Your first month is free in The New York Times app. Download now → Good morning. Senators are raising concerns about poor conditions aboard the aircraft carrier Abraham Lincoln. It’s been at sea for nearly nine months. We have more news below, but we’re going to start today with recent problems with political polling.
Number problems
Progressives thought they would clobber the moderates in primaries this month. The polls had set extraordinarily high expectations: a double-digit lead for Dr. Abdul El-Sayed in the Michigan Senate contest and more than 20 points for Francesca Hong in Wisconsin’s primary for governor. Instead, El-Sayed won by one point, and Hong lost. Primary elections have always been hard for pollsters. They’re uncertain and volatile contests. The turnout is low and voter preferences can switch on a dime. Even so, it is pretty unusual for polls to badly overestimate the same political faction twice in two weeks. Do the errors betray a deeper problem measuring the strength of moderates against the progressive left? It’s hard to say for sure based on just two races, but the answer may be yes. A Wisconsin fluke
The polls missed by a lot more in Wisconsin than in Michigan, but pollsters deserve a pass in the state where they did worse. This was a classic example of why primaries are so hard to poll. Progressive Democrats quickly consolidated behind Hong, a democratic socialist. The party’s establishment, on the other hand, struggled to put forward a candidate at all. Over the final month or so of the race, three establishment-backed candidates dropped out of the race. Incredibly enough, this included the eventual winner, David Crowley, who left, came back and ultimately won. With progressives unified and moderate voters divided and bewildered, Hong built an enormous lead in the polls — and the ingredients for a major polling error came into place. In the end, Crowley’s support surged as moderates coalesced behind him. (I talk more about this on “The Daily” podcast.) The polls did slightly overestimate Hong — she won 39 percent of the vote, while the polls showed her at 44 percent — and it’s possible this betrays a deeper kind of error. But it’s also possible she lost support in the final week of the campaign amid stories highlighting her general election vulnerabilities, such as her apparent opposition to Thanksgiving and Halloween. Michigan’s deeper problemThe pollsters don’t have easy excuses in Michigan. El-Sayed and Representative Haley Stevens, the establishment-backed candidate, were clearly the top two. They offered a straightforward contrast, received national media attention and benefited from tens of millions of dollars spent on their behalf. Voters had every opportunity to learn about them and the stakes. It’s hard to believe many voters changed their minds. So what happened? The pollsters got something wrong, and some explanations are more troubling than others.
One relatively innocuous explanation would be if pollsters were wrong about the makeup of the electorate. Maybe they overestimated the turnout among young voters, for instance, who overwhelmingly backed El-Sayed. This is a real possibility. The final Emerson College poll, a rare high-quality nonpartisan poll of the race, reported a much younger likely electorate than in other recent primaries in Michigan. If turnout was the problem — something specific about this race — pollsters wouldn’t be doomed to make the same mistake in the future. A more serious problem would be something called “nonresponse bias,” where one candidate’s supporters are much likelier to take a poll than others. This is the challenge pollsters have faced when measuring President Trump’s support, and it’s a hard problem to solve. In Michigan’s case, progressive candidates tend to fare better among highly educated and highly engaged voters, who are likeliest to respond to political surveys. Worse, pollsters increasingly rely on methods that might exacerbate this problem, like by polling people who sign up to take polls online or who are eager to click on a text message link to a poll. If the problem is nonresponse bias, then pollsters could be in trouble. Over the last decade, they’ve relied on heavy-handed techniques to represent Trump’s supporters among respondents, like ensuring the right number of voters who say they voted for him in the last election. There is no equivalent way to do that in a first-time contest between a progressive and a moderate Democrat. (By the way, I write a newsletter for The Times about political data, so if you want to keep up with these issues through the midterms and the presidential election, sign up here.) It’s possible that what happened in Michigan is just a one-off polling error — the wrong pollsters in the wrong place at the wrong time. But if not, it might offer a clue about just how bad many polls are without the crutches they’ve used to hobble through the Trump elections. For more: Socialists have wanted to take over the Democratic Party for years. They’ve never gotten this far. Our reporter went inside their audacious rise.
Politics
Around the World
Science
Weather
Other Big Stories
As extreme heat becomes normal, firefighters in Phoenix have adopted a technique that could save lives across the country. Click the video above to watch. The American left lost its way with D.E.I. and cancel culture, two writers argue. But they say the right’s anti-woke backlash is also spiraling out of control. Deeply reported journalism needs your support. The Times relies on subscribers to help fund our mission. Become a subscriber today.
We’ve made these links free for you. A different light: France has faithfully renovated most of what burned at Notre-Dame in 2019. But not these six windows. Diss track: A rapper in the West Bank released a song criticizing a fellow Palestinian musician for his Western lifestyle. It’s drawn comparison with the feud between Drake and Kendrick Lamar. Your pick: The most clicked link in The Morning yesterday was tips from ophthalmologists about caring for your eyes. Faking it: Marketers are creating A.I. personas that can host podcasts, star in YouTube videos and converse with followers. Times readers had thoughts.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
