Mad Poll Disease Redux

Don’t panic about new horse race polling: Americans know the difference between answering a survey and casting a ballot, even if the media doesn't.

In September, I published “A Cure for Mad Poll Disease.” In light of the latest New York Times poll that shows Biden lagging in almost every battleground state, it’s time for your booster shot. 

According to that survey, voters are ready to return Donald Trump to the White House with at least 303 electoral votes – despite his criminal attempts to overturn the election and his promises to invade Mexico, have shoplifters shot, and deploy the Justice Department against his political enemies – because they feel Biden hasn’t done a good enough job on the economy. So much so that Biden can be sure of garnering just 225 electoral votes, fewer than any Democratic presidential nominee since Michael Dukakis, 34 years ago. 

While the Times has gone to great lengths to substantiate its methodology, it has provided no other evidence to support the idea that the voters who rejected Trump-boosted and -boosting MAGA candidates just one year ago in Arizona, Georgia, Michigan, Nevada, Pennsylvania, and Wisconsin (by even greater margins than they had in 2020 and the 2018 Blue Wave midterms!) are now ready to go all-in for MAGA by 5, 6, 5, 10, and 4 points respectively.

Similarly, the Times reports that support for Trump among “non-white, under 45” voters increased from 29 percent to 42 percent. I’m pretty sure if there were actually another nearly 3 million such Trump supporters, we’d see evidence of it beyond this survey. At a minimum, the Times should have marked how unlikely that was rather than making it a subheading. (By the way, I’m pretty sure that nothing close to that margin share drop among a group as large as 13 percent of the electorate has happened any time recently in a real election. Clinton didn’t fall that far behind with white non-college men in 2016, for example.)

But, as I explained in “A Cure for Mad Poll Disease,” asking about one poll’s accuracy is the wrong question. Regardless of the methodology, pollster, or publication, all horse race polling is worse than useless. For one thing, polls simply aren’t accurate enough to predict close races (which, given the Electoral College, are the only races that matter) with precision. This is true no matter when a poll is fielded, but especially when the election is a year away. But more importantly, even if we could perfectly survey every voter in America about who they think they would vote for today, it can’t tell us what will happen between now and Election Day that might change their minds. And it can’t tell us how being in an actual voting booth might change their actions. Americans know the difference between answering a survey and casting a ballot, even if the polling industrial complex and pundits don’t. 

Horse Race Polls Don’t Measure the Horse Race Anymore 

Let’s begin with something that may be difficult to hear, but is the essential context for all horse race polling now – and why no matter how many times we take the patient’s temperature, the result doesn’t tell us anything about her cardiovascular health. 

A majority of Americans have been dissatisfied with the nation’s most important institutions – including both the Democratic and Republican parties, as well as the leaders of those parties – almost continuously for nearly 20 years. It has been, and will always be, the case that in that environment, the governing party will be deeply unpopular and a majority of voters will tell pollsters they want change. 

As I explained in April in “Don’t Panic About Biden’s Approval Ratings,” for three-quarters of the last 20 years, presidential approval has been underwater. When people take polls, they consistently vent their frustrations with their choices. But when they cast their ballots, they’ve shown themselves to be better aware of the consequences of choosing MAGA – which is why MAGA has lost 27 of the last 32 races in those six states. (To be clear, I raised the same logic in 2020, when polling by the Times and others showed Biden leading by 10 points or more in Wisconsin, etc; this isn’t about polls being bad for Democrats.)

Worse than Wrong: Making the News Instead of Reporting the News

There are four ways in which the regular drumbeat of horse race surveys from the most respected media institutions is changing the course of events by making news instead of reporting it by fully and accurately describing what’s really going on: a fascist movement that has already made alarming progress at taking over much our government, and that seeks to take it over fully. 

• Reinforcing the lie that the indictments are political. The surveys back up Trump’s claims that Biden knows he would lose a rematch, so the only way to stay in office is to prosecute Trump.

• Electability. It’s really surprising to me that none of the coverage of the vanishing prospects of Republican primary challengers has given credit where credit is due – to those mainstream surveys. Roll the tapes back to the aftermath of the midterms, just a year ago. Even in some Republican circles, Trump was seen as a liability. The path for Republican challengers was clear – run as a more electable Trump. But, as the MSM kept hammering on how electable Trump is, that argument evaporated.  

