The polls say Biden will very likely win. Why spend your time and effort voting?
Just a reminder: these Rationalist Papers posts are for the group I call the deciders: conservative, moderate, undecided, and third-party voters considering their choices in the 2020 US Presidential election.
The horse race
As they do every year, news organizations devote a great deal of time to “the horse race” — their analysis of how the candidates are doing relative to each other. As we close on the last three weeks of the campaign, there’s a lot of information available on that front, including hundreds of national and state polls.
The most advanced analysis of the polls comes from Nate Silver’s fivethirtyeight.com, which includes a model based on everything objective: polls, past voting tendencies, economic conditions, and the structure of the electoral college. Every day, they runs tens of thousands of simulations of the election, accounting for random variations beyond what’s in the polls, and publish the resulting projection of the likelihood of the candidates winning. Here’s what that looks like this morning:
So today, 20 days before the election, Nate Silver’s model thinks that there are 87 chances in 100 that Biden will win.
Here are the chances his model assigns to Biden winning in key swing states (the ones in italic, Trump won in 2016):
- Minnesota: 91%
- Michigan: 91%
- Wisconsin: 87%
- Pennsylvania: 88%
- Nevada: 86%
- New Hampshire: 85%
- Florida: 73%
- Arizona: 67%
- North Carolina: 66%
- Ohio: 51%
- Iowa: 46%
- Georgia: 45%
- Texas: 32%
Given these probabilities, the model estimates the most likely outcome is Biden beating Trump in the Electoral College by 346 to 192, and leading in the popular vote 53.5% to 45.2%.
It’s not just Fivethirtyeight. The sophisticated model at The Economist is even more extreme, predicting Biden’s chances of winning the election at 92%.
If you accept these projections, you could easily conclude the election is effectively over and we know who the winner will be.
Why the election is still in doubt — and your vote matters
The election models in 2016 also predicted that Trump would likely lose. Fivethirtyeight gave him a 29% shot on Election Day, while The New York Times rated his chances at 15%.
Polls have uncertainty. They can also miss trends that develop later. People who confused “29% chance” and “no chance” in 2016 were surprised when Trump won. But seven out of ten is not a sure thing. If there was a 29% chance of rain, would you bring an umbrella? If there was a 29% chance of getting run over, would you cross the street?
This year, things look significantly more skewed towards the Democratic candidate. People have seen four years of Trump, and they are telling pollsters that most of them don’t like what they’ve seen. But the election, like all elections, is still uncertain. Here are some possible reasons that Trump could win:
- Polling error. Polls measure the opinions of thousands of people, not millions. All polls have sampling error — if they don’t reach enough Trump voters, they’ll miss some of his popularity.
- Polling bias. The fivethirtyeight polling analysis corrects for “house effects” — polls that consistently predict outcomes too optimistically for one party or the other. But there are biases you can’t correct for. If Trump voters are shy about revealing their preferences — for example, because they don’t want to say they’re voting for a losing candidate — they won’t respond. The poll will miss them and fail to register their preferences.
- Good events for Trump. In the next 20 days, we could see a good debate performance, the announcement of a vaccine, good jobs numbers, the passage of an economic bailout package, the confirmation of a Supreme Court Justice, the arrest of protestors accused of killing police, or some other favorable outcome for Trump that we haven’t foreseen. Such an event could win back some of the Trump voters that have deserted him.
- Bad events for Biden. Biden could appear uncertain in the debate. He might make a gaffe, like confusing COVID-19 for Ebola. Or people could just decide his strategy of conducting fewer campaign events is not a good sign. In 2016, James Comey’s announcement of information about Hillary Clinton’s emails on Anthony Weiner’s computer may have cost her some votes, and it happened so close to the election that polls couldn’t measure it.
- Advertising. Both candidates are advertising heavily. Ads influence some votes (or else candidates wouldn’t pay for them). A devastating ad could shift the election in key states.
- An “October surprise.” The term “October surprise” refers to the release of damaging information about a candidate close to the election. If Trump’s allies have information about Biden doing something terrible in his past, they could release it and tear away some voters.
- Legal maneuvering. The Trump campaign is gearing up for legal fights, which are likely to focus on invalidating mail-in ballots as potentially fraudulent. (Democrats are more likely than Republicans to vote by mail.) If the result are in doubt in a given state and time is running out, the state legislature will decide what electors will vote in the electoral college — and could deliver the state to Trump. In Pennsylvania, for example, the legislature has a majority of Republicans.
Still think the election is in the bag for Biden?
If you are not registered to vote,
it’s it may be too late to do so. But if you are, you should still vote in this election.
Updated info: Check here for information on your state’s registration deadlines, which may include same-day registration.
Your vote could determine the makeup of the Senate, which is very much in doubt.
Your vote could deliver a landslide for Biden (or prevent one) which would make legal challenges and maneuvering much less likely to change election results.
If enough states are clearly in the Biden column, the legal challenges in disputed swing states won’t matter. This will make it far harder for Trump to challenge the fairness of the election — especially if those states are clearly in the Biden column based on election night results. So voting for Biden might make the results far clearer, far earlier.
Your vote could be the one that makes the difference in a state like Pennsylvania, Wisconsin, or Texas.
Your vote could determine the makeup of your state legislature, which will draw voting maps for congressional districts in 2021. And if your state is in dispute, your legislators will determine which candidate’s electors represent your state in the Electoral College in the 2020 presidential election.
Despite the long lines that are appearing in some early voting states, there is still plenty of time for early voting. There is time to mail in your ballot, if you requested one. And you can still vote on election day.
Forget the projections. Your vote matters. And if your candidate loses in November and you didn’t vote, you’ll have no one to blame but yourself.
3 responses to “The Rationalist Papers (17): Why bother voting?”
Dewey Defeats Truman!
If you’re allowed to vote in the US Presidential race, do so. It’s your right (provided the Republican Party hasn’t yet completely disenfranchised you), and this country can’t stand another four years of the incompetence and lack of compassion of the current president.
NOT too late to register! States have different deadlines for registration.
If you are not registered to vote, check with your state’s secretary of state or similar — you may even be able to register and vote on the same day!
Verify your state’s information immediately.
Don’t give up until you know!
All solid points.
There’s a psychic benefit of voting. I have never felt so satisfied expressing my outrage over the past four years. A landslide will hopefully send a message and stand up to the man-baby’s invariably whining.