We have every reason to think that the presidential race in Nevada, Pennsylvania and Wisconsin will be extremely close, yet the Democratic incumbent in each corresponding Senate contest holds roughly a 5-point lead or better over their GOP challengers. In Pennsylvania, Sen. Bob Casey Jr. enjoys a 6-point edge over Republican Dave McCormick, while in Wisconsin, Sen. Tammy Baldwin is up by about 5 points — but unlike most of the candidates in these key races, she is actually a hair above 50 percent in her average vote share. In Nevada, Sen. Jacky Rosen holds a larger 9-point lead over Republican Sam Brown, although we shouldn't be surprised if the race is closer by Election Day (Rosen's 5-point win in 2018 was the largest margin of victory in the past four Senate contests in Nevada.)

While Republicans have many Democratic-held seats to target on the 2024 map, Democrats have few places to go on offense. Democrats and independents who caucus with them control 23 of the 34 seats up this year, and of the 11 seats held by the GOP, very few seem to offer any chance for Democrats to flip them. In Texas, the closest of these races, Sen. Ted Cruz is ahead of Democratic Rep. Colin Allred by 3 points, so it could be quite competitive. Meanwhile in Florida, Sen. Rick Scott holds about a 5-point lead over Democrat Debbie Mucarsel-Powell. Additionally, in Nebraska, Sen. Deb Fischer only leads independent Dan Osborn by about 4 points — Osborn hasn't said which party he would caucus with — but with a huge share of undecideds and the Cornhusker State's red lean, it's hard to imagine the GOP actually losing that race.

We also have averages in a handful of other contests where one party is more clearly favored to win. In Maryland, Republicans have hoped that popular former Gov. Larry Hogan could potentially flip an open seat in that blue state, but Democrat Angela Alsobrooks is up by 6 points in our average. Incumbent Democratic Sens. Martin Heinrich of New Mexico, Tim Kaine of Virginia and Amy Klobuchar of Minnesota each lead their GOP opponents by around 10 to 12 points, while Democratic Rep. Adam Schiff is up by around 25 points in dark-blue California. On the GOP side, Sen. Josh Hawley of Missouri leads in his race by 10 points.

These averages will continue to update as we close in on Election Day, and we will of course add Senate averages for additional races as we get more polls, but we already have numbers for the contests most likely to decide control of the Senate. Given what 538's average says about the Montana race — and the fact West Virginia will almost certainly flip — it's easy to see why the GOP is favored to win control of the Senate.

Methodology

The 538 Senate polling averages follow a similar methodology to our averages for the 2024 presidential race. We start by taking all the polls from firms that meet our methodological and ethical standards. When polls ask support among multiple populations (say, likely and registered voters as well as all adults), we prefer the most restrictive population available. And like our other polling averages, we filter out "tracking" polls that share interviews with voters across multiple polls. Specifically, we remove the polls with overlapping dates (so if we saw three tracking polls from one pollster, conducted from Oct. 1-3, Oct. 3-5 and Oct. 5-7, we would throw out the poll from Oct. 3-5). Lastly, we filter out polls that ask about hypothetical candidates.

For each race, we then use those polls to calculate our polling averages. We assign more weight to polls that were conducted more recently, and polls that have higher pollster ratings, which score firms based on their empirical track record and how much methodological information they disclose. We then subtract any systematic differences among polls attributable to the pollster that conducted each poll (which we call "house effects"), the mode of interview it was conducted in ("mode effects"), and whether it surveyed likely voters, registered voters or all adults ("population effects").

538 also has a long history of gathering partisan surveys, too; our philosophy is that any data that gives us reliable information on trends in public opinion can be valuable if we make the right adjustments and weight the data appropriately. Because we found in our study of historical surveys that polls with a Republican- or Democratic-leaning pollster or sponsor tend to underestimate the other side, we include partisan polls by adjusting each poll for the partisanship of its pollster or sponsor. Please read the methodology about our presidential race averages for more information about how, precisely, our aggregation model computes weights and adjustments and combines polls.