Qualitative Research for Political Campaigns in an Era of Hyper-Partisanship

By Dave Metz
President
FM3 Research
Oakland, California
dave@fm3research.com

By Lori Weigel
Principal & Founder
New Bridge Strategy
Golden, Colorado
lori@newbridgestrategy.com

If you haven’t read it in the news, you have seen it around your Thanksgiving dinner table—partisan polarization in America is getting worse. American politics have always provoked sincere differences of opinion, but recently, they have felt like an existential battle between good and evil. Believe it or not, a few short decades ago, each approaching Congressional election was not always labeled “the most important of our lifetime.”

Something has shifted in our political culture—the values driving partisan differences seem farther apart, the stakes seem higher, and the messages stoking political conflict seem omnipresent. These aren’t just vibes; the data bears it out.

• In January, a Gallup poll¹ reported that “Republicans’ and Democrats’ ideology is the most extreme in 30 years:” 77 percent of Republicans identify as conservative, up from 58 percent in 1994. Fifty-five percent of Democrats identify as liberal, up from 25 percent in 1994. Meanwhile, a record low of 34 percent identify as moderate, down from 42 percent in 1994.
• There were only 13 Congressional districts where voters chose different parties for president and Congress in 2024—reflecting a steady decline in split-ticket voting. By comparison, there were 58 of these districts in 2004.

These shifts signal a more fundamental change in the way Americans integrate politics into their day-to-day lives. Party affiliation no longer reflects the set of policy preferences a voter holds—it has become central to Americans’ sense of who they are.

• In an FM3 survey across the 60-odd battleground Congressional districts last year, nearly two-thirds of voters (64 percent) said politics were “important to their personal identity.”
• A 2022 study² found that 72 percent of Republicans and 64 percent of Democrats now view the opposing party as “immoral,” up from 47 percent and 35 percent, respectively, in 2016.
• And a 2020 study³ found that 71 percent of single Democrats and 45 percent of single Republicans would not consider dating someone who voted for the opposite party’s presidential candidate.

So yes, it’s rough out there. Trends suggest it’s unlikely to get better: a fragmented media ecosystem has voters relying on online media sources that algorithmically reinforce (rather than challenge) their partisan perspectives. Moreover, the structural features of our political system—from partisan primaries to gerrymandering—reward political campaigns that deepen, rather than ease, party divisions.

This hyper-partisanship complicates governing, our civic life, and (apparently) our dating prospects. It also poses problems for political qualitative researchers. It makes all aspects of the job harder: recruiting the right participants, engaging a diverse and representative group in civil conversation, and securing honest responses that yield strategic insights.

Finding ways past these barriers is essential. As Jim DeBoo, former chief of staff to California Governor Gavin Newsom, puts it: “I know there is a group of consultants who have dismissed qual, but it’s more important than ever to measure emotion, tone, and language—you cannot get that from quantitative research.”

What Makes Political Opinion Research Different

Before we explore the headaches that hyper-partisanship creates for researchers, we should consider factors that make the field of political opinion research unique.

• Firms specializing in public opinion research for campaigns do little else—almost solely providing research to campaigns for candidates or ballot measures—or for businesses, unions, or NGOs that want to influence the political or policy process. Sure, we carry out the occasional membership survey or market research project—but it’s not the core of our work.
• Most political opinion research firms are anything but neutral; we distinguish and market ourselves based on our party affiliations. This is partly due to each pollster’s personal conviction that our chosen party has the better vision for America’s future, and our desire to help advance their cause. It’s also a practical business strategy—a Democratic candidate or committee is unlikely to hire a strategic advisor who is helping them in one race while advising a Republican in the district next door, and vice versa. Partisan clients only hire pollsters who share their mission.
• Because of these partisan dynamics, nearly all political pollsters begin their careers working within the political apparatus of their party and its campaigns and then come to learn the tools of opinion research later. As a result, the work of political opinion researchers is shaped by their first-hand experiences working for the kinds of organizations that ultimately put their data to use.

All of this said, Democratic and Republican researchers both use the same set of qualitative research tools—in most cases, the same tools as market researchers in other fields. But in our work, the target audience is persuadable voters, and the “products” are candidates, ballot measures, or legislative policy proposals our party wants to advance.

It’s also worth puncturing some of the more prominent myths that have arisen around the role of qualitative research in politics.

