In the lead-up to election night, Twitter is full of pleas to news outlets not to do what Donald Trump wants and announce definitive election results on Tuesday evening, before states have a chance to count their mail-in and absentee ballots. Trump has explicitly appealed to tradition (“that’s the way it’s been, that’s the way it should be”) in making his case for a call “night of.” But when did it become an American expectation that we would know who won on election night?
I spoke with Ira Chinoy, a former journalist and now a historian of journalism at the University of Maryland, about the history of media outlets “calling it” on election night. We talked about people camping outside of newspaper offices, desperate for information, in the 19th century; the fallout from some of the most famous failed calls; and why the media’s involvement in calling election outcomes is a good thing for democracy.
Our conversation has been edited and condensed for clarity.
Rebecca Onion: It seems like the history of results being announced on election night, or near to it, starts in 1848—the first time there was actually an “Election Day,” instead of a longer period of voting, and the first time it was possible for the AP to use the telegraph to collect the news of the way people voted. Is that right?
Ira Chinoy: Right, yes. [Morse’s] telegraph was about four years old then, and the AP had grown up around sharing dispatches from the Mexican-American War. But there wasn’t a transcontinental link until 1861. In between, there was the Pony Express to fill in the gaps in telegraph lines between East and West. But even when there was a transcontinental link, the whole country wasn’t completely wired, really. Results needed to come in from distant precincts on horseback, carriage, or train—whatever it happened to be.
And so, is it fair to say that in the 19th century, the news media kind of saw a hole to fill, and started reporting results on election night?
I don’t think the news media created an appetite for the results to be known on election night. I think this is the appetite: the democratic impulse. People want to know; everybody wants to know. In the 1850s, ’60s, and ’70s, people would hang around newspaper offices waiting for returns to come in—not to buy newspapers on that night, but simply for them to announce the results by posting sheets of paper with returns or making announcements. Take the famous election of 1876, which was Hayes vs. Tilden, and which didn’t resolve until 1877, when the parties came to some kind of agreement. For days after the election, the New York Times reported that people were milling around the newspaper offices, waiting for any scrap of information.
It’s funny—it’s portrayed like the news media are these people who have come in and taken advantage of the situation [in pushing to announce a winner as soon as possible], which is really not correct at all, in my mind. This is driven by public appetite. Elections are a case of social inversion—a time when people at the bottom have something to say. The world is driven by the power brokers in our society, corporations, government, churches, but this happens to be the day when an ordinary person has a say. So if you’ve got an excess of a hundred million people voting, you’ve got an excess of a hundred million people who really want to know the outcome. I don’t think it’s a demand that’s ever been created by the news media; they’ve simply answered a public interest in knowing as soon as possible.
What was the level of trust in what the AP was saying about elections in the 19th century—what authority did it have? Were candidates saying, The AP said I won, so I’m going to claim victory?
I’m not sure I can answer that question, but I haven’t seen anything where people were challenging the vote counts as reported. Challenges to media “calling” the elections are really not so much about the counts, but about the projections.
People like the editor of the Boston Globe, Charles Taylor, became kind of famous for the system he invented that was first used in the 1880s, based on the idea of key precincts and bellwether areas. He would divide the state up into different kinds of areas, projecting based on continuity or diversions from past votes. And nobody else was doing that, yet, but there were other newspaper people who were famous for keeping these big books of data, being able on election night to see how the vote was diverging from the past. But Taylor had this system that worked so well, even political officials would come down to the newspaper to find out what the projection was. But there were other journalists who were reluctant to project winners, before all the votes were counted.
So it wasn’t something that was automatically accepted; there was a debate over it. It took a lot of courage to say so-and-so has won, before all the votes were counted. Really the test is, how well does a final vote count compare to everything people were projecting. There were famous cases where it didn’t work out—1916, 1948—but in the vast majority of cases, the news media were right in how they called those elections. What we remember were the ones that are particularly problematic.
When radio came along, did radio take over the function of calling elections on election night?
