Before you share that poll: five things to look for

Every election season, our group chats fill up with charts: who's ahead, by how much, "according to a new poll". Some are real surveys. Many are screenshots with no source at all. During the last elections, I started asking the same five questions every time one landed on my phone.

The five questions

1. Who ran the poll, and who paid for it? A real poll has a name behind it. A chart with no logo and no author is not a poll.
2. How many people were asked, and how were they chosen? A few hundred people chosen at random can tell you more than thousands of clicks on an online form that anyone can fill in, as many times as they like.
3. When was it done? Opinions move. An old poll shared as "new" is misleading.
4. What's the margin of error? If two candidates are within it, the poll can't reliably tell you who's ahead.
5. Is it a poll or a prediction? Many "results" that circulate before the count are guesses dressed up as numbers.

A poll is a snapshot of a moment, not a result.

On election day itself, the only results that count are the official ones. Anything before that is, at best, an estimate.

The Pew Research Center's short explainer on the margin of error, linked below as the source, is the clearest one I've found.