• It’s not a hunt for “magic words.” Ever since Frank Luntz dubbed the federal estate tax the “death tax,” based on his research for Republicans in 1994, a folk wisdom has taken hold that wordsmithing can yield seismic shifts in voter opinion on public policy. Words matter, of course, but disciplined, well-resourced communications campaigns matter a lot more.
• We’re not using focus groups to wardrobe our candidates. In the 2011 film, “The Adjustment Bureau,” Matt Damon plays a Senate candidate who bemoans that he doesn’t even own his tie—which was chosen by specialists over “56 other ties we tested.” Such uses of research are not unheard of—President Bill Clinton reportedly had Dick Morris poll Americans on the most favorable vacation locations for his family in 1995—but they are limited to the most extravagantly resourced campaigns. Most research focuses on the basics: which voters to target and how to most effectively make their case.
• Qualitative research does not tell candidates what to say—it tells them how best to say it. The poll-driven politician is a common stereotype in the public mind—an empty vessel who chooses positions based on polling data and focus group feedback. In our collective decades of experience, this is largely a myth. People who choose to run for office generally do so with a clear idea of where they stand and the kind of leadership they want to provide. Research helps them frame a case for their policies that best meets voters where they are.

The best choice is often different from what expert strategists assume it will be. As Martin Hamburger, a long-time Democratic ad-maker puts it: “Nothing quite humbles a media consultant like watching their ‘brilliant’ message strategy crumble in real-time as actual voters explain why it makes no sense. Focus groups: where political egos go to die.”

The Challenges of Hyper-Partisanship for Researchers

The most prominent challenges political researchers face from hyper-partisanship occur in quantitative research. As the stakes of each election feel ever higher, anxious voters are more eager to know the score: is their “team” ahead or behind? To meet this demand, the market has been deluged with political surveys of wide-varying quality, all conveniently aggregated by voters’ favorite pundits and pumped hourly into their social media feeds.

The resulting attention is extraordinary—in California last year, FM3 polling showed that 68 percent of voters said they read the results of political polls on a weekly basis. But that degree of scrutiny means that when polling results differ systematically from election outcomes—as they did in 2016 and 2020—voters lose faith. Only 25 percent of voters in that same California sample thought political polls reflect voters’ true sentiments even “most of the time.”

This skepticism has bled through to the clients who commission political polls: often, the first question after a political polling presentation is something to the effect of “this is great, but isn’t political polling broken?” In our industry, great time and effort went into understanding and fixing polling misses in 2016 and 2020—drawing samples exclusively from official voter registration records, shifting to multimodal surveys, being more rigorous about bilingual interviewing, and setting quotas and weighting data by education and recalled presidential vote. These methods (and others) collectively helped improve accuracy in 2024. When it comes to qualitative research, however, the challenges of hyper-partisanship are different and require separate solutions.

Hyper-partisanship and Qualitative Research

The challenges hyper-partisanship causes for qualitative research are diverse, numerous, and stubborn.

• It has become more difficult to identify and recruit “swing” voters. In politics, researchers are largely uninterested in the views of the partisan bases—voters who reliably vote Republican or Democratic—but instead are focused on “swing” voters, those who may cross partisan lines based on the candidates and campaigns in a given race and often provide the margin of victory. These voters are increasingly hard to recruit, both because they are fewer in number (as reflected in the reduction in split-ticket voting noted above) and because they are harder to find. Some might think that because the share of voters identifying as “independent” is rising, as disillusionment with the parties grows, the number of swing voters is also on the upswing. But the fact is that much like partisans, the majority of self-identified independents rarely deviate from voting for one preferred party—so finding true swing voters, even among independents, is increasingly challenging.
• The most persuadable voters are the hardest to engage in political research. The factors that make voters open to persuasion—a disinterest in politics, low levels of political information, and less community engagement—also make them less likely to participate in research. In some cases, voters are actively hostile to engaging with political content—a University of Minnesota researcher has identified a group of “news avoiders” who actively resist engaging with political news. Getting these voters to participate in a focus group—even with compensation—poses a challenge.

There are also partisan differences in voters’ willingness to engage in political research: Trump supporters—especially those who are not college educated, live in rural areas, or are younger and male—are more highly suspicious of the research industry (like many other institutions) and are less likely to participate.

