Do political polls mean anything anymore?

When I was a kid, nothing made me more excited than seeing Jim Ranchino on KATV, Channel 7, sharing the results of his latest survey on Arkansas elections. (Yes, I know … a geeky kid!) Ranchino, a Ouachita Baptist University professor who wandered to Arkansas after completing his graduate studies at the University of Wisconsin, modernized the study of Arkansas elections and had a knack for explaining the numbers behind politics — both in polling and election results — to the public. As many older Arkansans remember, he died suddenly on election night in 1978 on the KATV set at the incredibly young age of 42. Even though I never met Jim, he was a real role model for me. Indeed, a George Fisher cartoon of him — signed by Ranchino — is right by my office door today.

So it’s been incredibly rewarding to play a role in survey work for the Talk Business & Politics/Hendrix College poll since 2010, a period of truly extraordinary change in Arkansas political dynamics. When the work began, Democrats still controlled almost every aspect of Arkansas politics. Within a few years, that dominance was replaced by a new one-party Republican rule that has only been cemented over the years. 

Watching a state move so emphatically in a new partisan direction and helping to tell that story through polling data and election outcomes has been fascinating. Moreover, polling has caught some of the shifts for the public a few months in advance of Election Day. By the time most of you read this column, we’ll know if any cracks have begun to show themselves in Arkansas’ Republican rule. 

We’ll also soon know the status of American public opinion work. 

On one hand, Democrats grew increasingly confident of gaining control of both houses of Congress and a majority of governorships across the nation starting around late August, based to a great degree on a bevy of polling data showing momentum for their party. On the other hand, ghosts of polling errors from the three past presidential election cycles that led to unexpected election night results still give party activists shivers. 

While national poll numbers have been off in elections when Trump was on the ballot, poll-based predictions have been solid in off-year elections across recent cycles, showing just one more way Trump is a distinctive American political figure. 

Still, concern over polling data validity was buttressed by major polling errors from the late summer in high-profile primaries — especially the Michigan Senate race and the Wisconsin gubernatorial race. Primaries are notoriously difficult to survey because voters can quickly change their minds based on campaign events, and when important candidates exit from races (as was the case both in Michigan and Wisconsin), numbers are thrown off.

No matter the final 2026 results, we know that survey work is dramatically different from the Ranchino era, when literally every home had a landline telephone as their key mode of communication with outsiders and they had a general trust of sharing their views on elections with those who surveyed them. Now, response rates are incredibly low even when respondents are offered compensation to participate. 

Moreover, there are now any number of modes of communication: landlines, cellphones, texting and the web. Subsets of the electorate are more at home communicating through each of these methods (or not at all!), making it difficult to create a truly accurate sample of the electorate. In this cycle, for instance, pollsters have consistently found that those who talk with live voices skew much more Republican, and text survey respondents skew decidedly more Democratic. Determining whom to communicate with for surveys with each method and then how to accurately mix them all together in a complete sample is more art than science for pollsters. Moreover, it’s ever-changing, and carrying out the research to determine the right mix for a given election is as costly as activating the array of methods. Importantly, most news outlets that sponsor the highest-quality polling have had to shrink their budgets rather than investing in newer strategies for enhancing the confidence of the results. An array of polls that lack transparency about their methods has made it harder for the public to separate the ongoing good work from the chatter. Still, there’s absolutely no doubt that survey work has a central role in public opinion research. It is important for use by campaigns in determining the messaging that has the most promise of resonating with voters, and it is important for enhancing the public’s confidence in election results in general.

But the way we gauge voter sentiments is changing fast. The most insightful public opinion research this election cycle has not been in the realm of polling. Instead, it’s been with a significantly different methodology: focus groups. Rather than attempting to replicate the electorate in a given election, focus groups bring together like-minded voters to share their responses to key questions about the political world. 

Increasingly, virtual technology is used to more easily bring these voters together, cutting the cost of focus groups. In every election cycle, there are a handful of voting groups who are crucially important for campaigns and analysts alike to better understand. In this election cycle, the most important group included those who cast votes for Donald Trump in 2024 but now disapprove of his performance in office. They make up about a fifth of the electorate and they are the reason that Democratic chances of victory have swollen in unexpected places like Texas, Kansas and Iowa. Their reactions to candidates and issues tell us as much as polls about what’s happening in the 2026 election.

The most visible practitioner of focus group data right now is Sarah Longwell, a former Republican who became publisher of Bulwark. Each week she shares snippets from focus groups, particularly from those disaffected Trump voters across the country. They are fascinating, leading to great analysis and an important reminder that real people, with their quirks and passions, drive this democracy.

While polling unquestionably has a place in understanding what makes voters tick and will never go away, even as it morphs, expect as we move forward to see more resources spent by campaigns and press outlets on focus groups. It’s a good investment.