Links for March 12-18, 2023 | A Trump indictment; The GOP against Ukraine; Fox's Arizona 2020 call; and Rasmussen is at it again
Plus, centrists are preparing a spoiler election candidacy for 2024, and comments on "the gerrymandering myth"
Happy weekend, all
This is my weekly post for paid subscribers discussing recent uses of (mostly political) data that I think are interesting and worth discussing. Comments are welcome below — and if you enjoy this, please share it with a friend!
There are a lot of links to cover this week, so I’ve included only a brief discussion of each in an effort to cover all the ground in a digestible fashion.
1. Natalie Jackson: “What polling says about a possible Trump indictment”
The New York Times reported over the weekend the former president is claiming the Manhattan DA will soon indict (and presumably arrest) him for the alleged crime of ordering hush-money payments to Stormy Daniels in 2006, an adult film actress. This raises an important question: What would happen if Donald Trump is indicted for a crime between now and the 2024 Republican presidential nomination?
Natalie Jackson, a pollster and the former Huffington Post election forecaster, looks at the potential fallout through the lens of the aftermath of the January 6th insurrection at the US Capitol. There is a lot of polling data suggesting Trump still has a tight grip over the party. Jackson writes “If we’ve learned nothing else by now, we need to have learned that Trump read the room better than the rest of us when he said he could shoot someone on 5th Avenue and still hold onto his supporters.”
The post is available at her new Substack newsletter.
2. Do Republicans really not care about Ukraine?
Ron DeSantis, Florida’s governor, said in a press conference last week that the US should distance itself from Ukraine:
“while the US has many vital national interests … becoming further entangled in a territorial dispute between Ukraine and Russia is not one of them.”
DeSantis is in step with the plurality of Republican voters on this subject, according to polling from Echelon Insights and reported by Kristen Soltis-Anderson in this blog post. The money chart:
And a recent YouGov/The Economist poll finds that a plurality of Republicans is opposed to sending weapons and financial aid to Ukraine. The link for that survey is here.
3. Fox’s way-too-early Biden 2020 Arizona call, 2.5 years later
Nate Cohn has a long newsletter read on how the Fox News Decision Desk ended up making its infamous way-too-early call of the 2020 presidential election in Arizona, which has made headlines again recently as a result of the lawsuit between Fox News and Dominion Voting Systems. There’s not much for me to really pull here (the article is good so please go read it yourself) but I suppose the upshot is that modeling election outcomes is very tough stuff, especially when you go from having a negligible partisan split in early and mail votes in your training data to a very, very polarized split by vote method in your testing data:
This is a very interesting story for polling nerds and probability buffs, and a profound one for people who care about election integrity and the media. Fox was 11,000 votes and 0.3 percentage points away from being wrong about the call, which Decision Desk head Arnon Mishkin said on air would have been a 1-in-10,000-chance event.
That would have given Donald Trump a lot of ammo in his subsequent claims that the election was stolen from him: If everyone was wrong in Arizona, what else were they wrong about?
4. Rasmussen is at it again
Sticking with Arizona for now: The polling organization Rasmussen Reports published an “exit poll” this week purporting to show that Kari Lake, the 2022 Republican governor candidate in the Grand Canyon State, actually won her race. Mark Mitchell, Rasmussen’s pollster, appeared on Steve Bannon’s web show and proclaimed (I’m quoting from Washington Post journalist Philip Bump here):
“A lot of pollsters, ones with maybe less courage, would really use [horse-race poll] results to weight the poll to match the outcome of the elections,” he said, “and then bury the questions and pretend that they never even asked them. And, you know, to be honest, these results were just so unbelievable that we had to report out.” Instead of weighting the responses to the results, he said, they weighted it to exit polls.
“What we found,” he added later, “is that voters told us that just four months ago they elected Kari Lake as the governor of Arizona by 8 points.”
In November, however, Lake actually lost by about half of a percentage point. Those results have undergone multiple audits, and challenges by Kari Lake have been thrown out in court. Rasmussen is essentially claiming that since it conducted a poll during the race that had her up above the margin of error, then it’s a statistical impossibility that she lost.