If they did, it would probably have been because they had access to the AP or some other kind of wire service. We didn’t really get robust radio network news departments until about the 1930s, 1940s. In that era, right into about World War II, there was a lot of conflict between the radio world and the newspapers—they called it “The Press-Radio War.” And some of it was a battle over reporting election returns. And radio kind of backed away from doing that for a while. They finally did develop their own robust news departments, but if they called it, it probably would have been based on something that came from the wires.
Even if we’re talking about newspapers calling elections during that period, they’re still doing it based on the wires. We didn’t have newspapers that have robust nationwide news operations in that period of time—or I should say, no news organization had reporters in enough individual places to do original reporting. It had to be a collaborative effort.
How did the advent of network TV news change the scene? Was there ever a moment where it became clear that the AP versus news networks were competing to call the election in different ways?
I’m not sure about that, but of course the networks had different decision desks, and they have their own criteria by which they’re going to call things.
In the 1980s NBC famously called that race for Reagan very early in the evening—before the polls closed in California. And since then, we’ve stopped calling a race before the polls close on the West Coast. I think Jimmy Carter even conceded when NBC made that call, but it became controversial because people were still voting on the West Coast, and there was a concern that it was putting a damper on people going out to vote in the West, when there were lots of other things on the ballot besides the presidential contest. There was drama around that, complaints that this had affected the results of the election; so news organizations have since then typically voluntarily decided not to call the race until the polls close in the West.
Are there other times where the way media has called the election on election night has caused controversy, or where a blown call, so to speak, has changed the way they do it? Did the “Dewey Defeats Truman” headline, or the way the networks had to walk back the 2000 decision, change anything?
After the 1948 incident [when the Chicago Tribune declared Dewey had won, only to have to reverse course], looking at the polls that were conducted in the run-up to the 1952 election, it looks to me like the pollsters were hedging their bets a little bit. They weren’t willing to say this was going to be a blowout for Eisenhower, the Republican candidate. And in fact, it turned out to be a huge blowout—remembering that there were fewer total electoral votes, he got 442 to Adlai Stevenson’s 89. But the news divisions were very skittish on election night in 1952—not wanting to rush to judgment at all.
One of the things that happened behind the scenes in 1952, which I’ve studied, was CBS had made arrangements with Remington, the maker of the UNIVAC computer; it was running projections for them. In the first projections that came off, the computer predicted this was going to be a blowout for Eisenhower. This wasn’t reported. The most likely reason is that the computer people didn’t want to go to CBS and tell them this, because the expectation was a close race. They may have been afraid that if they were wrong, this would set back the nascent computer industry.
I think what we see, as human beings, is tempered by what we expect to see, and I don’t think anyone wanted to call it early on, even though there were signs of a landslide that were very much there as the evening progressed—people out around the country, reporters, seeing a town that always went one way going another way. You could see all these indicators, but there was some reluctance because of 1948.
And I could be wrong about this, but I don’t remember, in the elections of 2004, 2008, 2012, and 2016, complaints that the media called the race too early; I think that’s in part because of what happened in 2000.
How does what happens in the United States, where the media projects and announces and the official count gets announced later, compare with what happens elsewhere? Is there anywhere where the media doesn’t do this, and waits for the state to announce?
It’s interesting to me! I feel like there is a frame in your questions that the news media are the boogeyman here.
Oh, I’m not trying to be that way!
Think of it this way. Without news organizations doing this, we are taking the word of the government for who won the election, without independent accounting for what happened. The news is serving an audit function here. Imagine if there was no independent journalistic reporting on election results. That would be outrageous.
I know! You’re completely right.
We do have a government system of counting—it’s called the Electoral College, and based on certifications by the secretaries of state, and we’ll get that in the middle of December. We have the Electoral College, and if you back up from that, we have a secretary of state, and if you back up from that, we have the county election boards, and if you back up from that we have precinct-by-precinct totals. While these precinct-by-precinct totals are being reported to the county and state level, they’re also being reported to the news media, that’s aggregating them, much faster than the state can.
It’s not a bad system. It’s not a terrible system. The news media is serving a watchdog role. In my mind, that’s one of their highest and best callings.