Media Bias(?)

Tim Noah links to a July 2003 Michael Tomasky paper put out by the Kennedy School’s Shorenstein Center study to argue that the conservative media is more doctrinaire than the liberal media. From the paper’s executive summary:

This study of the partisan intensity of the nation’s agenda-setting liberal and conservative editorial pages finds that while the pages are more or less equally partisan when it comes to supporting or opposing a given presidential administration’s policy pronouncements, the conservative pages are more partisan-often far more partisan-with regard to the intensity with which they criticize the other side. Also, the paper finds, conservative editorial pages are far less willing to criticize a Republican administration than liberal pages are willing to take issue with a Democratic administration.

The methodology used in the paper is pretty solid. It compares editorial responses for two liberal papers (the Washington Post and New York Times) and two conservative papers (the Wall Street Journal and Washington Times) on matched sets of issues — the Zoe Baird and Linda Chavez nominations, for example. Noah rightly quibbles with labeling the Post as a liberal paper but concludes:

Tomasky’s findings hold up when you compare just the New York Times and the Wall Street Journal. The Times supported Clinton 37 percent of the time and opposed him 37 percent of the time. The Journal, meanwhile, supported Bush 75 percent of the time and opposed him 3 percent of the time. The Journal opposed Clinton 83 percent of the time while the Times opposed Bush 68 percent of the time. The Journal praised Clinton 5 percent of the time while the Times praised Bush 8 percent of the time.

Tomasky is going to be the new executive editor for The American Prospect, so the right half of the blogosphere might be tempted to dismiss the study’s findings. Some of them are probably not as generalizable as Tomasky thinks they are — for example, Noah points out that editorial civility is likely to be a function of editorial page editor’s personality rather than ideology. However, the final graf of Noah’s piece has the ring of truth to it:

When the Brock piece came out, Chatterbox (then writing a media column for U.S. News) interviewed the conservative commentator David Frum about its thesis. Frum basically agreed with it. “What happens with the liberal press is that there are loyalties to causes,” he said. That’s correct. In Tomasky’s study, the Times editorial page supported Clinton on policy matters 52 percent of the time, a mere 7 percentage points less than the Journal supported Bush. But, Frum added, “[w]ith conservatives, I suspect there is much more of a loyalty to people.”

UPDATE: Let the debate commence!! Andrew Cline, Jim Miller, Jay Manifold, and PowerLineBlog all thake their whacks at the study. Their criticisms amount to:

a) Tomasky’s own rhetoric is biased and nasty; b) The sample size is too small; c) The ten “matched” cases are not really matched; and d) Tomasky is not on the cutting edge of rhetorical analysis.

(a) is correct but irrelevant — what matters are the comparison of cases, not Tomasky’s presentation style. (b) makes little sense — obviously, one would prefer as large an N as possible, but controlled comparison — which is what Tomasky does here — is perfectly appropriate. (c) is a judgment call. I looked at the cases, and they seem pretty comparable to me — but I’m sympathetic to arguments that some of the cases are not parallel. I have no doubt (d) is correct, and it’s probably the best critique, but it doesn’t necessarily vitiate his results.