Table of Contents
Most political pundits live in a bubble, but Sarah Longwell spends her days listening to real voters in focus groups, uncovering truths that might surprise even seasoned campaign strategists about what actually drives American elections.
Key Takeaways
- Undecided voters aren't conflicted intellectuals—they're busy, normal people who tune into politics late in the game and view it cynically
- Political focus groups reveal the "why" behind voter decisions through tone of voice and authentic reactions that polls can't capture
- The enthusiasm gap between candidates matters more than polling numbers because excited voters influence their less political friends
- Misinformation spreads through a "triangle of doom" between politicians, right-wing media, and voters who reinforce each other's beliefs
- AI could revolutionize political research by analyzing thousands of focus group transcripts to identify patterns and predict voter behavior
- Authenticity remains the most powerful tool for political persuasion, which creates challenges as AI becomes more sophisticated
- The Republican Party has fundamentally changed and won't return to its pre-Trump form because newer voters joined specifically because of Trump
- Focus groups work best when participants feel safe with their "tribe," leading to more honest responses than mixed partisan groups
- Technology like Zoom transformed focus group research during the pandemic, making it more scalable and accessible than ever before
- Compromise has become a dirty word in politics, but rebuilding trust in collaborative governance is essential for democracy's future
The Hidden Reality of Undecided Voters
Here's something that'll probably irritate every political junkie reading this: undecided voters actually exist, and they're not who you think they are. Sarah Longwell, who's been running political focus groups for years, gets this question constantly from highly engaged political podcast listeners who can't fathom how anyone could be "undecided" between candidates or parties we all know so well.
The thing is, those highly engaged listeners are the weird ones. They're the minority. Most Americans aren't sitting around consuming political content all day—they're living their lives, raising kids, working demanding jobs, taking care of sick parents, volunteering in their communities. When election time rolls around, especially for presidential elections (since many skip midterms entirely), they look up about a month out and think, "Alright, which of these two evils is the lesser one?"
It's not that they've thoughtfully considered both options and genuinely can't decide. They're what political operatives call "late-breaking attention payers." They believe voting is important, but they approach politics with a pretty cynical worldview. These voters represent a critical swing demographic that can determine elections, yet they're completely misunderstood by the political class.
What makes this even more fascinating is how these undecided voters actually process information. Unlike political junkies who've already made up their minds and are looking for confirmation, undecided voters are genuinely weighing their options based on limited information they encounter in their daily lives. They might catch a headline while scrolling, hear something from a coworker, or see a campaign ad during their favorite TV show. Their decision-making process is fundamentally different from people who follow politics closely.
Why Focus Groups Beat Polling Every Time
There's this ongoing debate in political circles about turnout versus persuasion, but anyone actually trying to win elections knows you need both. The 2020 election perfectly illustrated this—both sides turned out every voter they could while the race ultimately came down to soft GOP-leaning independents in swing states like Pennsylvania, Michigan, Wisconsin, Arizona, and Georgia.
Traditional polling can tell you the "what" but focus groups reveal the "why." When Longwell sits with voters, she's not just collecting their opinions—she's reading their tone, watching their body language, listening for hesitation or enthusiasm. She wants to know if they're being dragged kicking and screaming to vote for someone, or if there's genuine excitement there.
Here's what most people don't understand about how focus groups actually work: there are specialized firms that maintain massive databases of people with detailed demographic and voting information. Most of the time, these lists are used for commercial research—like whether people prefer cat food in blue bags or yellow ones, or what sounds appealing on Chili's new menu. But when political researchers need specific voter segments, they can slice and dice these lists incredibly precisely.
Want to talk to college-educated women who voted for Trump in 2016, Biden in 2020, and are currently undecided in Pennsylvania? They can find those people. The key is getting deep enough cuts to find exactly the demographic that matters for your research question.
One of the most valuable things focus groups reveal is what's actually breaking through to voters versus what's creating massive controversies in Washington. Longwell remembers doing a focus group in Ohio right after Trump's Helsinki summit with Putin, where he sided with Russia over American intelligence agencies. In DC and political media, this was treated as a massive scandal. But when she asked focus group participants about it, they had no idea what she was talking about. None of them had heard about it.
The Enthusiasm Factor That Polls Miss
Something shifted dramatically in Democratic politics when Kamala Harris entered the race, and it wasn't something you could easily measure in traditional polling. Longwell had been watching focus groups for months where Biden supporters would say things like, "Obviously I'm voting for Biden because I hate Trump, but God, is he old. I hope he doesn't fall over during his next speech."
That person is definitely going to vote because they hate Trump, but what about the less politically engaged person they're talking to? When someone expresses reluctant, fear-based support, it doesn't inspire their friends and family to get involved. But when someone says, "Oh my gosh, I am so pumped to vote for Kamala Harris. I love how she talks about the middle class, I love that she's a prosecutor," or even just "I'm high on vibes, I'm huffing vibes"—that enthusiasm translates.
The metrics that matter most aren't poll numbers that fluctuate by one or two points daily. They're surge in voter registration, surge in donations, surge in volunteering. These indicate genuine enthusiasm that can spread through social networks in ways that traditional polling can't capture.
Fear and anger are powerful motivators—probably the strongest human emotions for political engagement. But for too long, anti-Trump organizing was built primarily around these negative emotions, which can create fatigue over time. What Harris brought was something for people to be "for" rather than just "against." When you combine both—something you're really for and something you're really against—that's an incredibly potent combination.
How Misinformation Actually Spreads
The way misinformation takes hold isn't what most people imagine. Take the "stolen election" narrative as a perfect example. Right after the 2020 election, most Republicans didn't think it was stolen. Trump was saying it, but it hadn't penetrated broadly yet.
