|MSNBC's Steve Kornacki makes us want to believe. But should we?|
Why do we (meaning we in the media) continue to put so much stock in polls? As journalist Linda Ellerbee wrote in her excellent and inexplicably out-of-print memoir "And So It Goes: Adventures in Television" back in the 80s, "Ask yourself: If polling is so accurate, then why are there so many companies doing it?"
Two weeks ahead of Election Day, virtually every poll shows Biden beating Trump in terms of the popular vote. Meanwhile, the dreaded swing states (that is, the only places in this "democracy" where your vote seems to actually make a difference) continue to swing wildly. In other words, remember 2016? Let's not break out the bubbly quite yet.
Check out what Munr Kazmir writes in Medium's Dialogue & Discourse:
Polling is not reality. The polls showing Biden polling higher than Donald Trump, which are giving Democrats a dangerous dose of overconfidence, are often based on tiny sample sizes of 500 people. ... Things have changed since polling was a gold standard of accuracy, if indeed it ever really was one, which it wasn't. Polling is a representation of people willing to be polled."
Florida, as always, makes for a good study of the inherent flaws of divining the tastes of the American public, presidential or otherwise. As of this morning, FiveThirtyEight, an aggregation of hundreds of polls generating tens of thousands of potential outcomes, put Biden nearly 4 points ahead of Trump. Meanwhile, USA Today reports today that the numbers in the state are now favoring Trump. It's enough to make your head hurt.
Then, there's Robert Cahaly, chief polster of the Trafalgar Group, the only major polling organization that correctly predicted Trump's victories in Michigan and Pennsylvania in 2016. As Kyle Smith writes in National Review, Cahaly predicts this time around that Trump will also take Michigan, in addition to Florida, Georgia, North Carolina, Arizona and Texas. (This, despite the fact that a number of news orgs have already put Michigan, North Carolina and Arizona in the "safe" column for Biden.) Cahaly projects another ultimate win for Trump, with 280 electoral votes.
Pollsters reassure us that they properly identified the trouble with their crystal ball gazing last time around and have now "fixed" those problems. But what if they haven't? What if they're spectacularly wrong again? And the bigger question, again: Why do the media continue to worship at the altar of pollsters, people who are devoted to the patently absurd enterprise of predicting human behavior?
If the polls fail us yet again, you can bet one thing: It won't mean that one fewer player will be doing polling four years from now.