Two recent “studies” by the Media Research Center, a conservative media watchdog group, shine a bright light on the questionable techniques and absurd assumptions that guide the MRC's attempts to “prove” its claim that “liberal bias” is rampant in the U.S. news media. Looking through a funhouse mirror that renders everything -- even the facts themselves -- as manifestations of insidious bias, the MRC had no trouble finding what it was looking for.

MRC research director Rich Noyes summarized a May 9 “study,” titled "Extreme Conservatives vs. Unlabeled Liberals," as follows:

In the six months since November's elections, network reporters have zeroed in on “conservatives” -- especially “religious conservatives” -- as an energized and unwelcome force in American politics. As TV told it, George W. Bush won re-election because of strong support from “social conservatives” and would pack the courts with “conservative” judges. It was “conservatives” who pushed Terri Schiavo's right-to-life case, and “conservatives” like Tom DeLay and John Bolton were embroiled in controversy.

It's true conservatives have been making a lot of headlines, but even as the networks painted the right side of the spectrum as ideological, and even a tad fanatical, reporters rarely used ideological terms to define liberals. Since Election Day, network reporters branded politicians or groups as “conservative” 395 times, compared to 59 “liberal” labels, a greater than six-to-one disparity.

The basic premise of this “study” -- that if there are more mentions of the word “conservative” than the word “liberal” in a given period, then the news must be “biased” against conservatives -- is so ridiculous that a fourth-grader could pierce its logic.

If precisely the same number of actual conservatives and liberals had been discussed in the news, and conservatives had been identified as such while liberals hadn't, the MRC might have a legitimate gripe (though even this criticism would presume, as the MRC seems to, that “conservative” and “liberal” are inherently derogatory terms). But the real reason there are more mentions of “conservatives” than “liberals” is obvious: there has been more news about conservatives. In the wake of the Republican electoral victories, conservatives both in and out of government are wielding influence and getting more attention. One can't help suspecting that if the results of the MRC's Nexis searches had turned out the opposite of what they did, the MRC would be alleging that the greater repetitions of the word “liberal” showed that conservatives were outnumbered in the media.

Indeed, as Media Matters for America has documented, the news media have granted conservatives more opportunities than liberals to speak in a wide variety of network news forums. On NBC, Meet the Press consistently features imbalanced panels that favor conservatives; interviews on the Today show in April featured three times as many conservatives as liberals; and 19 Chris Matthews Show panels skewed right in 2004, while only 7 skewed left. In the 15 weeks following the 2004 presidential election, the CBS Evening News featured 65 clips of Democratic officials or commentators representing progressive organizations and 83 clips of Republican officials or commentators representing conservative organizations, not including President Bush; and on January 19, CBS anchor Bob Schieffer acknowledged that CBS' Face the Nation hosted more Republican than Democrat guests since the presidential election. Media Matters has noted imbalances in cable news coverage of political events as well, including the 2004 presidential debates and the inauguration.

The MRC highlighted a few examples of imbalanced labeling, presumably those that it considered most egregious, but even these examples show how faulty its logic is. To demonstrate the “imbalanced approach,” Noyes wrote: “On the April 26 Today, Katie Couric introduced a debate segment by branding just one side: 'Dee Dee Myers was President Clinton's first White House press secretary, and Tucker Carlson is a conservative commentator and host for MSNBC.' Were we supposed to believe Myers is non-ideological?” Apparently, the MRC believes that by labeling Myers by her affiliation with the Clinton administration, NBC was trying to keep Myers's ideology a secret. Of course, former officials of both Republican and Democratic administrations appear on television all the time, and are nearly always identified with the administration they served in, not as “conservatives” or “liberals.”

The MRC was also angry that the lobbying group USA Next was identified as conservative, while the AARP was not identified as liberal. As Media Matters has noted, USA Next is little more than a Republican front group funded by the pharmaceutical industry to attack Democrats and press Republican causes. The AARP, on the other hand, represents tens of millions of seniors on a wide variety of issues, in addition to offering services like health and life insurance. Its CEO worked on President Nixon's re-election campaign, and the organization endorsed the Bush Medicare prescription drug plan; calling the AARP a “liberal” group would hardly be accurate.

