Even good polls can give Democrats a pit in our stomachs. Looming in our heads at all times are the polling errors from 2016 and 2020 — especially 2020. Last time around, Biden was up by about 8 points in the national polling average and held comfortable leads in the swing states. Biden was so confident that he ended his campaign with stops in Iowa and Ohio, both of which Trump won easily four years prior.

Democrats had every reason to feel confident — even cocky — heading into Election Day. The polls suggested it would be an early night and the champagne bottles sitting on ice since Election Night 2016 could finally be popped. But alas, the election was much closer than the polls suggested. Biden barely squeaked by in an election that wasn’t called for days.

Just like in 2016, the polls dramatically undercounted Trump supporters. Democrats live in fear that something similar will happen this time. The race is basically tied in the battleground states so, if there is even a minor anti-Trump bias in the polls, Trump will be the next President of the United States.

It’s not just Democratic worrywarts. Most people in politics assume the polls will be wrong and will under-represent Trump’s share of the vote. This is why Republicans are so confident, Democrats are so anxious, and the not-so-subtle tone of the coverage suggests that many in the political press expect him to win.

Are people right to expect a polling error and would such an error necessarily benefit Trump?

Well, it’s complicated.

1. What’s a Polling Error?

The national polling average has Harris up one, but if Trump wins by one that doesn’t mean the polls were wrong. Polls are inherently imprecise tools. They have a margin of error — usually about three points in either direction. So, in the above example of a poll where Harris is +1, Trump winning by two or Harris winning by four would be in the range of outcomes.

Polls are informed by the pollster and from the composition of the electorate. Pollsters take the raw data and apply weights to make the data fit with their idea of what the electorate will look like. Will more Democrats or Republicans turn out? Will Black or Latino voters show up at the same levels as past elections? How many young people will vote? (This article is a great explainer for how the process works.)...