Three journalists of different stripes analyzed the media’s role in November’s midterm elections and presented their views for the future of political journalism during Saturday's Zeidman Colloquium on Politics and the Press.
Nate Silver, founder and editor-in-chief of FiveThirtyEight, joined Molly Ball, national politics staff writer at The Atlantic and Brian Stelter, senior media correspondent at CNN, for a panel discussion on election coverage and data journalism. The discussion was moderated by Bill Adair, Knight professor of the practice of journalism and public policy. During the 90-minute event, the panelists examined the successes and failures of pollsters and pundits during the 2014 elections and discussed the integration of quantitative journalism tools into the mainstream media.
“It’s okay to be surprised. That’s why we have news, because things happen that we weren’t anticipating,” Ball said regarding some of the unexpected outcomes seen this year. “I know that people think that we’ve made a factual error when we fail to predict the future, but that’s not true. The future is unknowable, and that’s why we do this.”
See The Chronicle's interview with Nate Silver on political predictions, media innovation and choosing the right career.
Underestimated Republican victory
Silver, Ball and Stelter spent time dissecting how news organizations and polling models underestimated the scale of the Republican victory earlier this month. Although most national media and polling organizations predicted that Republicans would have a good election night, Republicans were generally not expected to do as well as they did.
The three all agreed that polling inaccuracies did not have a major impact on the overall coverage. They said, however, that some lessons could be learned from the process. Stelter said that this year was much better for the media than previous election cycles, where some outlets made predictions that were clearly off the mark.
“There were no big mistakes this year,” Stelter said. “There was no giant egg-on-the-face, eat-your-shoes moment."
Silver pointed out that polls and election models are supposed to have some uncertainty, and that models cannot be judged simply by how often they pick the winner.
“At FiveThirtyEight we emphasize that we are making forecasts that have the uncertainty built in, and the uncertainty can be considerable,” Silver said. “If you’re making your 70-30 calls, you’re supposed to get those wrong 30 percent of the time. Not only will you, but you’re supposed to or you’ve done something wrong.”
Flaws in media coverage
The panel also analyzed the flaws in midterm election coverage this year. Ball pointed out that the severe cuts to local newsrooms and lack of focus on local politics may have resulted in the national media being taken by surprise in several races. The national media missed the sense of discouragement and unease on the ground that made voters reluctant to vote for incumbents.
“We misread the national mood,” Ball said. “There was a sort of sense of discouragement and disappointment. That was a phenomenon that mostly affected Democrats, it turns out. They didn’t see reason to vote.”
Silver concurred, adding that there also may have been a perception that there was less at stake in this election than in the past few election cycles since a switch in the Senate's majority party will not be able to change the overall divided government in Washington. This resulted in lower turnout, which generally hurts Democrats, he said.
“It was not the sort of existentially important election that you might have had two years ago or four years ago,” Silver explained.
Stelter identified the media itself as part of the reason why turnout was low. He pointed out that there was much less coverage of the midterms from the major network news organizations in the months leading up to the 2014 elections than there had been in the same period leading up to the 2006 elections.
“We need to put some of it on the press,” Stelter said. “We do drive interest. We drove interest in Ebola. We drove interest in ISIS. We did not drive interest in the election. Those may not be conscious choices, but it’s interesting to think about the ways we are acting as supposed to reacting.”
A polling 'crisis'
The panel also identified issues with the polls themselves as a cause of error in election predictions this year. They cited both instances of bad polling methodology and issues across the board. Ball said that the polling industry is itself having trouble as it tries to deal with a population that responds less and less to traditional polling methods.
“There is a crisis in the polling industry,” Ball said. “These days people don’t have landlines, or they don’t answer them, so pollsters have to call more cells phones, and there’s different rules on that, but people don’t answer their cellphones either.”
Silver pointed out that some of the incorrect forecasts may have been caused by questionable polling methods from some organizations, such as suppressing outlier polls or seeking to create polls that conform to the polling average. This reduces the independence of polls and makes systematic error more likely, he said.
Future of media in politics
The conversation also explored the future of political media. The panelists discussed how data journalism could be incorporated into the mainstream of political media. Ball said that she already uses data as a starting point for her reporting and sees part of her job as explaining why the numbers are what they are, rather than making predictions about what the outcomes will be.
“There’s a lot of talk about the supposed tension between what Nate [Silver] does and what I do, between first-hand reporting and data-based reporting, and I think they’re absolutely complementary,” Ball said. “I use polls very much to try and figure out what’s happening in an election and then go out and try and figure out why.”
Silver said FiveThirtyEight is still figuring the best ways of doing data journalism and using digital platforms. He noted that it is sometimes hard to figure out exactly which pieces will appeal to a wide audience and which will not.
Stelter added that the ongoing transition to data and digital is a reason to be optimistic about the news industry as a whole, and spoke about the transitions at his own company in particular.
“We’re more responsive to our audiences,” he said. “Here’s how I think CNN’s changing. CNN is primarily in the future a digital product that happens to have a TV channel. Right now the TV channel is where the money comes from, for the most part. But it’s increasingly going to be a digital operation.”
The panel agreed that these changes in the political media were not only challenges but great opportunities. News organizations have taken a multitude of approaches to integrate new tools and produce new types of content.
“Whether it’s a good time to be a journalist or not, I’m not sure,” Silver said. “It’s an interesting time. But I think on balance it’s been good for the consumer.”
Get The Chronicle straight to your inbox
Signup for our weekly newsletter. Cancel at any time.