• Hyper-partisanship inhibits full and honest participation. Every political researcher has seen the dynamic play out—a focus group of nominal swing voters will be proceeding smoothly, and then one participant will weigh in with a strident, passionate opinion on a political issue—potentially disparaging the character of the opposition in the process. Other participants will cast looks around the room, glance at their watches, and fall silent. We have all seen political conversations devolve into shouting matches, potentially at our own family gatherings—and many participants are understandably reticent to repeat the exercise with a group of strangers, when the alternative is smiling politely and waiting to collect their incentive payment at the end of the session.
• Hyper-partisanship yields self-censorship on some of the most salient political issues. Over the last several years, some of the issues most central to political debates are ones where Americans have had great trouble having honest and open conversations—abortion, race, religion, class, and more. Exploring these issues in qualitative research means creating an environment of comfort, safety, and trust—but the hyper-partisan nature of our public debate on these issues creates just the opposite set of emotions. Passions can rise quickly in a political discussion, and when political beliefs are increasingly central to personal identity, an honest difference of opinion can feel like a personal attack—shutting down conversation.
• A hyper-partisan news media gives voters unwarranted confidence in their own perspectives—and breeds unwillingness to consider others. A colleague of ours has a label she applies to a brand of swaggering male political consultants: “Often wrong, never uncertain.” Sadly, the same description applies to many American voters. The proliferation of partisan news sources, and the increasing use of social media to access them, means that voters live within ideological bubbles designed to reinforce—and never challenge—their core beliefs. In qualitative research, this means participants enter the discussion with varying factual descriptions of what is happening in the world. It’s hard to get voters to react to new political arguments when they don’t even agree on the facts underlying those arguments. Building group consensus on the underlying nature of reality is beyond the ability of even the most skilled moderator.
• Hyper-partisanship has soured some clients on the benefits of qualitative research entirely. Taken together, these challenges have made some clients more skeptical about the value of qualitative research. Good qualitative research now takes more time and costs more money—and even in a well-designed product, clients are less trusting of the insights it provides. For many, it is tempting to jump right to a quantitative survey, or to turn to cheaper alternative methodologies that promise qualitative insights without the same kinds of messy, antagonistic conversations—like “social media listening” or feedback from door-to-door canvassers.

Conducting Good Qualitative Research in a Hyper-partisan World

As Brandon Castillo, of California public affairs firm BCFS, puts it: “Focus groups continue to provide valuable insights, so long as you’re able to separate signal from noise. In today’s polarized environment, it’s critical to design your groups to avoid ideological or partisan triggers that can turn valuable insights into a dumpster fire of an evening.” Fortunately, practitioners have developed techniques to avoid these kinds of dumpster fires. These include:

• Using masked, asynchronous online discussion boards. As an alternative to live focus group sessions, online discussion boards offer some advantages in the political realm. Masking participants so that they cannot see and comment on each other’s thoughts removes some possibility of unhelpful conflict. A one-on-one online conversation with an anonymous moderator can also encourage frankness—with participants more willing to acknowledge deviations from the “party line” than they might when seated in a group of peers. Of course, sacrificing the group dynamic sacrifices some learning and makes it hard to keep participants from “doing their own research” during the course of the conversation.
• Probing more carefully in screening to identify true persuadables. This often means going beyond partisan or ideological self-identification to understand other attitudes and behaviors that may reflect ideological flexibility. Asking people if they have ever voted for a candidate of the opposition party, or whether they consider themselves a member of the Make America Great Again (MAGA) movement, for example, can be important ways to help identify swing voters.
• Dividing focus group sessions by ideology rather than other demographic factors. Once participants are aware that others in the room share their ideology, they tend to relax, and the conversation can become much franker. In addition, participants are generally more tolerant and willing to discuss differences within their partisan coalition than they are for differences across party lines. Such segments also make it possible to test messaging with explicit partisan content—including language, metaphors, and messengers.
• Using trusted channels for recruitment. The method of outreach to voters can matter as well; in recruiting low-income people of color for in-person focus group sessions, we have had some success partnering with community organizations that work in the community—in some cases, even hosting sessions at a local church. Social media recruitment for Republicans, in some cases, may work better through Truth Social or X than other social media platforms.
• Taking care in the presentation of the moderator. In political groups, suspicious participants often scrutinize the moderator closely for any sign that they are running the group with a political agenda. Small signals—like where the moderator lives or went to school or the clothing they choose to wear—may be read by some to have great political significance. As such, the moderator should always take care to present themselves as neutrally as possible—even when conducting a session among voters who share their partisan affiliation.
• Setting appropriate expectations at the beginning of the session. Focus groups that will delve into contentious political territory should begin by giving participants a sense of what’s ahead: that they will be asked about issues where there may be profound and passionate disagreements; that they should be open and honest about what they actually think; and that they should show each other respect when the conversation gets difficult. That type of introduction can help keep participants from being caught off-guard and reduce rash, in-the-moment emotional responses.
• Finding common ground early in the session to build rapport. In sessions focused on contentious political topics, it can be helpful to have participants introduce themselves early on with details about their family, their hobbies, their pets, or what they like most about living in their community. This kind of discussion can help the participants see each other more fully as people, with shared interests, passions, and experiences that cut across partisan divisions.
• Taking care to cut through reflexive cynicism and negativity about politics. In qualitative sessions on political issues, performative demonstration of contempt for the political process is a common initial reaction. Participants will, for example, uniformly say they “hate negative ads” about candidates—but after viewing a few in a focus group, they will inevitably repeat back key passages to explain their vote choices. Several approaches may help participants move beyond reflexively cynical responses.

