The Polls Were Wrong

And other observations following Super Tuesday.

Graphic courtesy of NBC News

Super Tuesday is over, and the polls were wrong. Again.

Going into Tuesday, Trump was expected to romp to victory across over a dozen contests. The polling averages, according to poll aggregator 538, had Donald Trump beating Nikki Haley nationally by a whopping 62 points.

So how did Trump actually do, compared to expectation? Not very well. He underperformed the polls in nearly every contest. And in many cases, that underperformance was in the double digit range. The pollsters got voter sentiment so badly wrong in Vermont that Trump actually lost the state to Haley, where he was supposed to win by around 30 points.

Moreover, as we saw in earlier primary contests, when interviewed in exit polls, Haley voters indicated a strong dislike for Trump, with many indicating they would not support him in the general election. And while Haley bowed out of the race today, she did not endorse Trump, at least not yet, saying instead that Trump must earn back the support of her voters.

That’s not very likely to happen, given that Trump is now demanding that the GOP be 100 percent MAGA and that all “Romney Republicans” be ousted from the party.

Joe Biden performed well on Tuesday, outpacing Obama’s support as an incumbent in 2012. So today, let’s take a closer look at the GOP results specifically and what some pollsters are now admitting about those poll misses. Then we’ll look at the exit interview results and see why they indicate continuing vulnerabilities for Trump. Finally, I’ll leave you with some thoughts on what is now a two-person race for the White House.

Quick pledge break! If you’ve been meaning to support my work with an upgrade to your free account, become a paid subscriber today!

Trump underperformance

Following each of the caucuses and primaries held before Super Tuesday, election numbers folks took a look at Trump’s actual performance versus what the polls had anticipated. As Professor Justin Wolfers of the University of Michigan reported, Trump had underperformed the polls in Iowa by 23 points, New Hampshire by 7, South Carolina by 8 and Michigan by a whopping 20.

This already indicated that something was pretty off when it came to measuring Trump’s popularity at the state level. And on Tuesday, that trend continued.

In Virginia, according to the 538 polling averages, Trump was supposed to beat Haley by 49 points.  Instead he won by a little over 28 points. That’s a huge polling miss of 21 points.

In Maine, Trump was supposed to win by 58 points, according to the latest poll from the University of New Hampshire Survey Center. Instead he won by 47. Again, a double digit miss in the polls. 

Professor Wolfers summarized the underperformance as follows:

By my count Trump's actual margin in the primaries has underperformed that predicted by the polls by: 

0-5%: AL, IA, TX

6-10%: CA, ME, NH, SC

10-15%: MA, MI, OK, TN, UT

16-20%: - 

20% or more: MN, VA, VT (an astonishing 34%)

The only state where Trump outperformed pre-primary polling was North Carolina (and we’ll get to what happened there, hoo boy).

So what gives?

Primary polling is notoriously off, but when it is consistently off in the same direction by massive amounts, something is wrong. Even polling aggregator 538 seemed to acknowledge this. Around midnight last night, G. Elliott Morris of 538 wrote up his thoughts on primary polling, and this caught my attention:

POTUS primary polls are by far the least accurate type of poll in 538’s pollster rating database going back to 1999. But error is different than bias, and when most of the polls are off in the same direction, something has gone awry.

That “something” is likely the difficulty in obtaining opinions of moderate Republicans from samples of “likely Republican primary voters.” Remember that fewer than one percent of people called for a poll actually complete the interview. That means the ones that do are statistical “weirdos” (excuse the technical language). Pollsters adjust for this by weighting their samples to known population benchmarks — like the percent of all adults who are white, over 65, have a college education etc. But in primaries, such benchmarks do not actually exist; pollsters are just making educated guesses about them.

My theory is that most of these primary polls pulling samples of voters from voter registration lists are missing moderate crossover partisans and first-time voters. Additionally, we know that people who are highly motivated to participate in polls (the “weirdos”) also happen to be the most politically and ideologically extreme Americans. That's a recipe for polling bias in primaries, where weighting to party, past vote and polarized demographic benchmarks does not control for the partisan consequences of overrepresenting politically engaged Americans.

So, to sum up, the pollsters are using the 1% “weirdos” who bother to complete polls, whom they know are political extremists, to create datasets and publish polls that are way, way off, creating totally false narratives and expectations.

Pollsters can’t be too openly critical of their own industry because that would put them out of a job. But these polls are so useless and wrong that the public is actually worse off from being told about them at all.

If the national polls are overestimating Trump’s strength at anywhere near the levels that the primary polls did, then Biden would be leading Trump in all of them.