In reality, polls are not this precise, especially when conducted online—where the laws of random sampling and the traditional margin of error do not apply. Of course, that’s not going to persuade Kari Lake or Liz Harrington, Donald Trump’s spokeswoman, who tweeted out the poll last week claiming the former president won the state in a landslide. (He didn’t).
5. How the ‘No Labels’ Gambit Could Wreck the 2024 Election
No Labels, a partisan organization for centrist political candidates, has recently made headlines for reportedly preparing to put forward a candidate for president in 2024. It has nearly $100m in funding, apparently, and hundreds of staffers in key states gathering signatures to put the organization's “unity ticket” on the ballot for next year.
They write on their website:
We are the voice for the great American majority who increasingly feel politically homeless. We’re . . . laying the groundwork to ensure the American people have a real choice in the 2024 presidential election . . . to bring our divided country back together and solve our most pressing problems.
But writing for The Bulwark, political commentator Norm Ornstein writes that
In fact, the great majority of Americans do not act as if they feel politically homeless. While a plurality of voters call themselves independents, as Geoffrey Skelley noted in FiveThirtyEight, “roughly 3 in 4 independents still lean toward one of the two major political parties, and studies show that . . . [i]ndependents who lean toward a party also tend to back that party at almost the same rate as openly partisan voters.”
This is right. At the least, No Labels is lazy in constantly arguing that a large number of political independents would support a “bipartisan” presidential campaign next year. (At worst, they are doing so because it makes them lots of money—like most other political organizations.)
The natural counter-argument is a question: If there were a supermajority of Americans who want a third party, wouldn’t there be one at some level of the government—where the stakes are a lot lower than in, say, a presidential election? Is it so hard to imagine that most people are either closeted or outright partisans who want their team to win?
3. The “gerrymander myth”
The Brookings Institution’s Bill Galson, who I know and respect, writes this week:
Journalists, pundits, and some political scientists argue that gerrymandering distorts representation and gives an unearned advantage to the Republican Party, which controls the majority of governorships and state legislatures. This belief has earned the redistricting process a secure place in the reform agendas of many activists. This may have been the case in the past, but a straightforward analysis shows that it isn’t true now and hasn’t been for several election cycles.
And he includes this table to back up his take:
The argument is that since the Republicans won a proportion of seats very close to its proportion of the national popular vote, there is no bias in the House electoral system.
I normally agree with Galston, but he makes three big errors here. First, he does not appear to use a calculation of the national popular vote which controls for an excess (lack of) votes cast in uncontested seats (depending on which party you look at). Republicans did not win 51.4% of the two-party votes for the House, as his chart implies, but actually 50.8%. With such a number the Republicans “should have” won 220 seats for the House, not the 224 that he calculates—a small difference, yes, but an important one.
The second and bigger problem is using aggregate totals of House seats to measure gerrymandering. What we know from studies of the 2022 Congressional maps at the state level is that there is a lot of gerrymandering across the states that create large biases for either party within state delegations. But those biases happen to cancel each other out when you aggregate results at the national level.
But the final error, and one I harp on constantly, is that House election results are not the right benchmark for studying gerrymandering. House elections have a lot of idiosyncracies at the state and district levels—whether a Senate candidate is polarizing or an incumbent is running, for example—that can push expected votes away from the partisan baselines in each constituency. For example, if a usually-Democratic seat is gerrymandering to the extent that Biden would have won just 45% of the vote in the 2020 election, but a high-quality Democratic House candidate runs against an election-denying Republican and wins 56% of the vote in a huge victory, is the seat still gerrymandered? We would probably say yes, but this analysis would indicate it is not.
So those are a few reasons why the “gerrymandering myth” is, actually, a myth.
That’s it for this week’s top links. Thanks for reading and being a member of the community supporting this newsletter. Consider sending a free trial to a friend you think will enjoy the subscriber-only content.
Have something interesting for me to write about? Send it to me on Twitter or via email (I’m gelliottmorris@substack.com).
Have a great week,
Elliott