The New York Times Proves Once Again That its Reporters Need to Learn How to Read Polls
Dear The New York Times, don’t take this the wrong way, but I think you should seriously consider not letting Shane Goldmacher anywhere near political polls ever again.
This is not the first time I’ve had to point this out. But apparently some newspapers never learn and keep assigning the same clueless reporter the same type of articles.
In yet another huge bucket of stupid masquerading as journalism, Mr. Goldmacher lets us know that according to a recent The New York Times/Sienna College poll, the advantage the Democrats once had in the upcoming Midterms due to the repeal of Roe vs. Wade has evaporated and now the Republicans have the statistical advantage.
Apparently women no longer care that their right to abortion is being taken away from them. They’re no longer care about losing control over their own bodies.
“The biggest shift came from women who identified as independent voters. In September, they favored Democrats by 14 points. Now, independent women backed Republicans by 18 points — a striking swing given the polarization of the American electorate and how intensely Democrats have focused on that group and on the threat Republicans pose to abortion rights,” Goldmacher notes.
There are a number of things going horribly wrong here:
Math. I know math is hard. But all you really need to know when dealing with polls is basic addition and subtraction. Independent women favored Democrats by 14 points in September and then shifted to Republicans by 18 points in October. That’s a 32 point shift in one month. That large a shift generally doesn’t happen in polls, unless something is horribly wrong either in the poll’s mechanics or in interpreting the numbers.
The New York Times really doesn’t want the Democrats to win. The Republicans are much more fun to cover. Our side doesn’t have any cartoon characters like Majorie Taylor Greene or Lindsey Graham. We have Nancy Pelosi and Chuck Schumer. And let’s face it, they’re pretty boring.
An over reliance on polls in the first place. Like using Twitter as your main source of news or story ideas, reporters really should get over their reliance on using polls to predict the future. Polls are a snapshot in time and not a reliable predictor of anything. Plus, you can get a poll to come out any way you want by asking certain questions in certain ways.
Not talking to enough people. Too many polls rely on too small a sampling of the population to be useful. The smaller pool of respondents, the higher the margin of error. The higher margin of error, the lower the utility of the poll.
Predicting the future based on a single poll. One single solitary poll is not (and I repeat, is not) a good predictor of anything. There are so many things that can be skewed in a poll that one single survey is not representative of anything.
Using a national poll to predict what will happen locally. This poll is a survey of voters across the country. It is not predictive of individual races in states or Congressional districts.
This Sienna College poll surveyed 792 people labeled as “likely voters” and “independent voters.”
According to Tom Bonier, CEO of analytics firm TargetSmart and a critic of polls, surveying “likely voters” at best will give you a “plausible outcome” not a definite outcome.
“Never pay attention to a single poll,” he recently told The Nation’s Joan Walsh. “View these polls as a ‘what if’ scenario. They’ve constructed a ‘likely voter model,’ with a potential outcome that’s plausible.”
He noted that the biggest flaw in the poll, which naturally was the part the media jumped on really hard, was the alleged 32 point swing in sentiment among independent women. This was based on a survey of 95 woman, he said, and had a margin of error of at least 10 points.
“Nobody should have reported that as truth,” he told The Nation.
The New York Times’ Goldmacher also notes that the economy is top-of-mind for respondents.
“The survey showed that the economy remained a far more potent political issue in 2022 than abortion,” he wrote.
Maybe among the 95 women who were surveyed. But I guarantee you that abortion is still a major issue among the majority of women in this country. The economy is also an important issue for them, but they are not forgetting about abortion, as The Times would like you to think.
The decisive “tell” that the poll was flawed was its finding that women are splitting their votes evenly between Republicans and Democrats, Joan Walsh notes in her piece in The Nation. “Do you really believe just months after losing a fundamental right, women will split their votes [between Republicans and Democrats]?” she quoted TargetSmart’s Tom Bonier as asking.
Like Tom, I really don’t think so.
“I just think we always need to be careful to over interpret one single survey,” noted Democratic pollster Anna Greenberg, in a recent The National Review article by Daniel Strauss. “But I am skeptical that Democrats and Republicans are breaking even among women. Have we ever since the ’90s had a situation where women didn’t vote more Democratic than men did?”
Strauss ended his article with this: : “Interviewing a handful of pollsters, perhaps fittingly, results in a handful of distinct responses. On this Times poll, though, there were a few consensus points: This poll is just one data point; it’s going to be a close election; anybody who says they know what’s going to happen doesn’t. But a one-month, 32-point swing based on 95 women maybe shouldn’t top the list of things Democrats ought to be worrying about.”
I say “Amen to that.”
But, I would also add that maybe skewing the results of a single poll to fit your agenda should not be how a media organization practices journalism.