We're Not Going To Find Out Tonight Who 'Won' The Debate

The 2020 election ’s debate season is officially getting underway, which means it’s time for a quick refresher on why this week’s Democratic primary debates won’t necessarily end with clear winners and why, even if they do, we won’t immediately know who they are.

Twenty of the Democratic candidates will take the stage Wednesday and Thursday night in Miami. Debates aren’t necessarily game-changing events , but the circumstances surrounding these ones gives them a better-than-usual shot at having at least some impact. As political scientist Julia Azari notes , debates seem to hold more importance in situations exactly like this: early on in a primary election with a crowded field of candidates, many of whom are unknown to voters. 

But the most immediate data we’ll have on hand things like social media reactions or focus groups aren’t really cut out to tell us how Democratic voters are going to respond, because they’re rarely an accurate reflection of the electorate in its entirely. To wit:

Reactions on social media like Facebook and Twitter are not representative of what voters think. Twitter users, according to one recent Pew study , are younger, more highly educated and higher-income than the population of the U.S. at large. In another study , Democrats who posted political content online were found to be more liberal, more educated and less diverse than those who didn’t do so.

Focus groups are not representative of what voters think. Back in the first debates of the 2016 presidential primaries, focus groups concluded that Republicans had turned against Donald Trump and that Democrats were deserting Hillary Clinton .

“While these discussions make for far more compelling television than dry survey statistics, they have important limitations,” Mark Blumenthal, formerly the polling editor at HuffPost, wrote in a 2008 column. “Every group is a small, non-random sample, and it is hard to know the degree to which the views of participants may be influenced by the atmospherics of the telecast, the probes of the moderator or the opinions expressed by others in the group.”

“Reader polls” that anyone can take online are not representative of what voters think. These polls the type you see posted on Twitter, or embedded in news articles don’t attempt to properly weight their responses along demographic lines and may allow people to vote multiple times. That makes them, at best, unrepresentative, and at worst, subject to intentional manipulation which is why they often end up rewarding candidates with enthusiastic online fan bases over those with more widespread support. (Note: the problem with these polls isn’t that they’re online there are plenty of scientific web-based pollsters with procedures in place to conduct representative surveys. The problem with reader polls is the lack of any provisions for sampling or weighting and the lack of even basic safeguards against being gamed by online mobs.)

Even after more reliable polling on the debates starts coming in, there are a couple of things to keep in mind.

First, what happens after the candidates leave the stage may matter at least as much as what happens during the debate. Not everyone in the Democratic electorate is going to tune in per one recent poll , 43% of Democrats said that they didn’t plan to watch the debates or that they weren’t sure if they would. Partly because of that, how the media covers a debate is likely to have a significant effect. “Debates don’t just affect those who watch them; they can also influence the political environment by how they are covered in the media,” Azari writes , adding: “A breakout moment is more likely to happen if the news media agrees that it happened. … If the post-debate media narrative is more muddled, we’re less likely to see a big shift in the race.”

Second, sampling error means that horse-race polling numbers naturally fluctuate to some extent, even in the absence of an underlying change in the campaign they’re tracking. (As a case in point, one poll HuffPost conducted recently found Joe Biden’s name recognition to be 4 points lower than it had been in a survey taken two weeks earlier. It wasn’t because people had forgotten who he was.)

“In presidential elections, even the smallest changes in horse-race poll results seem to become imbued with deep meaning,” Pew Research’s Andrew Mercer explained in a post about sampling error. “But they are often overstated.” Complicating matters further, people are sometimes temporarily more reluctant to answer surveys after their preferred candidate has a bad day.

It’s not at all a given that any candidate will clearly “win” or “lose” the debate, or that a particularly strong or weak performance will translate into movement at the polls. But if a candidate truly surges or plummets, by the very nature of those words, it probably won’t be subtle. The best way to figure out if that happens? Wait a few days, then watch for significant shifts in support that hold consistently across multiple different surveys.