By Dave Andrusko
For fun, let’s compare two stories and see what they tell us about the 2024 presidential race. The headline for an article in The Washington Times reads “Beware the polls: They’re all wrong: But just like weather forecasters, they stick around” while The Atlantic offers a less absolutist conclusion, “The Asterisk on Kamala Harris’s Poll Numbers: Pollsters think they’ve learned from their mistakes in 2020. Of course, they thought that last time too.”
I would encourage you to read both stories.
Joseph Curl, writing in The Washington Times, starts off asking how a survey from Fairleigh Dickinson University could find pro-abortion Vice President Kamala Harris up nationally by a whopping seven points (50 to 43)?! It’s to make the point that some surveys—like this one—are just wrong.
Even if Harris is up just 2 or so points nationally, Democratic Strategist James Carville warns that her lead in not real. “Traditionally, with Trump on the ballot … Democrats say, ‘Oh, James, you’re a Debbie Downer.’ I’m not. I’m just telling you, you got to win by three.”
That’s because we have an electoral college system which means a candidate can lose the popular vote but carry the 270 electoral votes needed to win.
Gilad Edelman, writing for The Atlantic, offers the many, many reasons the polls were off (in 2016 but especially in 2020) and why all the corrections that have been made since 2020 are not a guarantee of accuracy. It’s a long article but for a myriad of highly technical reasons (and not so technical—Trump voters often hide their vote and/or get angry when called by a pollster), it is exceeding difficult to get a representative sample. Edelman writes
Don Levy, the director of the Siena College Research Institute, which conducts polls on behalf of The New York Times, calls the phenomenon “anti-establishment response bias.” The more someone distrusts mainstream institutions, including the media and pollsters, the more likely they are to vote for Trump.
He concludes
Like Olympic athletes, political pollsters spend four years fine-tuning their craft, but don’t find out whether their preparations were adequate until it’s too late to do anything differently. The nonresponse bias that bedeviled the polls in 2020 is not an easy thing to fix. By definition, pollsters know very little about the people who don’t talk to them. If Trump outperforms the polls once again, it will be because even after all these years, something about his supporters remains a mystery.
As I was about to post this article, I ran into this written by the highly respected Nate Silver: “Pennsylvania may be a problem for Harris: The Electoral College/popular vote gap is increasing in post-DNC data.”
After writing about the national implications (which he thinks will level out), Silver says of this battleground state
There’s another, longer-term concern for Harris, though: it’s been a while since we’ve seen a poll showing her ahead in Pennsylvania, which is the tipping-point state more than a third of time in our model. Today, in fact, we added one post-DNC poll showing Pennsylvania as a tie, and another (conducted during the DNC) showing either a tie or Trump +1, depending on what version you prefer.
The model puts a lot of weight on this recent data because of all the changes in the race. And you can see why it thinks this is a problem for her: if she’s only tied in Pennsylvania now, during what should be one of her stronger polling periods, that implies being a slight underdog in November.
See you tomorrow.