The Volokh Conspiracy currently is hosting Timothy Groseclose, author of a book purporting to quantify liberal bias in the media. Basically, what he does is attempt to measure by "objective" (I'm going to problematize that below) criteria, how liberal or conservative a given media outlet is, then compare it to a perfectly "centrist" position. The upshot is that the media, under Groseclose's analysis, leans left -- excepting some niche conservative publications (like The Washington Times).
Groseclose provides three different metrics by which one could "objectively" measure the relative bias of media outlets. The first is with reference to think tanks, the second talks about "loaded phrases" (like "death tax" versus "estate tax"), and the third looks at the mention of "equally true facts", one liberal-leaning, one conservative-leaning. The media "scores" on this are then compared to a normalized ranking of Senators and Congressmen, so, we could say, the way The New York Times uses these think tanks, phrases, and facts, is most closely akin to that of Sen. Joe Lieberman (I-CT).
This sort of empirical analysis isn't really my area, but even I spotted some problems (actually, while running a google search for this post, I discovered I spotted some problems with the book's predecessor article in the first month of this blog. Ah, the memories). First, there are some question design aspects that worried me when I took my own "PQ" quiz (it labeled me as liberal -- which is absolutely accurate -- but I have no idea why the questions informed me how Democrats and Republicans voted except to push the data to extremes). Second, I don't think the metric disaggregates enough: my own line is that the media is socially liberal, economically moderate, and hawkish on foreign policy, and that hypothesis isn't really examined.
But the most gaping one is that it doesn't tell us who is leaning towards whom. A relative affinity between what think tanks the media cites, and what think tanks a Democratic Senator cites, may show the media is biased towards liberals. But it could equally show that liberals are more likely to cite independent, dependable think tanks of the sort relied upon by the media. In other words, we can use politicians as independent variables and use them to judge the media. Or we can use the media as an independent variable and use it to judge the politicians. Groseclose's methodology doesn't have us distinguish.
And there's a solid case to be made that the latter is a better explanation. Take Groseclose's own example on the "loaded phrases" front: "death tax" versus "estate tax". Republicans tend to use the former, Democrats the latter. Hence, under Groseclose's methodology, if the media tends to use the latter, it is exhibiting liberal bias. But another explanation is simply that "estate tax" is considered the more neutral, explanatory term, which is why the media uses it. "Estate tax" is, in fact, not loaded in the way that "death tax" is (or something like "plutocrat tax" would be). Practitioners don't talk about the "death tax", they talk about the estate tax. "Estate tax" is the name of the term in U.S. tax code dating back to 1916, "death taxes" refer to something else entirely. The term "death tax" didn't gain any mainstream traction until the Gingrich speakership, when it was part of a specific effort to muster up voter anger against the "estate tax".
Given all that, it makes perfect sense for the media to favor "estate tax" over "death tax". Groseclose's methodology would have us adopt the absurd conclusion that, because both the media and liberals (but not conservatives) refer to the estate tax by its historical and official name, that's a problem with the media's objectivity. One could hardly impoverish the meaning of "media objectivity" more if one tried.
What is most weird about Groseclose's methodology is that, for all its protestations of scientific rigor, it is really dependent on a very blunt sort of relativism. The methods he use only make sense if one effectively believes there is no way of determining which think tanks or more reliable or less reliable, which phrases are neutral and which ones are inciting, which facts are salient and which are irrelevant. Why not? Well, probably because the most obvious way of measuring that would be with reference to how a supposedly neutral arbiter, like the media, and then instead of concluding that there exists a liberal media bias, we instead conclude that, as the saying goes, "reality has a well-known liberal bias."
But if we dispense with the notion that we can evaluate the neutrality of various phrases or researchers on their merits, then it's all politics -- liberal ideas and conservative ideas, liberal phrases and conservative phrases, liberal facts and conservative facts. And the media tends to align itself with the liberals. If there's no truth behind it, then that's no strike against the right. It just marks the media as an adversary.