I have a certain way of going about election prognostication, and it is based on Real Clear Politics poll averages. This approach served me amazingly well when looking at the Senate elections, in which I got 100% of the predictions correct last year. I didn’t get everything right though, as I predicted a Democratic takeover of the House (they gained one seat). Although I correctly predicted that Trump would win, I did not predict that he would win Michigan and Wisconsin. I try hard to combat my personal bias in prediction and I will try to do so again this year. Let’s set some parameters of the possible in this election based on past elections:
I suspect an average Democratic bias in polls, but it won’t be as big as Republicans hope; there’s too much consistency, too much of a trend. Yes, its possible some races will flip in their favor if the Democrat is shown on average ahead by a point or two, but don’t expect a Wisconsin 2016 catastrophic polling error (Hillary Clinton was projected to win in average polling by 6.5!). That’s too ridiculous.
. Trust RCP classifications of Likely and Leans races: No Likely or Lean races have ever come out in favor of the other party.
. 5 out of 8 elections have had a net of Democratic bias in polling. Two of the three elections that had Republican bias were 2010 and 2012, the other was 2022.
. The largest margin by which a candidate was leading and the election went the other way in poll averages for the Senate was Nevada 2022.
. Leading in all late polls, although often an indicator of success, isn’t always: Russ Feingold in 2016 in Wisconsin and Chuck Cunningham in North Carolina in 2020 were leading all existing polls but the Republican kept his seat. Feingold led by 2.7 and Cunnigham led by 2.6. The most dramatic was in 2022 with Republican Adam Laxalt at 3.4 points ahead in polling, but Democrat Catherine Cortez Masto pulling the win by 0.9. The magic number here is 3.4. No Senate candidate has overcome a leading margin over 3.4 in polling. With the current polling an if this historical record sticks, this places victory out of reach for Mike Rogers in Michigan and Roger Marshall in Kansas, lest an unprecedented error occur. Dan Sullivan in Alaska, Josh Turek in Iowa, Jon Husted in Ohio, and Ken Paxton in Texas are still possible winners.
. Since 2014, average polling has been biased to the Democrats, with 2022 being the sole exception. In 2010 and 2012, the polls were biased to the Republicans. On a side note, I remain haunted by the failure of the concept of “unskewed polls” in 2012, and I have seen some strong coping energy from Republicans in the lead up to this race. However, it is also possible that the late surge in polling is similar to the 2022 surge for Republicans in the last week, which turned out to be illusory.
. There is going to be at least one upset in RCP polling averages for the Senate. It has happened every single election year! I defied this in 2022 and was wrong.
For the results, for the House, I think that Democrats get between 20-30 House seats. Those who think the House is staying Republican are frankly kidding themselves, especially given Trump’s approval rating.
My calls:
Gubernatorial Races
Florida
Byron Donalds (GOP) wins, he has consistently been ahead in polling and Florida’s fundamentals are too strong.
Georgia
Keisha Lance Bottoms (D) wins. Early voting is looking good for Democrats and she has been polling ahead recently.
Iowa
Rob Sand (D) wins, he has been leading all polls.
Kansas
Ty Masterson (GOP) wins against Democrat Cindy Holscher in the only gubernatorial pickup the GOP gets this year, and its going to be narrow. Despite a recent polling bump for Holscher I think the state’s fundamentals will work for Masterson.
Michigan
Jocelyn Benson (D) wins; John James (GOP) has not led a single poll.
Nevada
Aaron Ford (D) wins, defeating incumbent Joe Lombardo (GOP).
New Hampshire
Kelly Ayotte (GOP), who has been favored to win, wins despite recent narowing of the polls.
Ohio
Amy Acton (D) wins, she has been consistently polling ahead and Republican Vivek Ramaswamy’s ad has certainly not helped his position.
Oregon
Despite negative headwinds for her, Tina Kotek (D) wins. Oregon voters are going to think more about Trump than her this year.
Texas
Greg Abbott (GOP) wins based on the Republican fundamentals of the state; he doesn’t have the baggage of Ken Paxton.
Vermont
I get the feeling Phil Scott (GOP) does survive this one, his brand is too separate from Donald Trump’s; he endorsed Kamala Harris in 2024.
Wisconsin
David Crowley (D) wins, he has consistently led Congressman Tom Tiffany (GOP).
Senate
In Alaska, the second “Dan Sullivan” scandal and Mary Peltola’s campaign’s involvement in it gets Senator Dan Sullivan another term, just narrowly. Peltola wanted to establish that she was different from other politicians, but this scandal casts doubt on this idea. Right now I am calling this one as the one in which there is an upset in RCP poll averages.
