Overall Summary

As a rule, the press has tended to describe anonymous sources in the vaguest terms in covering the Clinton-Lewinsky saga, according to the results of a new study of what and how the press has reported.

Only occasionally has the press offered audiences a glimpse of what the biases or allegiances are that might be influencing what an anonymous source is revealing.

The study, a follow up to an earlier one in February, raises basic questions about whether the press has become too lax about offering readers as much information as possible, and whether journalists have allowed sources to dictate terms too easily.

At the same time, there are signs the coverage over time has moved more toward factual reporting and named sources and away from commentary.

The study, conducted by the Committee of the Concerned Journalists, involved a examination of 2,051 statements and allegations contained in the reporting by major television programs, newspapers, magazines and the Associated Press over four days in January and March. For comparison, the study included a list of tabloid publications and television programs. The goal for this, the second part of a study conducted by the Committee in February, was to find out how sources were described, how the mainstream press compared to the tabloids, and how the coverage may have changed by the seventh week.

Among the findings:

  • Six in ten statements from anonymous sources in the mainstream media (59% of all anonymously sourced reporting) were characterized in the vaguest terms, “sources said,” “sources told our news organization” or “sources familiar” with the event.
  • Less than two in ten statements (17% of the anonymously sourced reporting) offered even the slightest hint of the source’s allegiances.
  • Print was more forthcoming about the nature of its anonymous sourcing than was broadcast.
  • The mainstream press’ use of anonymous sources was not that different than those of the tabloid press, such as Inside Edition or the National Enquirer, though the tone of the two different kinds of media, which is not quantified by the study, varied considerably.

Characterizing Anonymous Sources

In the first study it became clear that a key question was how much news organizations were helping audiences understand about anonymous sources–not simply whether the press was relying on such sources.

Characterization of Anonymous sources, Jan 23, Mar 5 & 6

“Almost everybody we are talking to (on this story) has an agenda, and I don’t think we’ve been very straightforward with viewers and readers on where that information is coming from and how it might be tainted as a result,” Dotty Lynch, political editor of CBS News, said at a conference discussing the first study.

So for the second half of the study, we decided to look at how anonymous sources were characterized–including to what extent audiences were given information to judge for themselves if a source might have an ax to grind.

We looked at one day in January and two in March and one week’s editions of Time and Newsweek. We then broke down the characterization of anonymous sourcing into five categories:

  • How much was attributed to rumors.
  • How much was a blind attribution, `sources said,’ or `the news organization has learned,’ without any further identification of the source.
  • How much offered even minimal information about how the source would know what he or she was revealing (“a source familiar with the investigation”), but did not signal what if any bias or allegiance the source might have.
  • How much described in some manner the source’s official affiliation (a Justice Department official, a Capitol Hill source)
  • Finally, how much described what side of the dispute the source was aligned with, such as a friend or supporter of the President, a Republican source, or a lawyer for Linda Tripp.

The overwhelming plurality of the anonymous reporting in the mainstream press (43%) was essentially blind. It said simply sources said, or our news organization has learned, offering no effective characterization of the source.

Another 16% of the time, the sources were characterized as simply being knowledgeable in a fairly vague way, such as sources “familiar with the situation” or “close to the investigation.”

Taken together, that means that 59% of the time the sourcing was quite vague, offering no sense of where the source or sources worked or what slant there might be to the information.

Only 17% of the time did the press characterize anonymous sources in a way that offered at least some guidance as to the sources’ allegiances, describing the source as “a supporter” of the President, “a Democratic” or “a Republican source,” “a friend” of someone, or someone “close to” someone else.

And 13% of the time, anonymous sources were described in a way that offered a glimpse of where the source worked, such as “a Capitol Hill source,” or “a Justice Department source” but did not necessarily offer much guidance as to the source’s bias or allegiance.

At least explicitly, the press did not engage much in passing along rumors and innuendo in the days studied. Only 4% of the cases were “rumors” or “it is believed” or similar attribution cited as the source.

Vague characterization of sourcing may be one of the reasons that the public registers irritation with press coverage of this story. Certainly, some anecdotal evidence would suggest that. Washington Post ombudsman Geneva Overholser offered voice to some of these complaints in one of her columns. “Sources said, sources said…what sources?” Overholser quoted one reader as complaining. “Just who are these informed sources?” asked another.

The leak of the Clinton deposition may have lowered the amount of blind sourcing captured in the study because so much of the reporting by other media was attributed to the Washington Post rather than an anonymous source.