Newspapers remained the most factual, balanced and widely sourced of any news outlets studied. They also changed character less as the crisis shifted in topic over time.
For the first two phases of the study, facts accounted for more than 80% of all the newspaper coverage, including the opinion pages (85% in September, 81% in November). By December facts had begun to slide, though they still accounted for roughly three-quarters (73%) of the reporting.
Opinion and speculation never rose above 5%.
Are there substantial differences between papers, or between large and smaller-sized papers? Generalizations are hard to draw from looking at four institutions, but the numbers do suggest the answer is yes.
The Cleveland Plain Dealer and the Fresno Bee, for instance, relied less on anonymous sources than either the Post or the Times. The smaller papers began with factual accounts making up more of their coverage. Both shifted more toward analysis and punditry over time. The reason was that they began to rely more on syndicated columnists for their coverage, a reflection of the high costs of covering the war with reporters.
Interestingly, even from the start, both the Plain Dealer and the Bee relied little on wire service accounts. Instead, both leaned on staff writers to cover the crisis and often related the story back to events in their local communities.
In November, as the story moved to Afghanistan, the Times and Post began to show their muscle as news gathering organizations with a broad reach of sources.
The Plain Dealer and the Bee, by contrast, became even more local, increasing their reliance on community leaders as sources.
The New York Times was striking for the fact that in December it shifted more heavily toward interpretation. After more than 80% of its reporting was straightforward factual accounts in September and November, that number dropped to 66%. The shift was toward analysis rather than outright opinion or speculation, which never rose above 5%.
Beyond that, what stands out between The New York Times coverage and Washington Post coverage in the periods studied was how strikingly similar they were.
Time & Newsweek
When it came to coverage of the war in the three time periods studied, Time and Newsweek appear to be very different animals.
Across all three time periods studies, Time stuck much more to the facts and was less concerned with analysis or opinion. In September, it was twice as likely as Newsweek to contain factual coverage (62% versus 34%). And Time was only about half as likely to publish pure punditry (13% for Time versus 22% of Newsweek's).
By November the factual nature of Time stood out even more. Fully 71% of Time's paragraphs were strictly factual. That is more than four times the percentage of Newsweek's factual paragraphs, which made up a mere 20%. Newsweek, on the other hand, was nearly four times as likely to offer opinion (19% versus 5% for Time). In addition 60% of Newsweek's November coverage was analysis compared with just a quarter of Time's.
In December Newsweek became a little more factual and Time offered slightly more opinion, though the gap between the two still remained solid. Factual reporting in Newsweek rose to just over a third (38%), while in Time in dropped slightly to two-thirds (65%). The percent of punditry in Time doubled to 11% though there was still a greater percent in Newsweek (17%).
When it comes to sourcing, the two magazines had similar percentages of named versus unnamed sources (roughly 78% in September and between 54% and 69% in November and December), but Newsweek tended to provide more sources per story. Part of the explanation is that Time was much more likely than Newsweek to run short factual sidebars and background boxes that offered no sourcing. In September and November, nearly a third of Time's stories had no source compared to less than one in ten for Newsweek. (The numbers evened out a bit in December, 12% unnamed sources for Time, 9% for Newsweek).
Newsweek, on the other hand, consistently offered more sources for its coverage though these sources may have been offering their opinion rather than facts—in every time period studied.
The two magazines also differed in whom they chose as sources. Across all three time periods, Newsweek looked more to community leaders than did Time. Newsweek's reliance on these leaders continued to rise across time, rising to 19% in December, while Time use of community leaders never rose above 7%.
By November, however, Time became much more focused on what International officials had to say—three times more likely as Newsweek, and in December four times more likely. Newsweek, in contrast, focused more on what U.S. officials had to say.
In the first days following the attack, the network morning shows provided some of the most serious reporting around. It was full of facts and well sourced. Within the various broadcast genres studied, these shows had the greatest percent of factual reporting (74%) and only 12% opinion and speculation. In addition, it was clear where those facts were coming from. Fully 92% of the morning show sources were named, again higher than any other broadcast genre studied.
The morning shows also gave viewers more information from members of the public. On-scene sources, community leaders, families, friends and coworkers, and "people in the street" accounted for roughly half, 49%, of their sourcing in early days.
But if viewers in November were still relying primarily on morning shows for their news of the war, they were getting something quite different. Factual reporting dropped to less than half of the coverage, 46%. Analysis doubled to 26%. Punditry more than doubled.
And instead of offering the highest percentage of sources on the record in television, morning news offered the lowest.
Good Morning America
ABC's Good Morning America stood out among the early shows for its seriousness and adherence to fact. In every period studied, GMA's factualness outweighed that of the other morning shows. In December the gap in factualness between GMA and CBS's Early Show reached 21 points (65% versus 44%).
