Bad Polls Aren’t Bad for Biden

It’s good to be reminded that Trump can win.

Image: Hillary Clinton, 2016, by Gage Skidmore, CC

President Joe Biden’s polls right now are shit. Per 538, last week Biden was trailing Trump by 1.9 points nationally. He was also trailing in virtually every swing state; behind by 9.9 in Florida, 6.5 in North Carolina, 5.4 in Arizona, 5.3 in Nevada, 3.9 in Michigan, 3.2 in Wisconsin. He is ahead in New Hampshire, Virginia, Minnesota, and (barely) Pennsylvania, but that’s small comfort. If the polls are right, and the election were held today, Donald Trump would win.

____

Hello readers! If I were to paywall this, the paywall would be here. But! I hate paywalling things. So! Encourage me not to by becoming a paid subscriber if you find my writing valuable.

___

Obviously, if you are a Democrat, or a human being with a functional brain and heart, these numbers are alarming, to say the least. A second Trump term would be a disaster for the country, for democracy, and especially for marginalized people. I personally am very much not looking forward to Trump trying to ban trans health care for my daughter, just for starters. (And by “am very much not looking forward to” substitute, “am fucking terrified of.”)

When people are fucking terrified of something, they often try to explain why that something is not going to happen. And sure enough, on social media, every new Trump-positive poll is greeted with an outpouring of more-or-less panicked denunciation. Democrats will grab you by the digital shoulder and shake you across the screen this way and that while loudly explaining that this pollster is a hack, that those crosstabs make no sense, and/or that polls now are broken and never get elections right anyway.

These furious poll rejections are often framed as if they are themselves a part of an election strategy. Poll skeptics point to pollsters who give Trump a good chance of winning and insist that they those pollsters nefarious wreckers, releasing their low numbers in order to dishearten Democrats and boost Trump into office. Bad polls are treated essentially like anti-Biden ads, filled with slanders and misstatements which must be refuted for the good of the campaign.

As someone who is (again) fucking terrified, I understand the impulse to shoot the messenger, stomp on the messenger, and make sure everyone knows that the messenger was probably in the tank for Trump anyway.

However, I think that stomping the poor messenger is, for better or worse, mostly wasted energy—even if the messenger happens to be wrong. Bad Biden polls are not going to hurt Biden. If anything, it probably helps Biden if people underestimate him.

The Upside of Being the Underdog

It’s important at this stage to note that polls are always uncertain, and that they can be wrong. More, as G. Elliot Morris notes at 538, polls this early just aren’t very predictive. Jimmy Carter was up by 14 points in mid-March of 1980; in November, Reagan crushed him in one of the most thorough electoral landslides in US history. No matter how good the methodology, no matter how pristine the crosstabs, a poll released today just doesn’t tell us a ton about who’s going to win the election when the votes are all counted.

If it’s too early for polls to be predictive, it’s also far too early for polls to influence how people are voting. Unless you’re an extremely committed partisan, you are unlikely to remember in November what the polls said in March, much less base your decisions about voting on them.

What about polls closer to the election? There is some very ambiguous evidence that people may sometimes switch their vote to a candidate who is winning just because they like being on the winning side. Researchers have had a hard time pinning down this bandwagon effect, though, and it seems like its likely to have even less influence in an electorate as increasingly polarized as ours.

There’s better evidence that when people feel like they know who is going to win, they’re more likely to stay home and not bother voting. This is one reason turnout is so low in the United States compared to other countries; because of the electoral college, many people in very red or very blue states (rightly) figure their votes don’t matter that much to the outcome of the election.

More traumatically, there’s some reason to think that polls showing Clinton solidly ahead may have reduced turnout, especially among Democrats.  Overconfidence may also have led her to misallocate resources, so that she failed focus enough attention on the Rust Belt, where Trump narrowly won the election.

So misleading polls—and even more misleading mainstream punditry which insisted Trump could not win—probably did hurt Clinton. But this wasn’t because the polls showed her doing badly. It was because the polls showed her doing too well.

The Downsides of Poll Inaccuracy

As this indicates, there are cases where inaccurate polling can be bad for campaigns. Polls help individuals and parties decide where to spend money on contested elections; if the surveys are off you can end up spending a lot of money on a hopeless race or failing to spend enough on a winnable one. Misleading data on which demographics are voting which ways can also lead you to focus your outreach on the wrong groups. (Right now, some polls are showing Black voters fleeing the Democratic party in a catastrophic manner; we should probably be at least a bit skeptical.)

It's important to emphasize, though, that the main danger here is not that the polls are too negative, and so dissuade people from voting for Biden. The main danger is that the polls are wrong, in either direction, and so lead to strategic errors.

For reasons like this, it’s worth trying to assess poll accuracy (and even more worth paying attention to polling experts like G. Elliot Morris,  Lakshya Jain, and Tom Bonier.) But it’s not generally worth denouncing polls which show Biden not doing well. The main takeaway from those polls, especially eight months from the election, is simply that Trump can win.

Nobody wants to think about Trump winning. But there a strong argument that polls which keep Democrats scared are more useful for Democrats right now than polls showing Biden ahead—and this is the case regardless of whether the polls are accurate or not. In terms of election strategy, polls showing Biden slightly behind right now are probably the best spur to donating and organizing that the Democrats could ask for. Complacency is the enemy when you want to win an election. We should, therefore, be grateful for polls that remind us that the worst can happen, unless we work together for something better.