o Asking participants what most of their neighbors or friends may think often allows participants to reveal underlying views or values without being asked to identify those as their own personal views.

o Probing for examples, metaphors, or storytelling that push beyond the reflexive answer. For example, having voters describe the kind of animal that best represents each political party, or asking them how they might explain their views on an issue to a stranger at a barbecue.

o Focusing on hypotheticals that get participants out of their instinctive team-based responses, like asking them to describe the traits that an ideal candidate for governor might have, rather than evaluating specific individuals.

• Using a skilled, experienced, and sensitive moderator who can set their own beliefs aside. As always, it is critical to have an attentive, thoughtful, and empathetic moderator—one who can push past pro-forma and insincere answers, who can sense potential conversational minefields and navigate around them with grace. A moderator must be able to hear political opinions they strongly disagree with—or even find deeply offensive—and hear them, understand them, and accept them as part of the discussion without judgment. This can be challenging for many political professionals who, as noted previously, come to the field with high degrees of political engagement and strong partisan beliefs. To succeed, moderators need to set those beliefs aside and encourage their (usually equally partisan) clients to do the same.
• Keeping an open mind about the potential role of artificial intelligence (AI). Every field is reckoning with the role AI can play in revolutionizing how we do our work, and qualitative opinion research in the political realm is no exception. AI can potentially help address the perils of partisanship as well: in theory, once the technology matures, “silicon respondents” could be programmed to reflect the views, values, and beliefs of partisans—but without the anger, withdrawal, or irritation that mark the interactions of real people in qualitative research. The propensity for AI to “hallucinate” could even be a feature, rather than a bug—yielding unexpected and creative responses that could seed new ideas for political communications.

Looking to the Future

The bottom line is that while partisanship makes good qualitative research harder, it also makes it more essential. As Eric Jaye, founder and president of Storefront Political Media, puts it: “For years, I was a skeptic of qualitative research, having seen too many evenings and too many thousands of dollars wasted by focus groups taken over by a few opinionated respondents. But as our politics have become more partisan, qualitative research has become a valuable tool to make sure our language and base assumptions remain grounded in the vernacular and mindset of modern voters. I now leave many groups with copious notes on words and phrases used by voters to describe their feelings and experiences that I would not have otherwise tested in a quantitative survey, much less used in a script or a speech.”

In the end, the only way past our current political dysfunction is through it—all of us need to shout less and to listen more. Thoughtful qualitative research on politics is essential to learning how. Over the years, these conversations have left indelible impressions on us as researchers—we can both recall people’s faces from conversations that took place decades ago, and can remember what they told us, even though we cannot remember what the polls were telling us at the time. Continuing to hear and to learn from these human stories can help guide our politics back to a focus on what unites us as much as what divides us.

REFERENCES

1. Gallup. “U.S. Political Parties Historically Polarized Ideologically.” January 2025.

https://news.gallup.com/poll/655190/political-parties-historically-polarized-ideologically.aspx

2. Pew Research Center. “As Partisan Hostility Grows, Signs of Frustration with the Two-Party System.” August 2022.

www.pewresearch.org/politics/2022/08/09/as-partisan-hostility-grows-signs-of-frustration-with-the-two-party-system

3. Pew Research Center. “Most Democrats who are looking for a relationship would not consider dating a Trump voter.” April 2020.www.pewresearch.org/short-reads/2020/04/24/most-democrats-who-are-looking-for-a-relationship-would-not-consider-dating-a-trump-voter