Lori Falce: The virtual voodoo of political polls
So when do we decide polling is just the 21st-century version of witchcraft?
Regardless of what happens with the ballots and counts and final tallies of the 2020 general election, one thing is clear. The way it ended up doesn’t look that much like what the predictions were.
There were talks about blowouts. There were talks about landslides. There were talks about the sentiment carrying down the ballot and changing things up with the legislatures.
And when the big day came? It was more a purple fog than a blue tsunami or a red wave.
I have friends who are enjoying this chasm with a health portion of sour grapes — and others who are treating it like they were promised a pony and given a possum. A dead one. With fleas.
The reality, as with most things, is not so simple or so partisan.
One of the most prominent pollsters is Nate Silver of FiveThirtyEight, who took a real beating from some on the left in the wake of the 2016 election, which had been predicted to go to former Secretary of State Hilary Rodham Clinton.
Spoiler alert: It didn’t. Donald Trump powered past the polls and while Clinton took the popular vote, his were spread across the right states to pick up the Electoral College.
That couldn’t happen again, some thought as the 2020 election loomed. Silver’s predictions for 2020 were once again heavily blue. FiveThirtyEight claimed that after 40,000 simulations, 89 out of every 100 went for former Vice President Joe Biden to some degree while 10 out of every 100 broke for Trump.
The votes aren’t all counted. The decision isn’t final. But while it does look like the vote has tipped toward Biden at the moment, political commentators and pundits are shaking their heads about the failure of the polling.
But did it fail? Or is everyone asking for an estimate for a construction job and being shocked when they get a bill that includes overages for sinkholes and a bad foundation?
Because Silver didn’t fail. He gave a spectrum of possibilities that tended toward a certain outcome — an outcome, incidentally, that seems to be falling squarely in the middle of that spectrum. Out of that range of possible ends, after all, one was a 328 electoral vote victory for Trump. That is a mathematical impossibility at this point, but it was still on the table.
Polls are the only time we act like an opportunity is a promise. It isn’t. The PowerBall jackpot odds of winning are one in about 292 million. People buy the ticket knowing that it isn’t a contract. It’s a chance.
It’s ridiculous to ask someone to use a crystal ball — even one based in science and statistics — and then dismiss it for not being accurate enough. Polls shouldn’t be relied upon as much as they are, and we need to appreciate pollsters for what they bring to the conversation when it comes to elections, but it shouldn’t be overstated.
And we definitely shouldn’t burn Silver and other pollsters at the stake because the magic turned out to be math.
Lori Falce is the Tribune-Review community engagement editor and an opinion columnist. For more than 30 years, she has covered Pennsylvania politics, Penn State, crime and communities. She joined the Trib in 2018. She can be reached at lfalce@triblive.com.
Remove the ads from your TribLIVE reading experience but still support the journalists who create the content with TribLIVE Ad-Free.