In Florida, Senator Ashley Moody keeps her seat. Although the 2026 environment is unfavorable to Republicans, Florida is now too Republican and Democrat Angie Nixon is too radical to win the seat.
In Georgia, Senator Jon Ossoff retains his Senate seat against Trump favorite Representative Mike Collins. Collins has not led in a single poll. He’s not even in striking distance.
Congresswoman Ashley Hinson wins the Iowa Senate seat against Democrat Josh Turek narrowly. The RCP average is tied, but Iowa is still a Republican state, the state has a history of over-polling Democrats, and she doesn’t have some of the distinct negatives that the next Republican has.
Until last week, I thought Roger Marshall was going to hold on by the skin of his teeth in Kansas. I remember when Independent Greg Orman was thought to be a threat to Pat Roberts but Roberts held on by 10 points in 2014. However, 2014 was a great year for Republicans and Orman on average was only up by less than a point. 2026, however, is looking to be a rough one. Democrat Adam Hamilton appeals to a number of people as a moderate and Marshall has faced backlash not only as the top Trump surrogate in Kansas but also his frequent time spent at his second home in Florida (which he only recently sold) and he has gotten bad press for his medical practice taking legal action against a number of his patients.
Until last week, I thought Susan Collins holds on in Maine and that she’s done it before and she’ll do it again. The problem is that this race is far more frequently polled than the 2020 race in which Democrat Sara Gideon led in all polls. Indeed, this time the race is probably better examined by the pollsters and Collins no longer outpaces the partisan lean of her state.
Abdul El Sayed wins the Senate seat in Michigan…no polls as of late have shown Republican Mike Rogers as leading, and in order for him to win given the current polling, an error nearly as bad as the Hillary Clinton 2016 polling errors in Wisconsin would have to occur. This is despite an anti-Semitic ad from Cenk Uygur (which dates to 2024 funny enough) coming out on his behalf to which his response is not exactly satisfactory and Haley Stevens’ endorsement of Rogers. Most endorsements these days just don’t weigh much on the scale in the anti-establishment mood everyone’s in, with Trump endorsements in Republican primaries being a major exception.
Although Minnesota looked interesting, the state’s voters are going to overlook the fraud scandals under the Walz Administration and elect Lieutenant Governor Peggy Flanagan to the Senate. The state’s Democratic lean as well as Trump’s unpopularity is just going to keep the DFL up for now.
Although the notion of Republican John Sununu making a comeback in New Hampshire is interesting, the state’s voters are once again going to be inclined to vote Democrat for Congressman Chris Pappas. Sununu has not led or even been close in a significant poll as of late.
In Nebraska, Senator Pete Ricketts looks like he’s going to survive despite a scare from Independent Dan Osborn.
In North Carolina, Democrat Roy Cooper winning the Senate seat was and remains a foregone conclusion as he was a popular governor of the state.
In Ohio, Democrat Sherrod Brown pulls off the rare feat of a defeated senator winning back his seat. This is not only because of this year being a backlash to Trump but also because Brown has far more name recognition than Jon Husted, who as an appointee of Governor DeWine has not made much of an impression aside from being a Republican and Trump’s recent observation about his posterior. We live in interesting times!
In South Carolina, Darline Graham Nordone holds on to her brother’s seat. South Carolina is too Republican to flip, even under these ideal conditions for Democrats.
In Texas, James Talarico prevails…too many people don’t like Ken Paxton, who in my opinion is the worst choice Republicans have made in a primary in the South since they nominated Roy Moore in Alabama back in 2018. Yes, conservatives will say that Talarcio is “weird”, but I hold that Paxton’s issues likely outweigh Talarico’s among the general electorate. If I was to look at elections without thinking about party, I would consider El Sayed and Paxton to be the worst candidates this cycle, and it is entirely possible both of them win, but my money’s on El Sayed and Talarico. Flaws in Democratic candidates are just largely going to be overlooked this year because the issues of Trump and inflation are just more in the minds of the public.
In conclusion, Republicans lose the Senate, holding 48 seats, putting them in the worst position they have been in the Senate since after the 2012 election. Some things that are pretty incredible have to happen for this outcome, but with a president with around a 39% approval rating based on inflation, a war in Iran that he didn’t make the case with the American public beforehand for that is contributing to said inflation, and tariffs contributing to said inflation. The picture overall looks grim for Republicans, but my message for Republicans is…don’t let my prognostications discourage you from going to the polls…perhaps YOU can play a part in proving me wrong! And for Democrats, if you think you can defy my predictions on Alaska and Iowa or elsewhere, be my guest! By the way, I have below a history since 2014 of polling bias in RCP averages and the actual margins of victory. A positive bias is one towards the Democrats while a negative is one towards the Republicans. They are quite illuminating!
