Happy Election Day!
A roundup of articles to help you follow the returns.
It’s FINALLY here — Election Day 2024!
I hope you’ll excuse me if I sound relieved; my colleagues and I have spent literally years preparing for and writing about this day. And as I write this at 1am, mere hours before polls open, I find that I’m … pretty much out of things to say? (Lucky you.) So I’m going to make this quick: Here’s a list of resources I’m going to use to follow along with the election tonight.
538’s forecasts for the presidency, Senate, and House, which will give you a good sense for which party is favored to win which state/district—so you’ll know the upsets when and where they happen.
538’s guide to when to expect election results in every state. I’m really proud of this one—I reached out to election officials in all 50 states, and we got data on how quickly votes were reported in the 2022 election in 48 out of 50. TL;DR: In all likelihood, we will not know the winner of the election on Tuesday night.
My colleague Geoffrey Skelley calculated vote-goal benchmarks for each candidate in every county in every state. It’s a super useful way to track who’s on pace to prevail in a given state when we only have results from a few counties.
Federal races aren’t the only thing on the ballot. 538’s Cooper Burton wrote about the 10 state-legislative chambers that could flip control today. I also tried and failed to pick the most interesting ballot measures from the 146 that are on the ballot today (I got it down to a mere 49!). I also recommend Bolts’s uber-comprehensive guide to local races that are flying under the radar.
OK, I lied when I said I was out of things to say. I’ll leave you with a seven very top-level things to keep in mind before we all venture into the great unknown.
According to the consensus of the data we have, Harris and Trump have roughly equal chances to win. Anyone who tells you with certainty that they know who will win is not making a data-based prediction.
“What the polls say” can vary depending on which polls you look at, but any comprehensive aggregation/average of them points to extremely tight races in the seven main swing states (Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania, and Wisconsin).
That said, it’s totally normal and expected for polls to be off by 3-5 points. As a result, a comfortable win for either Harris or Trump wouldn’t be a huge surprise.
This polling error is just as likely to benefit Harris as to benefit Trump. Some people think you can guess the direction of polling error in advance. Half of those people will be wrong; the other half will be lucky.
If polls significantly underestimate Harris (say, by more than 3 points), it won’t be because of GOP pollsters “flooding the zone.” They account for less than 10% of presidential polls this year (meanwhile, 85% of polls are nonpartisan) and have moved 538’s averages by, at most, 1 point.
If polls significantly underestimate Trump, there will be legitimate questions about whether polls can adequately reach Trump voters. But as long as the error is within the normal 3- to 5-point range, it won’t mean polling is “broken.”
Forecasts like 538’s aren’t trying to “call” a winner. There is no significant difference between Harris having a 51% chance and Trump having a 51% chance. We aren’t “wrong” if the underdog wins—and we won’t claim to be “right” if the favorite wins. (You can hold us to that!)
This is not to say that 538 or pollsters shouldn’t be held accountable for bad models or polls! But for forecasts, the best way to grade us is by what percentage of the time candidates with an X% chance actually win. And for polls, just remember that polls have margins of error for a reason and some amount of error (less than ~5 points) is acceptable—even expected.