Why Post-Debate Instant Polls Are Terrible

Online snap polls are one of those things that journalists hate, media companies love, and the public is largely unaware of.

Shaken after his loss in the first presidential debate to Hillary Clinton, Donald Trump and his campaign surrogates grasped onto the only positive news they could find--instant online “snap” polls that seemed to show Trump had “won”.

On Twitter, Trump quickly set the tone for his surrogates as they flooded the airwaves during the following day’s news cycle, using the post-debate night snap polls as proof that it was actually Trump who had won in the minds of the public, not Clinton. Not only did Trump win, they claimed, but these online polls proved an even more important point: the media are biased in favor of Clinton and have it out for Trump.

Despite the widespread consensus among reporters and political analysts that Clinton bested Trump in the debate, journalists were left to defend against snap poll wielding Trump surrogates who insisted otherwise. Two days after the debate, MSNBC’s Chuck Todd clashed with Trump campaign surrogate Jason Miller over this very issue, becoming visibly agitated at Miller’s insistence that post-debate online polls proved Trump had won:

MILLER: Well I have to set you straight on that one. The polls that happened that night, the night of the debate, the snap polls, the ones that happen online, those all showed Mr. Trump winning in a huge way. 
TODD:  What scientific poll had Donald Trump winning?  Give me one scientific poll. Those are fan polls, man. Those are what computer programmers can mess with. Those aren't real.

What are online snap polls?

Online snap polls are one of those things that journalists hate, media companies love, and the public is largely unaware of. The internet and social media is full of them ― snap polls are instant online surveys open to anyone to vote in.

Online snap polls are not the same as a real scientific poll, which must have a statistically significant sample size and most importantly, be based on random sampling (along with other controls like weighting for demographics). Scientific polling is expensive and time consuming.

Online snap polls are gimmicks designed for ‘user engagement’.

“Who Should Get Custody, Brad or Angelina?” “Will you buy the new iPhone 7?” ― Online snap polls are everywhere. News media sites use them frequently, not for their news value but as a clickable piece of interaction to keep users engaged.

Who doesn’t mind taking a second to click and vote in a poll?

When news outlets were trying to expand their web presence in the early 2000s, some started including online snap polls to encourage user engagement, or even just to showcase the newness and excitement of “interactive news” on the internet.

They might be fun, and normally online snap polls are little more than harmless clickbait, but when it comes to politics, they are a microcosm of the worst tendencies of election coverage.

Snap polls have little value in measuring real opinions.

This can’t be emphasized enough: online snap polls don’t measure anything useful, especially when it comes to politics. They are “opt-in” and basically just reflect whoever happened to stumble upon the poll. They can easily be “gamed” and voted in multiple times by refreshing a web browser, or even hacked and swamped with votes by a bot program.

Online snap polls might reflect the views of users of a particular site, such as the Drudge Report snap poll below taken after a 2016 Democratic primary debate, which says more about the rightwing, anti-Hillary mania of Drudge readers than anything meaningful about how the general public might have perceived the debate.

Online snap poll from Drudge Report
Online snap poll from Drudge Report

Post-debate snap polls are constantly gamed by activists.

In the 2008 Republican primary, Ron Paul won virtually every single post-debate online snap poll, sometimes by as high as 90%, but he never won a single state and rarely broke single digits when it came time to vote. Paul, the libertarian-leaning Republican Congressman had a small, passionate army of activists online, often dubbed “Paul Bots”, who made it a point to post links to post-debate snap polls and get all their comrades to flood the poll with votes. It got so bad that news organizations routinely had to take down their online polls because they were so out of sync with reality and were conflicting with their on-air assessments.

In 2016, we’ve seen the same phenomenon from online activists supporting Bernie Sanders and Trump, many of whom organize through message boards like Reddit or on Facebook, posting links to online snap polls for fellow activists to overwhelm with votes. After the first presidential debate between Trump and Clinton, Trump’s supporters on Reddit, 4chan, and Twitter led a coordinated campaign to bombard dozens of online snap polls to make it seem like Trump was the winner, using the hashtag #TrumpWon to spread their message throughout social media.

The media (and the public) are obsessed with instant results.

Actual scientific polling can take days to reflect the real effects of a debate or speech. In the world of non-stop breaking news and trending topics, online snap polls fuel our addiction to instant reactions, and many media sources are all too happy to feed that addiction; unfortunately the desire for ratings or clicks so often supersedes journalistic integrity.

Immediately following a presidential debate, there are usually a few attempts to have semi-scientific polling or surveys, like focus groups or post-debate sampling polling, but even their results come out slower than snap polls, they are fewer in number (because they cost money), tend to have very small sample sizes, and are ultimately no substitute for the real polling that is done over the course of the following few days.

Unscientific polls feed the trolls.

Thanks to the hyper-fast rumor mill of social media, the “results” of online polls are widely circulated as if they’re truth. Many people, understandably, don’t know the difference between an unscientific snap poll and a real poll. They provide ammunition for partisans to push their agenda online, and fuel conspiracy theories over widespread media bias or unfairness.

Like Ron Paul’s supporters in 2008, many Bernie Sanders supporters used his domination of online snap polls to argue that there was a deep media bias against Sanders, and that “viewers” all seemed to be claiming that Sanders won. Except, snap polls don’t actually measure anything, except whose supporters are more willing to spend time trying to game online polls all night after a debate.

They’re just confusing.

Most people don’t know the difference between scientific and non-scientific polls--and they shouldn’t be expected to know. Regular polling is reported on so often by the news media that the very word “poll” evokes a sense of new data, of useful information. Even if news organizations clearly spell out that these polls are non-scientific, their very existence will still be used to spread misinformation. It’s the job of the news media to avoid misleading viewers about the validity of information. When using the term “poll”, news anchors should explicitly clarify the difference between scientific and non-scientific polling—but so often they don’t.

Snap polls are a distraction from what matters.

When it comes to election coverage, the media loves focusing on the “horse race”; stories about who's up and who’s down make for bigger ratings and more clicks than bland analyses of political policy.

Snap polls feed into the frenzy, providing instant “data” (of dubious validity) to help frame the horse race narrative--who did better in the debate, who won and who lost. Instead of helping viewers understand complex policy ideas and disagreements, we get misleading and useless information in a form of what looks and feels like online-democracy in action.

At their worst, online polls undermine faith in our democracy.

Imagine a Trump supporter who is getting all his or her news from pro-Trump media: Fox News, Breitbart, Trump’s Twitter, pro-Trump facebook groups, etc. Immersed in that world, the “results” of online snap polls help cement a false narrative where Trump is actually winning the election, and the media are trying to stop him by telling the public the opposite.

If Trump loses, many of his supporters will falsely believe that the system is rigged against him--a claim Trump has been making with more vigor in recent days. In their world, the results from online snap polls that showed a unanimous Trump victory are further proof of the delusion of a grand conspiracy against Trump and his movement.

Ultimately, instant online polls are a needless gimmick that have been abused and misused so often to the point that their use undermines a sense of fairness in our democratic system. Media organizations should stop creating and them and broadcasting their results, especially after political debates. Alternatively, like cigarettes or unpasteurized milk, online snap polls should come labeled with a very clear warning: