With increasing frequency in recent weeks, friends and family members have been coming to me for reassurance, asking me with trepidation in their voices, “Donald Trump's not going to win, right?” I understand why they fear a Trump presidency, because I share those same fears myself.
But a Trump victory, as catastrophic as it would be, remains very unlikely, and for a simple reason: The polls strongly favor Hillary Clinton.
As of Friday, the Daily Kos Elections forecast gives Clinton a 91 percent chance of winning, and indeed, her odds have been in the 90s for an entire month straight. Our forecast is based on a mathematical model that takes in all of the state-level polls—literally thousands of them—and uses them to simulate the election outcome. There’s no magic or wizardry involved, just some very straightforward math. And that math keeps saying Clinton is very likely to win.
I trust the math because the man behind our model, data scientist Drew Linzer, had the most accurate forecast of the 2012 elections, predicting Barack Obama’s exact electoral college margin. Then he did it again, producing the most accurate forecast of the 2014 Senate elections. Drew, in other words, got it right in a good year for Democrats, and he got it right in an awful year for Democrats. I have every reason to believe Drew’s model is going to be right again.
And here's the critical thing: In 2012, most polls underestimated Democratic performance, while in 2014, they overestimated it. But despite those systemic failings, Drew's models still got the outcomes right, because those models incorporate as much polling data as possible. And when you aggregate polling data, you’re much less likely to be misled by outliers—polls that stick out from the rest like proverbial sore thumbs.
But pundits love those outliers, because they quickly tire of telling the same story. And when those outliers point toward a close contest, pundits love them even more, because they crave a competitive contest, whether one exists or not. After all, that’s what sells papers and drives clicks.
I’m here to tell you, though, that the race is a lot more staid than most reporting would have it. And we aren't the only ones saying so: Nearly every other mainstream forecaster agrees that Clinton is the strong favorite, and even the most pessimistic analyst still thinks she’s winning. This doesn’t mean that Trump can't win, because there are no guarantees in politics. But it does mean he very probably won’t.
I say all this to calm some jitters, to push back against bad punditry, and help spread the gospel of poll aggregation. But I do not say any of this to encourage complacency—absolutely not. If you haven’t voted yet, go and vote now if you can, or on Election Day if you can’t vote early, and make sure your friends do the same. If you're planning to volunteer, get out there right away and help the Democrats running in your area, or phonebank for those elsewhere. Whatever you do, don’t let up until the polls close Tuesday—but don’t panic, either.