Lies, Conservatives and Statistics
A study that questions the alleged liberal bias of the media gets trashed, but conservatives' favorite study on the issue has problems, too.
By Terry Krepel
When, in late July, a study was released suggesting that George W. Bush was getting more positive presidential election coverage than Al Gore, the Media Research Center wasted no time in bringing out the long knives.
The study was conducted by the Project for Excellence in Journalism and the Committee for Concerned Journalists ("funded by the liberal Pew Charitable Trusts," MRC wants you to know) and was based on one sample week each month over five months in early 2000, focused on media coverage of specific character-related "themes" of the Bush and Gore campaign. It concluded that coverage of Gore tended to focus on negative themes, while coverage of Bush tended to focus on positive themes.
The MRC hit first with a July 31 "Media Reality Check" criticized the study's methodology, its narrow focus and the way the study's results were reported on CNN. MRC chief and Bush apologist L. Brent Bozell III followed up in an Aug. 24 column; the study, he wrote, "skipped entire weeks, even months, when Governor Bush took high and inside fastballs from the national press."
"This study is about as comprehensive with statistics as James Carville has hair," Bozell adds.
More importantly, Bozell notes, this study was publicized "while ignoring nearly every scientific study proving a liberal bias." And who has generated the vast majority of those studies? Why, the Media Research Center, of course.
Conservatives' favorite "proof" of media bias, however, doesn't have the MRC taint. In 1996, the Freedom Forum with the Roper Center released a survey of reporters who cover the federal government that revealed, among other things, that 89 percent of respondents voted for Bill Clinton in 1992.
Because it was done by a group that isn't an obvious right-wing shill like the MRC, the study got, and continues to get, a lot of play. The survey has been cited so much by conservatives that it threatens to become apocryphal. This year alone, Bozell himself mentioned it in a June 8 speech in New York, and the Wall Street Journal's Robert Bartley mentions it in a Sept. 11 column.
Fortunately, the study has been dissected -- and it reveals some serious questions, not unlike the ones MRC raised about the above PEJ study.
Robert Parry of Consortium News reported back in 1997 that out of 323 surveys sent out, only 139 were returned -- a little over one-third. The list of recipients was complied from the press credentials list of the Congressional Press Gallery. And there are a lot more than 323 journalists covering Washington.
The problem? Many conservative journals are organized as non-profit corporations so they can accept tax-deductible donations, Parry writes. And non-profits have difficulty getting credentials from the Congressional Press Gallery.
The result of this was clear when Consortium examined the names of news organizations represented on the the original mailing list and the number of questionnaires they received. Missing from the list were many major conservative journals; The Washington Times got four questionnaires, Human Events one, and The New York Post one. "The other big-name right-wing publications got zero," Parry writes.
The survey "appears to have dramatically undercounted the scores of conservative journalists in Washington, despite their significant influence in setting the national agenda," Parry writes. "Many of these conservatives appear regularly on TV pundit shows and their opinionated columns resonate across the country through conservative radio hosts and on the op-ed pages of newspapers."
This revealed another problem: "Only 60 questionnaires -- or less than 20 percent of the total -- had gone to the likes of The New York Times, Washington Post, Los Angeles Times, USA Today, Wall Street Journal, CBS, NBC, ABC, CNN, PBS, National Public Radio, Time, Newsweek, US News, The Associated Press and Reuters," Parry writes.
"The bulk of the newspapers on the list were regional dailies," Parry writes, "such as The Modesto Bee, Boston Globe, Denver Post, Dallas Morning News, Atlanta Journal/Constitution, Richmond Times Dispatch and San Jose Mercury News." News services for newspaper chains such as Knight Ridder and Newhouse are counted here as well.
But, Parry reveals, "more than 80 of the list's newspapers -- roughly a quarter of the total -- were much smaller, often with only one reporter or 'bureau chief' in Washington." This includes papers like the Sheboygan (Wis.) Press, the Grand Junction (Colo.) Daily Sentinel and the Thibodaux (La.) Daily Comet. Also included here are specialty and obscure journals like Indian Country Today and El Pregonero.
The survey, Parry concludes, "gave much greater weight to the voting choices of reporters from small publications who have next to no influence in the nation's capital. These work-a-day reporters rarely, if ever, appear on TV and their stories concentrate on the humdrum actions of local members of Congress, not on national affairs."
So, to sum up, the Freedom Forum study uses a small sample from a pool not representative of the total Washington press corps, and the vast majority of the reporters who responded focus their coverage mostly on local issues and have no influence on the national political agenda.
But you won't hear that from the MRC. Why? The study's results dovetail nicely with the MRC's agenda, and to analyze it would kill the goose that laid the golden, non-right-wing-sourced egg.
In his Aug. 24 column, Bozell writes: "Simply judge these studies by whether they prove what they claim, and you'll find what any serious journalist examining them knows: the evidence isn't there." Too bad he only uses that principle to attack what he doesn't agree with.