Study: Use of anonymous sources peaked in 1970s, dropped by 2008

Newspaper ombudsmen and media critics complain often about excessive and unnecessary use of anonymous sources, and yet the press uses them less frequently now than in the so-called “golden age” of journalism.

The use of unnamed sources peaked in the 1970s in the wake of Watergate. By 2008 it had dropped to the same relative frequency as in 1958, according to a paper to be presented at AEJMC this week.

“Going into this, I really did think that I was going to find that anonymous sourcing was used more than in the past,” said Prof. Matt J. Duffy, a professor at Zayed University in Abu Dhabi who worked on the study with Prof. Ann E. Williams of Georgia State University.

The other key findings:

  • Nowadays journalists almost always describe anonymous sources in some way rather than simply calling them “reliable sources.” In 1958, 34 percent of stories with unnamed sources used such vague language; that dropped to under 3 percent in 2008.
  • Reporters are doing a better job of explaining why they grant anonymity. In 2008, about a quarter of stories offered some explanation. While Duffy said that’s still low, through 1998 such explanations were provided in fewer than 10 percent of stories.
  • Journalists haven’t changed their practice of independently verifying all information from anonymous sources. They do so in most cases, but not all.

Duffy and Williams’ study is the first to track anonymous sourcing over such a long time with the same methodology. They did it by analyzing a representative sample (14 days) of front-page stories in The New York Times and The Washington Post, looking at one year each decade. (Focusing on just those two papers has its weaknesses, but others have used the same method to judge the overall performance of the press.)

Because half as many stories appeared on 1A in 2008 than in 1958, the rates are presented as a percentage of all front-page stories for that year. So at the peak in 1978, almost half of all front-page stories included anonymous sources. In 1958 and 2008, the figure is about 25 percent.

Andy Alexander, former ombudsman of the Post, said he wasn’t surprised that the rate has dropped. “Over the past few decades, I think there’s been a greater awareness throughout the news industry that excessive use of anonymous sources damages credibility,” he told me via email.

“But before we celebrate, let’s keep in mind that the study — correctly, in my view — said the use of unnamed sources remains unacceptably high.”

Len Downie, who was executive editor during part of the period studied, said he was pleased at the progress. Twenty-five percent “probably feels like it is so high, but again, we’re constantly battling the convention in Washington,” he said.

The paper quotes Bill Keller, executive editor at the Times, saying in 2005 that in a year he hoped his reporters would see the use of anonymous sources “is not a routine, but an exception.”

He told me via email that he was “heartened” by the improvement noted in the study, although since 2005 “I suppose I’ve come to feel more like Sisyphus. Excessive anonymity is one of those things that simply require constant vigilance.”

In the absence of identification, Keller said, readers deserve to have enough information to judge whether a source is believable or has an ax to grind.

The study notes improvement in describing sources and explaining why they aren’t named, which can help on both counts. But it didn’t judge the quality of the description or the merits of the explanation.

And as Clark Hoyt noted while ombudsman, in some cases those reasons are pretty thin. (Hoyt, who now works for Bloomberg, declined to be interviewed for this story.) Among the questionable examples Hoyt cited in a 2009 column:

In an article about the decoration of New York apartment building lobbies, a woman was granted anonymity to describe a particularly edgy one as ‘a den of hell.’ She had been visiting a friend in the building and would not give her name, the article said, ‘for fear of offending the hostess.’

Columbia University journalism students found that the paper didn’t follow its own guidelines on anonymous sourcing in 4 out of 5 cases, according to Hoyt. Hence the image of Bill Keller pushing that enormous boulder up the hill, only to have it roll back down.

Alexander also concluded that the Post often didn’t follow its own policies, in print and online. “I found that unwarranted use of anonymous sources often occurred in stories played inside the paper.”

“In many cases,” he continued, “it’s not sufficient to say simply that a source was granted anonymity ‘in order not to offend’ or so they could ‘speak more freely.’ Those are reasons, but they’re often not very good ones. As The Post’s internal policies say, granting anonymity ‘should not be done casually or automatically.’”

Though he found the study encouraging, Post Executive Editor Marcus Brauchli acknowledged shortcomings in meeting the paper’s standards, saying in an email, “We’re not perfect at this – many days, we fail to keep from using the noun ‘source’ to describe someone who has spoken to us – but our editors are pretty harmonized on the need to use anonymous sources only when there is no other way to get information.”

In another paper to be presented at AEJMC, Duffy challenges news outlets and the Society of Professional Jouranalists to raise their standards for anonymity.

The paper identifies Watergate as one reason for the rise in unnamed sources in the ’60s and ’70s. Downie, who was involved in that coverage, agreed.

“There was a period when reporters, especially young reporters, were tempted to use anonymous sources with that vague reference [“informed sources”] just because it sounded good,” he said. “We had to break them of those habits.”

The problem, Duffy said, is that journalists use “idealized” examples such as Watergate to defend the practice. “But I think in reality the pedestrian, day-to-day use of anonymous sourcing is pretty far away from that ideal.”

We have made it easy to comment on posts, however we require civility and encourage full names to that end (first initial, last name is OK). Please read our guidelines here before commenting.