Don't Let Polls Poison Your Brain

What both centrists and progressives too often misunderstand.

Photo by Annie Spratt on Unsplash

Last week, a bunch of Democratic centrists got together in a Washington hotel to talk about pulling the party back from what they perceive as its ill-considered dash to the left. At the event, called WelcomeFest, there was lots of what we used to call hippie-punching, long a favored pastime of the Democratic establishment, and praise for the moderates who found ways to win in red-leaning districts. At one point, one of the organizers told the audience, “We need five more Joe Manchins,” a questionable idea for any number of reasons.

The prevailing theme (you can watch the whole thing here) was that the way to win is to figure out what voters care about and what they think, then convince them that you too care about those things and agree with them. Which does sometimes work. But as ever, centrism is a content-free ideology; the centrist doesn’t know what he believes until you tell him where the right and left are, so he can place himself between them.

I’m not going to adjudicate the argument between the center and the left right now; my sentiments lie with the progressives, but I also have plenty of critiques of how some on the left go about doing politics. What I want to discuss instead is the way polls and focus groups — the “figuring out what the voters think” part of this equation — have so often formed a cage that imprisons Democrats.

Back in the early 2000s, I wrote a number of articles and reports about how, if you looked at polls, Democrats had an advantage on almost every policy issue (and I was hardly the only one saying this). On item after item, a majority of respondents would say they preferred the Democratic position, whether it was increasing the minimum wage or expanding health insurance or protecting the environment. And yet, Democrats at the time were struggling to win elections. I argued that they had to stop being so apologetic and start displaying some confidence, since they were in an advantageous position.

The problem with being right

I haven’t been able to find it online, but the wise Michael Tomasky, who was then the editor of the American Prospect, wrote a piece arguing that Democrats were actually being undermined by their advantage on issues. Their pollsters presented them with the data, and their conclusion was “Great, all we need to do is talk about policy.” Republicans looked at the same data and understood that they had to think more creatively, explore new frontiers of demagoguery, and build campaigns around identity, not issues.

We’ve gone through a lot of political ups and downs in the last two decades, but this remains true: For Republicans, polls are a starting point that tells them what they need to change — not about themselves, but about the world — and what they should work around. For Democrats, polls too often form the bars of that cage they build around themselves.

Both the centrists and the progressives are convinced that their policy preferences are the more popular ones, and if only Democrats would advocate for those positions, success would inevitably follow. They both marshal data to make their case, and there are so many polls and studies out there, and the evidence is complex enough, that if you want to you can construct a seemingly persuasive case for either position. Americans want single-payer health care, if you ask the question the right way. But they also want tough policies on crime — again, if you ask the question the right way.

The “popularist” strategy advocated by the centrists is to advocate for popular positions and do popular things. All other things being equal, of course it’s better for a candidate or a party to take a position that’s more popular than one that’s less popular. But — and I cannot stress this enough — all other things are never equal.

Politics is not a controlled experiment, and it’s not a static environment. The experience of being a voter is nothing like taking a survey or sitting in a focus group, artificial situations that don’t resemble the way people actually imbibe political information. Every message politicians send comes accompanied by a hundred competing messages, exactly what one tries to control for when exploring public opinion and testing messages.

This isn’t a knock on pollsters; in my former life as an academic I spent a lot of time writing and analyzing surveys, and the people in that profession understand very well the limitations of their methodologies. But this is one of the ironies of polling as a means of determining political strategy: The more carefully you work to isolate the effect of different variables or determine the impact of particular messages, the farther you travel from what’s actually happening in the world.

Why message testing didn’t help Kamala Harris

As Jeff Hauser points out in his insightful critique of WelcomeFest, a group called Future Forward, which some of the same people present at that conference were involved in, spent almost a billion dollars in 2024 testing messages and ads on behalf of Kamala Harris to find and air the most effective ones. During the campaign, the effort received some wide-eyed media coverage, premised on the idea that Democrats had moneyballed the election and must surely be on their way to success: A bunch of really smart people with almost limitless resources were finally applying rigorous methodologies to discover the most effective messages.

Yet they failed. How could such a thing happen? Well, imagine I offer you a piece of penne and a piece of fusilli, each prepared with a different combination of spices. You spend a few moments munching on each one, then tell me you like the penne better. OK, I say, and I drop a few pieces of penne, prepared in the same fashion, into a stew that contains literally hundreds of other ingredients, then I offer you a bowl of that stew.

Will you enjoy it? Maybe, or maybe not. It’ll depend on what else is in the stew.

Look, for instance, at the policy issue driving the crisis of the moment: immigration. Democrats are terrified of talking about immigration, because they’ve seen polls saying that more people trust Republicans on the issue, and that, generally speaking, majorities want undocumented immigrants to leave (they’re also frightened of any issue where they think Republicans look “strong” and they look “weak”).

But the truth is that the public’s opinions about immigration are complex and contradictory. You can find polls saying that a majority of Americans want less immigration and they think immigrants make America better and they want undocumented immigrants to leave and they think Dreamers ought to be able to stay. To people immersed in politics, this is hard to understand, because political obsessives think in terms of “issues” about which they have coherent and consistent beliefs. Ordinary people don’t.

How have Republicans responded to that complexity? By attempting to shape public opinion in order to make their preferred policies more popular. For instance, they would like people to believe that every undocumented immigrant is a terrifying terrorist gang member who can barely find enough hours in the day to eat and sleep between all the time he spends raping and murdering. So they invest a great deal of time into convincing people that’s true. Does it convince everyone? No, but it helps open up the space for them to pursue their preferred policies when they have the chance.

There are lots of issues where we can see this contrast: Democrats look at public opinion as something static, a guide that tells them what they should say and what they should avoid, while Republicans see it as something dynamic that can be pushed in their direction, even if only by small increments.

Where does this leave us? We have to be aware of some important dangers polls present. First, you can see polls showing something you’d rather not be the case — Oh no, this belief I have is unpopular! — and decide that they represent an unchangeable feature of the world, and you might as well not advocate for the things you believe. Second, you can see polls that show your preferred ideas are popular, and decide that not only is the work of persuasion basically done, but that there is only one path to success.

That’s the pathology that seems to afflict the “popularist” centrists. They look at polls showing the popularity of some of their positions, and the fact that some moderates have won elections in some places, and insist that the argument is over and theirs is the only way to win. The truth, however, is that there are lots of ways to win. And the argument is never over.

Thank you for reading The Cross Section. This site has no paywall, so I depend on the generosity of readers to sustain the work I present here. If you find what you read valuable and would like it to continue, consider becoming a paid subscriber.