Axios, Internals, and How Not To Report on Polls

From left, Rep. Rashida Tlaib, D-Mich., Rep. Ilhan Omar, D-Minn., Rep. Alexandria Ocasio-Cortez, D-N.Y., and Rep. Ayanna Pressley, D-Mass., respond to remarks by President Donald Trump after his call for the four Democratic congresswomen to go back to their “broken” countries, during a news conference at the Capitol in Washington, Monday, July 15, 2019. All are American citizens and three of the four were born in the U.S. (AP Photo/J. Scott Applewhite)

Polling, it’s fair to say, isn’t exactly high in the public estimation following the 2016 election, when most major polling outfits got the result of the general election wrong.

Even so, the pollsters of the world were mostly on the money in the four presidential elections prior to 2016 and were again mostly right about the midterm elections in 2018. There’s a good chance that the polling failure, like many other things, was an outlier in 2016.

With the 2020 race getting underway, polls are once again in the news on a near-daily basis, and candidates and their supporters continue to brandish them to tell certain motivated stories about the success of their campaigns.

There are certain rules about polls that both observers of politics and those who report on it would be wise to follow: Never take too much stock in one individual poll. Pay more attention to averages of several polls than to anyone in particular. And, always read the fine print about timing, methodology, and margins of error.

There was one story this week that played like a master class in how not to report on or react to a poll. And within days, it made its way to the Twitter account of the President of the United States.

Veteran political reporter Mike Allen of Axios ran a story on July 14 with the headline “Exclusive poll: AOC defining Dems in swing states.”

The story stated that “Top Democrats are circulating a poll showing that one of the House’s most progressive members — Rep. Alexandria Ocasio-Cortez — has become a definitional face for the party with a crucial group of swing voters.”

What is this poll? Conducted by an unnamed “Democratic group,”  it’s described as a survey of “1,003 likely general-election voters who are white and have two years or less of college education.”

And the poll contains some supposed bad news for the quartet of Congresswomen who have become known as “The Squad”: Alexandria Ocasio-Cortez is approved by just 22 percent of those surveyed, while Ilhan Omar is favored by just 9 percent. As for the concept of “socialism,” just 18 percent view it favorably. The poll, Axios says, “Is making the rounds of some of the most influential Democrats in America.”

To those who know anything about polling, there are quite a few things here that should ring the alarm bells.

For one thing, while the word “internal” doesn’t appear in Allen’s story, this is an internal poll. And not only are internal polls inherently to be taken with a huge grain of salt—they only ever leak to the media if the leaker wants to share their results—but we know nothing about the identity or reputation of the pollster, the methodology used, or how the questions were worded.

For example, the Axios headline refers to “swing states,” but there’s nothing in the story saying anything about the geography of the poll respondents, what qualified as a “swing state,” or whether they’re in swing districts.

Furthermore, the Axios piece positions white voters with two years or less of college education as natural swing voters, when everything we know is that that cohort is a solidly pro-Trump bloc. There’s nothing in the Axios story, other than the blanket assertion, that indicates that the people who responded are “swing voters.”

And besides, none of “The Squad” are on the presidential ballot in 2020. It’s virtually certain that the “face” of the Democratic party in the next election will be… whoever wins its nomination, not four freshman members of Congress.

Even veteran GOP pollster Frank Luntz questioned the poll on Twitter.

Not everyone on the Republican side, however, has been so honest about the poll.

Various conservative media outlets have aggregated the Axios story about Ocasio-Cortez and Omar’s approval ratings, either minimizing or omitting altogether the part where it’s an internal poll of only non-college-educated white people.

Gateway Pundit: “TRUMP IS RIGHT: Democrat Internal Poll Shows Only 22% Favorable Rating for Socialist Nut AOC and ONLY 9% Approval of Ilhan Omar”

The Blaze: “New poll has extremely bad news for Democratic Party, reveals how voters feel about AOC, Ilhan Omar”

The Daily Wire: “NOT A TYPO’: Democrat Poll Shows What Voters Think Of Omar, Ocasio-Cortez”

bongino.com: “OUCH: Internal Dem Poll Reveals AOC, Omar Wildly Unpopular”

Eventually, on Tuesday, the idea made its way to President Trump, who after the poll was released began a vicious feud with the same Congresswomen, which has since exploded into national theater.  This feud erupted from a series of Tweets from the president this weekend widely denounced as racist:

President Trump used the Democratic dirt first reported in Axios in his own mudslinging:

This poll, now being wielded by political opposition across the aisle, is what historian Daniel Boorstin described as a “pseudo-event.” It was almost certainly only commissioned in order to be leaked and reported on. And reporting on it was falsely, and possibly even maliciously, distorted.

It is important to expect veracity out of individual polls. Polling, done right, can be a useful tool for predicting political outcomes and gauging public sentiment. Even when a poll is done right, something that cannot be said of this latest internal poll clearly calculated to put pressure on the four Democratic Congresswomen it targeted, it should be weighed as a single data point—the average of several of these polls will always tell the more accurate story.

The poll at question here is polling at its worst—zero transparency, and obvious political motivations—and is frankly the kind of poll that stokes public disillusionment with polling as a practice.