Avoid Getting Played by ‘Polls’

by July 1, 2020
Business journalists should always remember to take a careful and analytical approach to the information gathered by polls. Here’s where to start. (Photo via Pixabay user Pexels)

Who confuses the coronavirus family and Corona beer? Earlier this year, it would seem like many people, if you believed a poll that had come out.

“Survey finds 38% of beer-drinking Americans say they won’t order a Corona,” a headline on CBS News ran.

CBS had company as well. CNN, the New York Post, the New York Times, and others either ran a piece on the poll itself or mentioned it in broader coverage.

And they mostly all got played.

PR firm 5W was the one that had spread the supposed poll around. Originally there had been a release online, according to the CBS story that linked to it, but now the URL brings up a 404 page-missing error. According to that story, 5W had made the survey calls itself.

Still, played. Here’s how Atlantic contributing writer Yascha Mounk explained a basic problem:

The original press release from 5WPR notes that in a survey of 737 beer-drinking Americans, 38 percent said they “would not buy Corona under any circumstances now.” By presenting this finding in the context of other questions that are explicitly about the coronavirus, the press release creates the impression that Americans’ reluctance to drink the beer is due to the coronavirus. “There is no question that Corona beer is suffering because of the coronavirus,” Ronn Torossian, the CEO of 5WPR, says in the press release. “Could one imagine walking into a bar and saying ‘Hey, can I have a Corona?’ or ‘Pass me a Corona.'”

But imagination isn’t enough to support a survey. There must be questions and answers that hold together—information that reportedly wasn’t included with the release. Specifics about the sourcing and demographics to help explain whom the respondents actually represent.

Also missing was a careful and analytic reading of the information on the part of multiple journalists. When you see a number that suggests a percentage of the public won’t buy a given product, recognize that you’re not being told why they wouldn’t. The 38% number could have been a group that had a pre-existing preference for another brand.

Something I had noticed wasn’t addressed by some who were critical of the response to the poll is why someone put the poll together and who ultimately might have been behind it. Was this a stunt? Perhaps a competitor to Corona wanted to undermine the brand. There were the most critical questions that too many journalists apparently never considered.

These are the most important questions a reporter can ask. You want any wider context and that is often illuminated by the interests of those sponsoring a campaign. As with so many aspects of business reporting, remember to always follow the money.