Hillary Clinton makes an appearance at Temple University in Philadelphia on Sept 19. (Melina Mara/The Washington Post)

On Saturday, a lead headline in the New York Times announced, “Clinton Team Steels Against Email Fallout.” With FBI Director James B. Comey’s letter the previous day announcing the discovery of a new trove of emails linked to Hillary Clinton’s private server, a potential “October surprise” had finally arrived.

Yet it shouldn’t have been a “surprise” at all. Why not? The answer, somewhat ironically, appeared in a Washington Post headline the next day: “FBI email probe has upended the ‘Clinton is coasting to victory’ media narrative.”

The media’s urgency to maintain drama in an election that was increasingly looking like a blowout made this story all but inevitable.

By definition, to be “news,” a story must be new. Pursuit of novel stories is thus a core media news value. A dramatic horse race in which the outcome is uncertain and continually fluctuating is perpetually novel. Additional stories about the candidates’ long-standing policy positions? Not so much.

So media coverage of presidential elections disproportionately emphasizes the horse race. Research by Harvard University’s Shorenstein Center, for instance, shows that during the 2016 primary season, 56 percent of news coverage of the campaign focused on the horse race, compared with just 11 percent on substantive policy concerns.

What does this have to do with the email story? This story raises the possibility of halting the “blowout” narrative and replacing it with an unpredictable (hence novel) horse race. This keeps the audience engaged.

We counted the number of stories mentioning the phrase “Clinton email” between Aug. 1 and Oct. 27 — the day before Comey’s latest bombshell — appearing in all news broadcast transcripts (television and radio) available in Lexis-Nexis. We then compared this trend with the polling average on RealClearPolitics (RCP).


(Graph by Matthew Baum and Phil Gussim)

The two trends are strongly and statistically significantly correlated (.33 where the maximum positive correlation is 1.0). As Clinton’s lead in the polls goes up, the number of stories mentioning “Clinton email” follows suit. As her lead declines, the frequency of such stories declines as well.

Thus, when Clinton’s lead increases, making it seem more likely that she will decisively defeat Trump, news outlets seeking to sustain the horse-race narrative find it opportune to reexamine what has consistently proven to be the most damaging story line for the Clinton campaign. Conversely, when her lead shrinks and the horse race appears competitive, the email story loses some of its luster — at least absent a new revelation, like Comey’s letter to Congress.

Could email stories be driving the polls, rather than the other way around? Only if one believes that more stories focusing on Clinton’s email troubles tends to push her poll numbers higher, which seems unlikely on its face. Consistent with this logic, we find that polls predict news coverage more strongly than news coverage predicts Clinton’s poll numbers.

But perhaps the real culprit is WikiLeaks, strategically releasing hacked emails, and thereby demanding media attention, whenever Clinton’s lead expands. For instance, the Oct. 7 release of emails belonging to Clinton campaign manager John Podesta followed a week during which Clinton’s lead in RCP’s polling average expanded from 2.7 to 4.7 points. On the day of the release, stories mentioning “Clinton email” doubled from the previous day. Yet over the next several days, attention to her emails fell off sharply, suggesting that WikiLeaks failed to drive the media narrative, at least beyond a single day.

So while the latest turn in the saga of Hillary Clinton’s emails was unforeseeable, the media’s renewed interest in the email story was not. Given Clinton’s recent lead in the horse race, the fact that we are again talking about this issue a week before the election is no surprise at all.

Matthew A. Baum is Marvin Kalb Professor of Global Communications at Harvard University.

Phil Gussin is a professor of political science at College of the Canyons.