If it's not Republican spin, it must be liberal bias

The MRC's April 27 “study” on coverage of the Social Security debate, "Biased Accounts: Networks Guarantee Liberal View of Social Security" -- under the auspices of the MRC's Free Market Project, whose mission is “devoted solely to analyzing and exposing the anti-free enterprise culture of the media” -- shows the same pattern one usually finds in MRC's attempts to document its endless claims of liberal bias. No matter what the facts are, the MRC can design a study with the scales weighted to show “liberal bias”:

This study of 125 news stories on Social Security between Nov. 15, 2004, and March 15, 2005, found four out of the five major networks biased toward liberal talking points. CBS and CNN had almost three times as many liberal stories as conservative. Overall, liberal talking points outweighed conservative ones by a margin of 2 to 1. Reporters favored extreme examples that made liberal points, while failing to explain key economic terms and concepts that would inform the debate.

A look at the supporting evidence shows the presumption that drives this “study” is the same that underlies virtually all of the MRC's work: Any news story not dominated by unadulterated Republican spin is, by definition, a case of “liberal bias.”

If this is your starting point, it is not hard to find examples that support your pre-determined conclusion. If its “study” of Social Security coverage is any indication, the MRC apparently believes that even any fact that does not coincide with Republican spin constitutes a “liberal talking point” and proof of “bias.”

For instance, in the study, MRC coded any mention of the transition costs of President Bush's plan for private investment accounts as a “liberal talking point.” The flip side of this contention is that, according to the MRC, when a news organization fails to report Republican spin as fact, it has shown bias. For instance, the report cited the following as an alleged example of bias: “On Feb. 12's 'Evening News,' CBS's Russ Mitchell said, 'Mr. Bush said he's open to any good idea to fix a system he claims is heading for bankruptcy' (emphasis added).” CBS' attribution of the bankruptcy argument as a claim of Bush's and not as simple fact, according to the MRC, shows CBS' bias. Of course, the idea that Social Security will be “bankrupt” is not merely a point of contention, it is highly misleading -- even under the pessimistic projections of the Social Security trustees, the system will not go broke (it will be able to pay, at a minimum, between 70 and 80 percent of promised benefits, even if nothing is done to change it). Anything less than uncritical Republican spin, even when that spin is clearly misleading, constitutes “liberal bias” as far as the MRC is concerned.

The MRC looked for conservative arguments in favor of Social Security privatization, then claimed that the absence of those arguments from news reports constitutes bias. The study quoted conservative activist Stephen Moore, an advocate of privatization, to contend that “personal accounts would reduce the burden on the government and eventually create surpluses rather than more debt. But in unbalanced stories, this point was absent.” In fact, even the Bush administration has declined to endorse this claim; the administration has admitted that private accounts will do nothing to address Social Security's long-term solvency. Moore is apparently assuming spectacular future stock market gains to make his prediction. But even if you agree with Moore, his is a highly contested, speculative argument. The MRC's claim that not including it in a news report makes that report “unbalanced” exposes the assumptions that make its “study” so absurd.

The MRC has not posted complete data for the Social Security “study” on its website, so it's impossible to know what criteria the group used to classify statements in news stories it examined to arrive at its conclusions. Nor does it provide a list of guests, sources, institutions, or ideas that were unfairly labeled “conservative,” or a similar list of guests, sources, institutions, or ideas that ought to have been labeled “liberal” but were not.

But the absence of data has not prevented conservative media from citing MRC “research” as though it said anything meaningful. Free Market Project chairman Herman Cain summarized the Social Security coverage “study” in a May 10 National Review Online article. And on the “Political Grapevine” segment of the May 10 edition of Fox News' Special Report with Brit Hume, Fox News Washington managing editor Hume reported the “results” of MRC's “study” of the number of times media labeled conservatives versus liberals.

The MRC is certainly free to cry “liberal bias” at every news report it sees. But when it labels something a “study,” it is presenting its analysis as more objective, and should thus be held to a higher standard. Using coding rules that people who disagree with the MRC would accept as objective, and publicly releasing the data on which it bases its conclusions, would be two good places to start. But again and again, the MRC comes back to the same point: Any fact, observation, or argument that does not precisely coincide with the Republican viewpoint constitutes “liberal bias.”