• Rationalizing both-sides coverage. Imagine the world exactly as it is, except there are no mainstream media surveys. There would be no reason to believe that roughly half of America buys into what MAGA is selling – and no reason to treat the MAGA GOP like a legitimate participant in the democratic process, instead of the unhinged fascist movement that it is very clearly behaving like. How might the public’s attitudes change in a media environment where “everyone knows” that the Republican Party is a danger to democracy? Instead, we have a self-perpetuating cycle of self-fulfilling prophecies. The media’s obsession with horse race polls normalizes the abnormal.  

• Breeding fatalism. The more the media treats polls as the most relevant election news, the more people will think that what’s going to happen next November is possible to know – and that their own actions are therefore not relevant. To show how “responsible” they are, outlets caveat their findings with a vague “things could change.” But putting the results on the front page with color graphs that are impossible to unsee belies the idea that those caveats are anything more than pro-forma CYAs.

Narrative Frames Have Consequences

The worst part of the Times’ and others' coverage of their own polling is their absolute confidence, despite evidence to the contrary, in their declarations of what matters to voters. While the media cannot tell voters what to think, it can tell them what to think about. 

After the second impeachment trial, the media had two frames to choose from in covering the midterms. The first, especially given the positions of the challengers, would be as a national referendum on the Republican Party’s complicity in the criminal conspiracy to overturn the results of the elections. That’s the kind of frame we saw after Watergate, and Republicans paid a price for it even though they tried to distance themselves from Nixon. But leading up to the 2022 midterms, the media chose the second frame – the poli-sci idea that midterms are always referendums on the incumbent administration, especially when the incumbent holds majorities in the House and Senate. In the past, the second frame wouldn’t even have come up. 

Consider the differences in the way the Senate and House midterms were covered. In the Senate, it was clear to voters in Arizona, Georgia, Nevada, Pennsylvania, and Wisconsin that they were not choosing between two candidates. They were choosing between two profoundly different directions for the country, because the most important vote their senator would cast would be for majority leader. On the other hand, because the savvy media treated a coming Republican House majority as inevitable, the droves of anti-MAGA voters who stayed home in California, New York and New Jersey had no idea that their staying home would allow Republicans to achieve that majority. In those places, to the extent that House races were even covered significantly, the coverage contrasted the candidates’ positions on particular issues, as if it weren’t the case that the only vote a member of the House makes now that matters is for Speaker of the House. The media plays an indispensable role in sustaining the fiction that it makes sense, say, to vote for a Republican who espouses moderate issues positions closer to your own when those moderate issue positions have no chance of even getting a hearing in a GOP-majority House.  

History Has Its Eyes on You

I’d like to ask members of the media this question directly: If Trump wins – and if he fulfills any of his long list of deranged promises, some of which involve breaking America beyond repair – how do you think history will judge how you covered this election? Will future generations ask whether voters had a good enough idea about how other voters would vote, or would they ask whether voters had a good enough idea of what was in store for them?

The foremost journalistic institutions in the country should know very well that a lot will happen next year – including Trump’s trials. And no poll can ever tell you how you vote when something that hasn’t happened yet occurs – or that the media knows but still hasn’t reported! Remember that in the time between the Dobbs leak and the Kansas abortion initiative, the media was telling us that their polling showed Dobbs wouldn’t matter because, as with J6, few voters were changing their minds about Roe. But it turned out that a hell of a lot of voters were changing their minds about how important Roe was. It’s one thing to rank abortion in a list of other issues on a survey when the conventional wisdom is that Roe isn’t seriously threatened. It’s quite another to enter a voting booth with the knowledge that one of the two parties is an imminent threat to your own or your loved ones’ right to choose.

The media needs to decide whether they are covering this election as if it’s an election like any other, or the election that will decide whether the MAGA movement succeeds in ending American democracy. As long as the media chooses the first option, it is creating a self-fulfilling prophecy. It shows that the media has learned nothing from the consequences of asking “what about her emails” far more often than they asked “what about Trump’s fascist aspirations,” or not covering the uncertainty about the House majority that they did about the Senate, or condescendingly dismissing as catastrophizing for political effect those who contended that Gorsuch, Kavanaugh and Coney Barrett had their fingers crossed behind their backs when they pledged allegiance to stare decisis at their Senate hearings. 

And as long as we have more confidence in the media’s ability to see the outcome than in our own ability to affect it, we surrender before the battle for our freedoms begins. 

Weekend Reading is edited by Emily Crockett, with research assistance by Andrea Evans and Thomas Mande.