Last week, news broke from Axios that Joe Biden, like his opponent Donald Trump, does not believe the polls except for the ones that show him winning. This week, Isaac Chotiner of The New Yorker interviewed presidential campaign strategist Simon Rosenberg about his view of the polls and the race in general, and hoo boy does it reinforce the narrative that at least parts of Bidenworld are wholly insulated from objective reality.
Before getting into yet another interview where Isaac Chotiner asks some pretty straightforward questions that force his interviewee to self-immolate—and this one is an all-timer, you really need to read the whole thing—we need to back Chotiner up and establish some well-known ground rules.
Most polling indicates that Joe Biden is losing to Donald Trump right now. Not by a lot, mostly still within the margin of error, but both high quality and low-quality polling reveal a race where Donald Trump is slightly ahead.
But not if you ask Simon Rosenberg, Democratic strategist and author of, I kid you not, “Hopium Chronicles.” He actually brings up a good point before beclowning himself, and it’s that partisan polling outlets have kind of ruined aggregate averages.
In 2022, there was an effort—and this has been documented again and again—by Republicans to flood the polling averages with bad polling, to push the polling averages to the right, which was then successful.
If you just read this passage, you’d think this was a reasonable interview and I’m being the jerk here. Well, here’s Chotiner’s next question in bold with Rosenberg’s responses in regular text.
As FiveThirtyEight makes clear in their piece, “While the polls in a few closely watched races—like Arizona’s governorship and Pennsylvania’s Senate seat—were biased toward Republicans, the polls overall still had a bit of a bias toward Democrats. That’s because generic-ballot polls, the most common type of poll last cycle, had a weighted-average bias of D+1.9, and polls of several less closely watched races, like the governorships in Ohio and Florida, also skewed toward Democrats.”I’m ending the interview. I’m ending the interview because what you’re doing is ridiculous.
Wait, wait—why?
Because I have definitive proof that what you’re saying is not true. And I don’t care. I know what FiveThirtyEight wrote. I live this every day. And so, the point is what you’re saying is wrong. I am on record saying that what FiveThirtyEight has written is incorrect, and I’ve given you definitive proof otherwise. So if you want to keep coming back at this, do it. But this has become one of the most ridiculous interviews that I’ve ever done my entire professional career.
This meltdown occurs at the beginning of the interview. If the goal here was to paint a picture of a calm Biden reelection campaign with a plan to change the narrative around what is empirically the least popular incumbent in modern history, Rosenberg’s “definitive proof” did the opposite.
The sensitivity on display here is difficult not to read into. I have no knowledge of how influential Rosenberg is in the Biden circle outside of him holding the same beliefs Biden has, per last week’s Axios report, but if this attitude is representative of the broader Biden camp, these people are cooked.
His chief contention is that polling missed Democratic enthusiasm in 2022, and the RealClearPolitics Senate average predicted 53 seats for the Republicans when they actually won 50 seats. This is true, but polling also missed Republican enthusiasm in 2016 and 2020 in predicting less support for Trump than he wound up getting. Polling is not exact and when it’s wrong it has been wrong all across the ideological spectrum. The reason polling is getting less accurate in general is because the local news outlets that used to conduct them are all dying or becoming Gannett-fied zombies.
Rosenberg contends that the Supreme Court’s overturning of Roe changed everything, and this is a reasonable position to take. The surge in Democratic voting in 2022 definitely was influenced by this decision, as women’s rights are unquestionably on the ballot in every race with a Republican in it, and Gen Z and Millennials voted at higher rates than previous generations did at that same age.
That said, Pew found that “the GOP improved its performance in 2022 across most voting subgroups relative to 2018.” If you only want to limit your analysis to the last election, then you must assume improved Republican performance along with it. Anyone asserting a sample size of one is enough evidence is not a serious person, let alone a wannabe data scientist. Which brings me back to Rosenberg, who said that “any comparisons to 2016 and 2020..are not valid.”
If I had told my political science professors in college that it was wrong to analyze a presidential election next to the two most recent presidential elections, and in fact one midterm election was the better comparison, they would have failed me. This is some shambolic jackassery that makes it look like Rosenberg spends most of his days with his head firmly buried in the sand.
Midterm electorates are very different from general electorates. They have way less voters who on the whole are more educated and engaged. This creates all sorts of different incentives that are favoring Democrats in the age of Trump and their growing suburbanized coalition.
It’s one thing to say that data from 2022 is more important because of the new political dynamics unleashed by SCOTUS’s Dobbs decision. It’s quite another to say it’s the only thing that matters. Turnout in 2022 was actually lower than turnout in the 2018 midterms and the GOP made inroads with key Democratic voters. It’s not wise for Rosenberg to travel down this path he’s laid out for himself.
Here’s some good analysis from two guys who actually understand this stuff.
The rhetoric around Trump’s primary performance after Haley dropped out is frankly out-of-whack. He’s doing historically well for a non-incumbent. (Best in the 2000s)
Biden, meanwhile, is doing the worst in his primary for an incumbent president in the 2000s. pic.twitter.com/4q2vPqm3bA
— (((Harry Enten))) (@ForecasterEnten) May 23, 2024
The clearest, most dramatic snap shot of the 2024 “engagement gap” I’ve seen yet.
Since 2020, Biden’s lost ~2 points among ’20 voters and ~20 net points among registered voters who didn’t cast a ballot in ’20. Just a massive 4-year survey swing among less likely voters. pic.twitter.com/s5Qd9Zdgq5
— Derek Thompson (@DKThomp) May 24, 2024
This isn’t the only area where a Dem strategist tilts the playing field in favor of the Dems and then claims an inherent advantage, as Chotiner challenged Rosenberg on so-called zombie voters, and he pointed out that The New York Times “found that nineteen percent of Republican primary voters were still voting against Trump. But that’s a smaller percentage than in any other Republican or Democratic primary in the past twenty-four years.”
Rosenberg’s response?
It doesn’t matter. It doesn’t matter. None of that matters.
Rachel Cohen of Vox noted that she “published a piece that took several weeks of work interviewing (mostly female) pollsters about why abortion could be less salient in 2024. Simon Rosenberg responded by email that ‘there simply is no reason to expect’ November will be any different.”
This is what shitty and self-serving political analysis looks like. It’s pure hackery, and it’s no different from Trump just making stuff up and saying he’s winning. Chotiner and Cohen asked Rosenberg for proof of his assertions and the best he can come up with is a condescending “it doesn’t matter” reply. This is how Hillary lost.
I used to believe this election is a pure toss-up, and that it will likely be decided by stuff we don’t factor into our political calculations that still has a measurable impact, like rain patterns on election day. But after reading every sad and desperate line of this interview with a presidential campaign strategist, I’m convinced that Joe Biden is actually down by 50 points.