REALLY Junky Stats!

Hyman has no love for people in academia . . . unless they say something that happens to support his talking points.

Exhibit A is the recent editorial on a study done by Timothy Groseclose, a political science professor at UCLA, and Jeffrey Milyo, an economics professor at the University of Missouri that claims to have identified a thoroughly objective way of measuring media bias.

Hyman touts this study because it falls in line with the conservative party line on the media: it’s systematically left leaning, a fact that Hyman says “reasonable people” already know from watching television and reading the papers.

The problem? The study, even by its own standards, is thoroughly biased.

To begin with, although Hyman calls the study a “UCLA” study, only two professors (and only one from UCLA) worked on the project. That by itself is not so much the issue as the fact that both men have been fellows at conservative think tanks and receive significant funds from them. A few years ago, the same men paired up to write a piece for the right wing magazine The American Spectator. On top of that, although their study appeared in the respected Quarterly Journal of Economics, it apparently was strongly pushed by one of the journal’s co-editors who is notably conservative himself. (All this raises the question: What happened to the supposed stranglehold liberals had on academia?)

This is especially interesting given the methodology of their media study. In an attempt to find an objective way of measuring bias, Groseclose and Milyo make think tanks the fundamental measure of bias.

The tortured reasoning goes something like this: Americans for Democratic Action is an organization that assigns a score (0-100) to members of Congress based on their votes. The number represents where they fall on the conservative-liberal scale: the higher the score, the more liberal a Congressional representative is.

What Groseclose and Milyo do is count the number of times a member of Congress cites statistics provided by certain think tanks to support an argument. They then assign a 0-100 score to the think tank based on how liberal or conservative the Congressmen who cite it are. Then they count the number of times this think tank is cited in news stories by various television and print sources. The more often think tanks that Groseclose and Milyo have assigned a “liberal” score to are cited by a newspaper or news broadcast, the more liberal the news source’s bias (and vice versa).

If, like me, you found this series of steps a bit hard to follow, rest easy in knowing that the fault is not yours. The study makes no sense.

If you want proof, here you go. According to Groseclose and Milyo’s scoring system:

The Wall Street Journal is the most liberal news source in America. (Of course, the paper has contested the findings by pointing out numerous basic flaws in the study, but you’d expect that sort of spin from a leftist rag like the WSJ, wouldn’t you?)

The ACLU is a conservative organization.

The RAND Institute is a liberal organization.

The NRA is barely to the right of center.

The Council on Foreign Relations is a liberal organization.

Dopey? Laughable? Useless? Yes, yes, and yes, but that’s what Groseclose and Milyo’s formula kicked out, so according to them, it must be true. You can see more of the topsy-turvy results of the study in the excellent analysis of it by Media Matters for America.

Given these bizarre assignations, it’s not surprising that the results are both dramatic and meaningless.

And, as I alluded to above, given the close ties to conservative think tanks these two men have had, their own logic suggests that they themselves have a conservative bias (unless, of course, their odd scoring system pegs the Heritage Foundation as more liberal than the AFL-CIO—a distinct possibility if you take a glance at the other results they come up with).

But it’s the methodology itself that’s most damning. In addition to the twisted logic involved in assigning scores, Groseclose and Milyo gather wildly different sets of data from different news sources and treat them equivalently. For example, they used 12 years of data in analyzing CBS, but only 4 months when analyzing the Wall Street Journal. They also ignore a significant number of important think tanks. Individual authorities and experts cited in news stories are not taken into account at all. And the notion that a reporter citing a think tank somehow equates with a news story being biased is utterly loopy. (For more on the methodological issues with the study, see Eric Alterman’s fine piece on it).

As a rhetorician, I can’t help but scratch my head at scholars whose attempt to analyze communication leaves out both content and context. And the methodology ignores basic facts about rhetoric that should be obvious.

For example, if I’m a liberal member of Congress trying to convince centrist and conservative colleagues to accept a proposal on an assault weapons ban, as a matter of argumentative tactics, I’d love to cite statistics from the NRA if possible, as a way of demonstrating that I was using data from a source that my conservative colleagues could not doubt. But according to this study, my citing of the NRA would make it receive a more “liberal” score as a think tank. Then, when Brit Hume does a story on FOX about opposition to an assault weapons ban and mentions the NRA, the story ends up being coded more liberal (or at least less conservative) because I, a liberal Congressman, cited the NRA in a speech in favor of the assault weapons ban.

I feel like I’m taking crazy pills!

And given how many times Hyman cites the ACLU in his editorials, maybe this study would say “The Point” is liberal. But wait--the study says the ACLU is actually mildly conservative. So maybe “The Point” is actually centrist. And maybe pigs will fly. And maybe there’s snowboarding in Hades. And maybe the moon is made of a creamy Camembert.

And maybe that’s The Counterpoint.

Hyman Index: 3.31