What happened next illustrates what Longwell calls the "Republican triangle of doom"—a toxic, symbiotic relationship between right-wing infotainment media, elected officials, and voters. Trump and some extreme elected officials started pushing the stolen election narrative. Certain media outlets picked it up because there was voter appetite for that information. Viewers who wanted to hear it gravitated toward those outlets, which gave politicians incentive to appear on those shows and repeat the claims.
Pretty soon, you go from this being a fringe Matt Gaetz and Trump thing to Kevin McCarthy saying the election was stolen. When a more serious, mainstream figure like McCarthy endorses the claim, a whole other set of voters who didn't initially believe it start thinking, "Well, if Kevin McCarthy says it, maybe I should believe it too."
This cycle continues until 70% of the Republican Party believes something that started as a fringe conspiracy theory. Fox News had to compete with Newsmax and OAN, so they promoted claims they knew were false. Everyone worked together, consciously or unconsciously, to make the lie dominant.
AI's Promise and Peril for Political Research
The technological transformation of focus group research started during the pandemic when everything moved to Zoom. Suddenly, researchers were teleported into participants' living rooms—sometimes people were smoking, vaping, babies were crying, dogs were barking—but they still got about 85% of the valuable feedback they needed.
More importantly, Zoom sessions created perfect transcripts using AI transcription tools like Descript or Otter. Now imagine having transcripts from hundreds of focus groups and being able to ask AI questions like: "When we ask about immigration, what are the top three responses from two-time Trump voters?" The AI could analyze patterns across massive amounts of qualitative data that would be impossible for human researchers to process manually.
The next level would be predictive analysis. Feed the AI hundreds of different voters talking about various issues, then ask it to predict how certain demographics will respond to new developments. For example, after RFK Jr. dropped out of the race, you could ask the AI to find all previous discussions about RFK and predict how two-time Trump voters would react to his endorsement.
But the most exciting possibility is scaling qualitative research to thousands of people while maintaining the richness of individual responses. Instead of yes/no polling questions, you could ask open-ended questions to thousands of people simultaneously and get back summaries with direct quotes, word clouds showing common themes, and searchable responses that capture human inflection and nuance.
The challenge will be maintaining authenticity. Political persuasion depends heavily on authentic messengers—real people sharing their genuine experiences and opinions. As AI becomes more sophisticated, there's a risk that people will lose trust in any messenger, assuming everything might be artificially generated.
The Personal Journey Behind the Research
What makes Longwell's perspective particularly valuable is her personal journey navigating political identity and authenticity. Growing up as a conservative in small-town Pennsylvania, she entered the conservative movement right out of college, working for Rick Santorum during the height of the gay marriage debates—while struggling with her own sexual identity.
She describes a pivotal moment at a Santorum event where gay protesters showed up, including two moms with their tween daughter holding a sign that said "My two moms take me bowling." Standing there as a closeted young staffer, she thought, "I quit. I'm out of here. This is I can't do it."
That experience taught her what it felt like to be hated by everybody—Democrats didn't like gay Republicans ("oxymoron, like jumbo shrimp"), and Republicans didn't like her because she was gay. She had to learn how to be comfortable with everyone being mad at her, not having a place in either tribe.
Working on marriage equality within Republican circles from 2006 to 2013 showed her that massive opinion shifts are possible. She watched the entire country change its mind about gay marriage in real time. This experience shaped her belief in persuasion and her understanding that people can change their minds when approached correctly—usually through personal connection and storytelling rather than abstract policy arguments.
The Future of Political Persuasion
Looking ahead, the biggest challenge might be rebuilding basic civic norms that have eroded. Voters now want policy outcomes but don't want politicians to compromise—the very mechanism by which policy actually happens. They're mad about the "uniparty" and have no patience for reaching across the aisle.
When you no longer believe in compromise as a legitimate way to achieve goals, you start reaching for strongman solutions—leaders who promise "I alone can fix it." Rebuilding trust in collaborative governance might be the most important work happening in politics right now.
The Republican Party, according to Longwell, isn't going back to what it was before Trump. The voters don't want it to. People who joined the party in the last decade did so because of Trump, not in spite of him. Young Republicans today don't understand why someone like Nikki Haley would be appealing—they believe Republicans should be isolationist, pro-tariff, anti-immigration populists.
There might be a healthier version of the Republican Party that becomes a multi-racial, working-class coalition focused on populist economic issues, as long as they abide by rule of law and accept election results. But it won't look like the old Republican Party of Romney or Liz Cheney.
Technology, Truth, and Trust
As AI becomes more prevalent in political research and communication, the fundamental challenge will be maintaining the authenticity that makes persuasion work. People are experiencing a collapse of trust in institutions—politicians were never highly trusted, but faith in media, courts, and other traditional authorities has dropped significantly.
The most trusted messengers are still "people like me"—ordinary folks sharing relatable experiences. This is why testimonial-based political advertising often outperforms slick campaign productions. But as AI-generated content becomes more sophisticated, maintaining that sense of authentic human connection will become increasingly difficult.
The solution isn't to abandon new technologies but to use them thoughtfully. Just as good uses of AI will be needed to police bad uses of AI, the key is ensuring that beneficial applications develop faster and remain more accessible than malicious ones.
One thing focus groups consistently reveal is that despite our polarized politics, people share more common ground than our media environment suggests. Participants are funny, volunteer in their communities, take care of sick parents, worry about their kids and the country's future. They often mention that online everyone seems angry, but when they go outside, people are pleasant and wave to each other.
That more positive world exists—we just need to figure out how to seize it and bring it back. We don't have to be angry all the time.
