You Should Be Skeptical of Campaign Internal Polls

If you’re a politico and on Twitter, you may follow one of the poll aggregator accounts like @IAPolls2022.1 With the election right around the corner and general election matchups set in the vast majority of states, candidates are releasing internal polls to build buzz around their campaigns. These poll aggregator accounts are posting those internals. The average person on Twitter may not realize the difference between an internal poll and an independent poll.

It’s pretty clear that the outlook is bleak for Republicans, but some of the internal polls are hard to believe. The RealClearPolling average shows Democrats with a 7-point lead on the generic congressional ballot.2 VoteHub gives Democrats a 78 percent chance of taking the House.3 Still, the partisan way districts are drawn and the partisan gerrymandering we saw in 2025 and 2026 may limit the number of Republican offices.4 That said, some of the internal polls are wild.

Old political hands used to say that an internal poll should be discounted by 3 to 6 points. That wasn’t because campaigns were making up data. It reflected the simple reality that campaigns almost always release the survey that tells the best story. Even applying that old rule of thumb, though, some of the internals making the rounds this cycle require a healthy dose of skepticism. Here are some examples.

• In IN-05, an internal poll released by Democratic candidate J.D. Ford shows him trailing Rep. Victoria Spartz (R-IN) by only 2 points. The district has a PVI of R+8 and is rated “Solid Republican” by the Cook Political Report.

• In SC-07, Democratic nominee John Vincent trails Rep. Russell Fry (R-SC) by 4.2 points. This is an R+12 district, and it’s rated “Solid Republican.”

• A poll released by Democratic candidate Chris Jones in AR-02 has him leading Rep. French Hill (R-AR) by 3 points. AR-02 is R+8. It’s another “Solid Republican” district.

• In the Arkansas Senate race, Sen. Tom Cotton (R-AR) only has a 3-point lead over his Democratic opponent, Hallie Shoffner. Look, I can’t stand Cotton. I think he is the most statist member in either chamber. However, this is an R+15 state. It’s “Solid Republican.”

I saw another poll, this one in IA-02, that had the Democratic nominee, Lindsay James, leading her Republican opponent, Joe Mitchell, by 1 point. This poll was sponsored by House Majority PAC. It’s a partisan poll, but I’m more inclined to say it’s closer to reality because it’s an R+4 district, and the Cook Political Report currently handicaps the race as “Lean Republican.”

Look, candidates in both parties need to show viability to get money in the door to run or to get the Democratic Congressional Campaign Committee or other outside group to invest in the district. A poll showing a close race is one way to get donors to open their wallets.

I want to be clear that I’m not saying internal polls should be ignored. In fact, many are conducted by respected firms using sound methodology. Campaigns need accurate data to decide where to spend money, which messages resonate, and whether they’re winning or losing. The catch is that campaigns almost never release the polls that show them getting blown out. They release the poll that puts them in the best possible light. Sometimes that means publishing a survey after favorable messaging has been tested. Sometimes it means releasing the strongest result out of several polls. And sometimes it simply means highlighting one crosstab while leaving out the rest.

When you see an internal poll making the rounds on social media, ask a few questions. Who conducted it? Was the full memo released, or just a single graphic? Does the result make sense given the district’s partisan lean and recent voting history? Most importantly, ask yourself why the campaign chose to release that poll instead of keeping it private. Internal polls are valuable pieces of information, but they’re also campaign advertisements.

None of this is to say that 2026 won’t produce some genuinely surprising results. Margins may be closer in many or even most Republican districts just because of the atmosphere. Trump’s approval ratings remain terribly weak, Democrats continue to hold an advantage on the generic congressional ballot, and history suggests the president’s party often struggles in midterm elections. At the same time, today’s political map is far less elastic than it was a generation ago. That’s what makes this election so fascinating. We could see Republicans lose the national popular vote for the House by a significant margin yet hold seats that would have been swept away in previous eras. In an environment like that, skepticism—not cynicism—is the right way to approach eye-catching internal polls.