Exit polling 

Exit polls are generally more accurate than pre-election polling. After all, pollsters are asking actual voters how they actually voted and ascertaining what they actually think.

And once again for Trump, the exit polling from Super Tuesday contains some real danger signs. Fox News’s Jessica Tarlov, one of the lone liberal voices on the network, put it bluntly. “You look at the exits for GOP voters tonight who won't guarantee their vote for the nominee. North Carolina, 35%, Virginia 36%, California 33%. That has been consistent, there are 30-50% of people who don't want Donald Trump but identify as Republican or are voting in that primary.”

In Virginia, while Trump dominated among Conservative voters 73 to 26, Haley walloped him among moderate voters 68 to 30, according to NBC exit polling. As we’ve seen in other contests, one of the biggest divides in the GOP electorate is over the question of whether Biden won the 2020 election legitimately. Trump voters overwhelmingly said “no” (76 to 15) but Haley voters were an even more emphatic “yes” (87 to 4). 

Virginia is an open primary, which means independents and even Democrats could vote in it if they chose to. But Trump can’t win the general election on conservative voters alone, nor does his election denialism sit well with the moderates who back Haley. Those are the very kinds of voters who determine election outcomes.

In North Carolina, which is a partial primary that allows “unaffiliated” voters to cast ballots, there were more warning signs for Trump. As the Times reported, citing NBC data:

A majority of Ms. Haley’s primary voters said they were voting against her opponent more than for her, a sign of anti-Trump motivation that could last until November. And even in defeat, she was leading among moderate voters by nearly two to one. Her problem was that moderates make up only 20 percent of the voters in a G.O.P. primary. But in a close general election, those voters may matter more.

Overall, roughly one in four Republican primary voters in North Carolina said they would feel dissatisfied if Mr. Trump won the nomination.

“In state after state, there remains a large bloc of Republican primary voters who are expressing deep concerns about Donald Trump,” Olivia Perez-Cubas, a spokeswoman for the Haley campaign, said late Tuesday.

In North Carolina, 81 percent of Haley voters told CNN’s exit pollsters that they’re not an automatic vote for the GOP nominee, who of course is now Donald Trump. Four out of five is a very significant number of possible defectors or stay-homers among Republicans in the state.

One other observation from last night: North Carolina is a potential pick-up for Democrats, particularly as the state attracts more white-collar, educated professionals to the Research Triangle area. The likelihood of a flip to blue went up yesterday after the GOP’s most extreme candidate, Mark Robinson, the party’s first Black nominee for governor, won that state’s GOP gubernatorial primary.

Robinson is likely to inspire a significant protest vote in much the same way that extremist Doug Mastriano turned off voters in Pennsylvania in 2022. As Democratic strategist Sawyer Hackett observed,

Republicans in North Carolina just nominated a man for governor who:

-is a literal Holocaust denier

-said gays are “the end of civilization”

-called school shooting victims “prosti-tots”

-said shootings are “karma” for abortion

-called civil rights movement a “communist plot”

Trump has praised Robinson as “Martin Luther King on steroids.” Voters are more likely to view him as yet another conspiracy-driven, dangerous Trump acolyte. 

Now it’s down to two

Now that the race really is down to Trump and Biden for all practical purposes, there’s a good portion of the country, one which has been in wilful denial of this rematch, that has yet to start to pay attention. It may be hard to believe, but many Americans simply are not yet tuned in to the election yet. This is particularly true among lower income and lower propensity minority voters that make up a good chunk of the Democratic base.

Biden’s job will be to bring those voters back into the party fold by highlighting his many accomplishments while warning of the dangers of Trump’s authoritarianism. That task kicks off officially tomorrow with his much-anticipated State of the Union address. 

Trump’s task will be to unite his party behind him, even after telling moderate Republicans that he doesn’t want any Romney RINOs, just MAGA faithful, in the party—all while navigating a criminal trial in New York that starts later this month and possibly another in D.C. in the summer.

Based on what the exit polls are telling us, the fundamentals on the economy, and the basic leadership styles and practices of the two candidates, I’d much rather be us than them.

Later this month, I will be asking my readers to contribute to the Biden Victory Fund, of which I am a committee member. It will be my “birthday” ask on the 28th of March. I’m flagging it early because I want folks here to think about how much they can afford to give. It should be enough that it hurts a little. This is an election where none of us can afford to sit on the sidelines. I hope you join me in helping ensure Joe Biden and Kamala Harris are returned to the White House next year.

Beyond our votes, we can all take some part and some responsibility in keeping our Republic and our cherished democratic freedoms.

Jay