These findings reaffirm an earlier study of October coverage by the Project, which found GMA's story topics to be much more serious than the other network morning shows.5 Here, the broad list of topics looks quite similar. The seriousness emerged in how the story line was developed.
The Early Show
CBS, on the other hand is notable for its greater tendency toward punditry. Even in September, opinion mongering accounted for 18% of the coverage compared with 6% on GMA and 12% on the Today Show. By November punditry had risen to roughly a third of all the Early Show coverage (32%) and remained at that level in December.
While the Early Show had more talk than other networks, its coverage was more likely to include views that dissented from the administration. Of all the stories dealing with an administration view, 17% contained a mix of views and another 13% mostly dissented. In short, one in three contained dissenting voices compared to 19% on GMA and 17% on the Today Show.
The Today Show
The Today Show on NBC once again emerges as something of a hybrid between the more serious style of Good Morning America and the looser style of the Early Show. It was not as factual as GMA, for instance, but was more so than CBS. It had more opinion mongering than ABC, but less than CBS.
One place where Today did stand out was that it had more sources per story—in all three time periods studied. In September, for instance, a quarter of its stories had at least four sources, compared with roughly one in ten stories on both GMA (13%) and the Early Show (10%).
The reasons for this are hard to know. But it might be worth studying to see if it has anything to do with the synergy NBC has with its cable cousins, CNBC and MSNBC. Do correspondents who know their stories may run multiple times across three networks have more time for more voices when they put those packages together?
Network Evening News
What about those who still turn to the four traditional broadcast network evening newscasts each night, ABC, CBS, NBC and the Lehrer News Hour?
Overall, these shows were less given to punditry, but more likely to use anonymous sources, than other television newscasts.
Their commitment to the story also changed less over time than was true on morning television, which moved away from the story by November, then returned to it somewhat more in the December period.
If one wanted to keep up with the war on terrorism on television, evening news was the most consistent source of information. In general, about two thirds of the evening newscast coverage was factual reporting, and one quarter was analysis. Unlike morning TV, that level of factual reporting remained consistent every month studied.
Evening news also had the fewest sources on the record of all broadcast genres (just six-in-ten sources in September and December, though it spiked to nearly eight-in-ten in November).
Much of the anonymity came from those programs' reliance on U.S. officials for information. Roughly half of the U.S. officials cited on the evening news were unnamed. Another reason may be the fact that stories on the nightly newscasts tend to be shorter than elsewhere.
Jennings, Brokaw and Rather
The numbers for individual programs in this study are limited. Still in this snapshot the three commercial network evening broadcasts looked strikingly similar in their coverage of this crisis so far. They were all highly factual. They all had similarly low levels of punditry. They all had similarly high numbers of sources in each story.
Perhaps, after nearly 40 years of close competition, they have so assimilated each other's virtues that they have become indistinguishable in the way they are put together.
One of the few areas of difference, though the numbers of relevant stories over nine broadcasts are admittedly small, came in how likely each program was to offer viewpoints that did not necessarily support U.S. policy. CBS was the most likely to air stories that contained no dissent whatsoever, doing so 64% of the time when the subject of U.S. policy came up. NBC fell in the middle at 54%. ABC was the least likely to do so, 45%.
The one program that stood out was the PBS Newshour with Jim Lehrer, perhaps the most consistent broadcast program studied. The NewsHour represents a striking mix. It has about the same level of straightforward factual accounts as the other networks. But the program seemed to avoid analysis, in which conclusions are attributed to facts and evidence, and moved more into outright punditry. Roughly 17% of the program during the time periods studied was punditry, more than double that of other network evening newscasts.
This is almost certainly due to the show's format, packages followed by interview segments. It is interesting, however, that these interviews involve less analysis and more opinion.
The Lehrer show also stood out for its use of named sources. In the three periods studied, 81%, 85% and 97% of its sources were on the record.
Although the number of relevant stories was small, The NewsHour was the evening news broadcast most likely to entirely support the administration's stance. Overall, more than three quarters, 77% of its applicable stories entirely supported U.S. policy, the same as CNN's NewsNight with Aaron Brown.
So much for the supposed liberal slant of PBS—at least in the periods covered by this study.
Talk Shows comprise an increasingly large part of the television news landscape. In an effort to reduce its cost-per-minute of news time, even CNN, which once made its worldwide bureaus its hallmark, has fired scores of journalists and added more hours of talk to its broadcast day.
While many talk shows generate relatively small audiences by broadcast standards, the genre cumulatively makes up a key part of the prime time lineup of most cable networks.
For this study, we examined two types of talk programs, the Sunday show lineup on ABC, CBS, NBC and CNN and a sample of weekday talk shows (Larry King on CNN, Hardball with Chris Matthews on MSNBC and Charlie Rose on PBS).
