Skip to main content
ABC News
Harry Talks Polls With The Guardian’s Tom McCarthy

We’ll be reporting from Cleveland all week and live-blogging each night. Check out all our dispatches from the GOP convention here.

CLEVELAND — If you’ve never attended a political party’s national convention, let me tell you that there are a lot of media here. While Donald Trump needed only 1,237 delegates to clinch the nomination, 15,000 members of the media came to town to cover it. Among the masses: Tom McCarthy, who writes about politics for the Guardian US. I use data all the time, but I’m always curious about how more traditional reporters work with polls. So I sat down with Tom and asked him how he uses polls and how readers react to them. (The transcript below has been lightly edited for length and clarity.)

Harry Enten: How much do your readers trust the polls?

Tom McCarthy: People are skeptical of polls and pollsters. They are traumatized. Readers took false hope [when the polls showed the Brexit vote failing]. As a result, readers get in the comments and say, “Why are you bothering?” This is especially the case if there is a stand-alone piece on a poll. Readers are suspicious. Once bit, twice shy. … Guardian readers feel about pollsters as Trump voters feel about the media: They are angry. You wouldn’t want to meet a Guardian reader in the alley if you are a pollster.

Harry: So why do you bother?

Tom: It’s a love-hate relationship. They don’t trust them, but they can’t quit them. … They drive traffic. … There is a paradox. Nothing gets them going more than a Quinnipiac poll showing Trump up in Florida. I’m in a position where I don’t want them to shoot the messenger.

Harry: How do you choose which polls to write about?

Tom: I noticed us not covering the last two Rasmussen polls that came out. It feels like misinformation at this point. I would say I don’t trust the outliers, especially ones with an established bias. There are certain polls, such as those very in-depth Pew Research Center polls with a wealth of demographic information, that we’ll will jump on and do a stand-alone. We place it in context.

Harry: So you trust Pew Research Center polls? What other ones?

Tom: SurveyUSA, Ipsos, Marist and the ABC News/Washington Post poll are among my most reliable. Basically, I think that pollsters tied to the largest news organizations are the most reliable. I’m not going to distinguish among the different media pollsters (Ipsos, Marist, etc.), but I think this is a case of the so-called mainstream media having better information. Those pollsters are better than a Breitbart poll. I will say, I think Gallup is an underrated pollster. What happened with Gallup in 2012 was unfortunate. It’s very disappointing they aren’t running a tracking poll [this election].

Harry: Has the way you’ve use polls/data changed in the past eight years?

Tom: One thing we do now is discuss the polling average whenever we mention the top-line number in the poll. There must be context. It’s OK to mention a trend. I think we take blind faith in these polling averages. My colleagues may have done this, but I have not opened the hood to see how these different averages (Huffington Post/Pollster.com, RealClearPolitics, etc.) work. And these numbers have differed. I tend to trust whoever the data people around me trust.

Harry Enten was a senior political writer and analyst for FiveThirtyEight.

Comments