Where could Democrats pick up seats this November? This is the question on everyone’s mind heading into the midterm elections, as the GOP seems to be on its heels trying to fight off a blue tsunami. While national attention has largely gone to high-profile Senate races like those in Texas or Maine, most congressional races tend to get left behind in the midterm commotion. If the Democratic Party is gunning for that blue tsunami, it’s time to pay attention to Arizona’s Second Congressional District.
Cook Political Report rates AZ-02 as a Lean Republican seat right now (down from Likely Republican before Cook’s October 23rd House rating adjustment). In 2025, Cook gave AZ-02 a Partisan Voting Index of R+7. One might assume that a national environment of D+7 or higher might indicate that AZ-02 could flip blue. Fortunately for Arizona Democrats, Nate Silver, a notable elections forecaster, has generic congressional ballot (GCB) polling at a very convenient D+8.7. In fact, Silver has modeled a continued climb for Democrats in the final days of the campaign, with projected GCB polling of up to D+9.2. While one should be wary of relying on forecasted polling data, it’s certainly not a bad sign for Democrats nationwide going into November. While that’s one data point that might suggest a more competitive race in AZ-02, what do the local polls say? This is where the data get sparse. Two of the three publicly available polls for AZ-02 are Democratic internals, which show the GOP nominee leading the Democratic candidate by two to three points. The third, and most recent, actually has the Democrat up by two points. While these are loose indications of a competitive race, data from just three polls unfortunately aren’t enough to get an accurate read on the state of the race, so local polling data is essentially a wash.
I can hear potential criticisms a mile away: “But the polls weren’t accurate in 2024, what if they’re wrong again!” It’s a valid contention, but let me attempt to assuage these concerns. First, people have misconstrued the 2024 presidential election’s polling error as a colossal failure. The weighted-average error (WAE) in the 2024 presidential election was 3.4 points. WAE measures the average error of all polls published in the last 21 days of the 2024 cycle, giving diminishing influence to each subsequent poll an individual pollster publishes. While a 3.4 point WAE might seem like a lot, it’s actually the second-lowest WAE for a presidential election in the last 20 years (second only to 2004’s 3.2 points of WAE). 2020 was actually a higher WAE at five points. The difference between 2024 and 2020 was the outcome; both polling errors had the Democratic candidate ahead by more than their actual results (an average of D+4.1 and D+2.9, respectively), but in 2020, Biden was still able to make it across the finish line despite being overestimated in the polls—Harris wasn’t. The difference in outcome, coupled with a few egregious examples of polls being way off, has created a narrative that polls were incredibly wrong in 2024. In reality, they have actually performed quite well against the past two decades of presidential elections. The bigger issue is the aforementioned partisan lean in both years, as almost all of the error across the two elections overestimated the Democrats’ numbers.
This is a legitimate cause for concern, worsened by the fact that partisan error was even bigger in the down-ballot congressional races those years, peaking at D+4.9 in the Senate and D+6 in the House in 2020. However, more important than the partisan error in those races is the context in which voters selected their congressmen: These were races where the president was on the ballot, not midterm elections. When you look at the results of the last two midterms and their partisan polling errors, they actually slightly favor Republicans by a measly R+0.2 (I’ve chosen the last two cycles because those are ‘Trump-era’ midterms; if you insist on looking back at the last 20 years of midterm elections, the average stays largely similar). The WAE also falls back to 4.8, suggesting polling precision may be lower than it was in 2024; however, accuracy is unlikely to reach 2020/2024 levels of partisan error. This is all to say that I wouldn’t discount the GCB polling data based on the errors in 2020 and 2024, and if more district-level polls come out in the final stretch of the campaign, I would pay close attention to them too.
Now, let’s return to Arizona. The two candidates in this race are GOP incumbent Eli Crane and Democratic nominee Jonathan Nez, who challenged Crane in 2024 but lost by nine points. Are Democrats putting up a losing candidate once again and hoping for the best? Looking at the loss in context, Nez’s nine point loss to Crane was already an overperformance compared to Trump’s 15-point victory across the district. That means essentially one in every 17 AZ-02 voters voted Donald Trump for President and Jonathan Nez for Congress. While six points might not seem like all that much of a swing, in the age of ultra-polarization, and in a district that only has an R+7 partisan lean to begin with, that kind of candidate overperformance can make a huge difference, especially in midterm elections where we see (on average) about a 33 percent decrease in voter turnout. Nez’s past overperformance in the district is a good sign for Democrats in an election where only the most motivated voters show up—the ones who really like their candidate.
If voters in AZ-02 either liked Nez more than they disliked Harris or disliked Crane more than they liked Trump, which was it? Well, probably a bit of both. Let’s start with Crane. He’s a hardline conservative in the House, even serving on the House Freedom Caucus, which is widely considered the most conservative voting bloc in Congress. Crane is also heavily tied to the Trump administration, voting in line with Trump’s stated position a staggering 98 percent of the time in 2025. Normally, this wouldn’t be a huge problem for Crane, since Republicans are conservative and tend to vote conservatively. The issues arise when considering just how historically unpopular the Trump administration is. Trump’s second-term approval rating is almost 23 points underwater in the aggregate (even lower than Trump 1.0) and shows no signs of improving anytime soon. Crane’s association with hardline conservatism might also be more off-putting to voters under an ultra-conservative Trump administration than under a Democratic president. Why might voters like Nez? His biggest appeal is his background. Nez is the former president of the Navajo Nation, serving in the role from 2019 to 2023. It just so happens that Navajo Nation lies directly in the middle of AZ-02 (and spills slightly into other states). When you look at the official demographics of AZ-02, Indigenous peoples make up almost 20 percent of AZ-02’s population. In comparison, the median congressional district is less than one percent indigenous. Conventional wisdom suggests that voters like candidates from their own communities; for many, Nez checks that box and then some.
The takeaway is that Democrats have a real chance at picking up AZ-02. What’s the point? Why should we care about this one congressional race? AZ-02 is a microcosm of how the midterms are going overall. If Dems can flip Crane’s seat, or at least become competitive in the district, it signals a blue tsunami is coming. If Democrats are competitive in AZ-02, I would stop asking whether they’ll win the Senate and the House and instead start asking how much they’ll win by. Forget Alaska and Texas; if Democrats are winning AZ-02, they should start paying attention to Kansas and Nebraska. Those are long-shot races for Democrats, but so is AZ-02, and they also feature charismatic Democrats running against relatively unpopular Republicans—they could be competitive, too.
Nobody knows whether Democrats will actually win in AZ-02 this year. I think the wisest predictions for this race still call for a one to three point Crane victory. Even with the aforementioned points in Nez’s favor, the district is, at the end of the day, still R+7; that’s a really difficult hill to surmount in our modern political environment. I do, however, think that this race, and others like it, are being undervalued in the final stretch of this campaign season. Democrats can position themselves to compete for dozens more congressional seats and a few select Senate seats by paying attention to these races. Of them, Democrats only need to win a few to start putting a serious dent in the GOP’s presence in Washington, particularly in seats the GOP thought were safe for at least one more cycle. If Democrats want a blue wave, they should focus their attention and money on AZ-01 and AZ-06; if they want a blue tsunami, they should pay attention to AZ-02 and the other ruby-red districts previously thought unwinnable.