The Selzer poll is an exception to that rule. If it is correct, Trump is in a far better position than assumed, and we could be in for a much closer race than many expect.
Few, if any, of Biden’s pathways to 270 electoral votes run through Iowa. But if Biden isn’t doing much better in Iowa than Clinton did four years ago, then Democrats might be worried that he isn’t doing better in other places as well.
Indeed, it is possible that Trump wins Iowa by a wide margin that foretells trouble in other states such as critical Midwestern battlegrounds such as Michigan and Wisconsin.
But as I noted at the top, the average poll in Iowa paints the picture of a race that is too close to call. There are plenty of polls taken within the last few weeks that show Biden or Trump slightly ahead or slightly behind (Monmouth University, New York Times/Siena College and Quinnipiac University).
The reason why the Selzer poll gives Democrats chills and Republicans hope is the history of the Selzer poll. Four years ago, the final Selzer poll
had Trump up by an identical 7 points in Iowa. Trump won by 9 points.
Just like then, this 2020 poll comes in the midst of polling in a lot of other states that have Trump in trouble.
Moreover, the 2016 example is not the only time in
which Selzer was unafraid to publish a seeming outlier that proved to be correct. (The 2008 Democratic Iowa caucuses were another example.)
All that being said, there are many reasons to doubt that this Iowa poll is really telling us all that much. While Selzer is a fantastic pollster, every pollster has the occasional outlier. Selzer has been wrong before
. Pollsters aren’t magicians. It’s the sign of a good pollster that they are willing to publish polls that don’t look like other ones.
Speaking of outliers, an ABC News/Washington Post poll
in Wisconsin published in the last week had Biden ahead by 17 points. There isn’t a galaxy in this universe where Biden is up by 17 points in Wisconsin and down by 7 points in Iowa. But like the Selzer poll, the ABC News/Washington Post is at the outer ranges. The truth is almost certainly in-between. This is why averages work.
But it goes deeper than just outlier talk to understand why the Selzer poll this year may be different than it was four years ago.
This year we have a lot more state polls of high quality staying in the field until the last minute. I honestly cannot think of a single one that reflects what Selzer is seeing. Biden’s lead is larger in pretty much all the states that matter. The congressional district level presidential polling
, which was a warning sign for Democrats in 2016, is much better for them this time.
Perhaps most interesting is the national polling. Last time around, it was actually that national polling that led me to believe that Trump was just “a normal polling error
” away from winning. It ended up being just 1 or 2 points off the mark (i.e. very accurate).
This year Biden’s margin is about 5 to 6 points larger than Clinton’s was in the closing days nationally. To believe the Selzer poll, you probably have to believe the national polls are in for their worst error at least 40 years. That is, unless something very unique in Iowa is happening.
When you dig into the national polls, you see more reasons to believe Selzer is off. Iowa is a state with a lot of Whites without a college degree.
Do the national polls suggest that Trump’s holding his own among this key group? No. Instead, a comparison with the pre-election polls
from 2016 suggest his margin with them is lower by about 10 points.
If that’s right, the other pollsters in the Midwest (including Iowa) and not Selzer will end up being much closer to the final results.
Of course, that’s an “if.”
Another “if” is if Trump ends up pulling off a bigger shocker than in 2016, the Selzer poll will have been our best clue.
Before we bid adieu:
The theme song of the week is the old CNN election theme