Some journalists have speculated that one reason public approval of early press coverage of the crisis was so high was that even the talk shows seemed more factual than before. "I think even the cable shows themselves are somewhat more information focused," New York Times Washington Bureau Chief Jill Abramson suggested at a Brookings Institution forum. "It isn't just a partisan or ideological debate about the war. That debate really isn't happening…. I don't sense as much that tone on shows like Larry King."6
Was this really true?
For a short time, the answer was a relative yes. In September, just over half the statements on the talk shows (54%) were people offering factual information. The level of outright opinion and speculation was 30%.
Even so, the talk shows in September were nearly four times more likely than the evening news and three times more likely than morning news to engage in opinion mongering.
But by November the talk shows abandoned even this measure of factuality. Factual accounts dropped to just 32% of what was on the talk shows. Punditry surged to 40%. Analysis nearly doubled to 28%.
In December, the talk shows were basically a three way split between punditry, fact, and analysis.
At times, the speculation was as wild as ever. Consider the reaction on CNN's Capital Gang to the idea that the November crash of American Airlines Flight 587 in Queens might actually be an accident.
"If it's an accident, there's never been an accident quite like it," columnist Robert Novak theorized.
"Within 30 minutes of the crash, before anybody really knew anything much, the government was asserting there's no evidence of sabotage, because they are so anxious for it not to be," agreed National Review Editor Kate O'Beirne.
It was curious, Wall Street Journal columnist Al Hunt added, "that just within a matter of weeks or months after this, that suddenly we have an unprecedented crash of this sort."
With universal agreement on the conspiracy, even Connecticut Senator Joseph Lieberman conceded, that, yes, there had been similar air crashes, "but nothing exactly like it."
As for balance, as noted above, the talk shows were slightly more likely than other broadcasts to provide some dissent about Administration policies, though not nearly as much as in print.
But the dissent existed only within a limited range. Sixty to 75% of the viewpoints on the talk shows were predominantly or entirely pro-Administration depending on the time period. Outright criticism of U.S. policy never rose on the talk shows above 7%.
There were, despite the basic format of journalists interviewing officials, clear differences among the four Sunday interview shows studied, ABC's This Week, CBS's Face The Nation, NBC's Meet the Press and CNN's Capital Gang.
CNN's Capital Gang had the most outright punditry. Even in September, there was as much opinion mongering as fact on the program, 38%—the only show to do so soon. By November punditry outweighed facts, accounting for more than 40% of the program in November and December.
The supposedly staid Face the Nation was close behind. The CBS show saw a steady rise in punditry from 24% in September to 41% in November and finally 43% in December.
On This Week and Meet the Press, punditry fluctuated over time but only once (Meet the Press in November) did it exceed 30%.
As a rule, talk shows on weeknights (Larry King, Hardball and Charlie Rose) were even less factual and more engaged in punditry than the Sunday talk shows—except for Capital Gang. Larry King started out the most factual of the bunch (more than two-thirds fact in September compared to 49% on Hardball and 30% on Charlie Rose.)
By November facts took a beating even from Larry King, falling to less than half (49%) of what was on his shows, though that was still more than twice that of from Chris Matthews (20%) and four times that from Charlie Rose (12%). In December, the three shows had flip flopped. Hardball was the most factual (47%) followed by Charlie Rose (31%) with Larry King bringing up the rear (23%).
Apparently it's all about the guest.
ABC's Nightline stands out among media outlets, especially on television, for bucking the trend. As the press grew more interpretive, less factual and not as heavily sourced with time, Nightline moved in the other direction. The program's reports became more factual. The level of punditry declined. More of its reporting was on the record, and the number of stories with multiple sources increased.
Nightline also stood out for being far more likely than any other television broadcasts to provide views that dissented from the Administration's—levels equal with print.
Nightline also gave viewers an international view of the war that others on television did not. In November nearly a third of its sources were international officials, well ahead of any other type of broadcast news show, and more than twice that of the talk shows. By December these foreign voices accounted for more than half (56%) of Nightline sources, compared with 20% on the evening news and a 15% on the talk shows.
Prime Time News Magazines
On prime time magazines, such as Dateline, 20-20 and 60 Minutes II, we can see the arc of the crisis and the return to media normalcy most clearly. In September, news in prime time became highly serious and dedicated exclusively to the crisis. By November, the war on terrorism was losing its place on these shows, and what did air tended to focus on emotional stories about heroes (Rudy Giuliani's exit interview as N.Y. mayor with Barbara Walters and a fire fighter who donated bone marrow) or victims (in this case of con artists preying on survivors of the Trade Center attacks). One researcher, having noted a habit of the shows to assemble everyday survivors in such stories for in studio group interviews, called these stories "victims on risers."
5. "Before and After, How the War on Terrorism Has Changed the News Agenda," Project for Excellence in Journalism, November 19, 2001.