The liberal group Media Matters published a report earlier this month seeming to show that conservative opinion columnists get far more play in daily newspapers than liberal ones do. The piece attracted a lot of comment in all corners of the blogosphere, which you are welcome to search for and chew on if you wish. I don’t actually disagree with the finding ? it’s probably true that conservative columnists are more carried and more read than liberal ones, though I suspect it has to do with factors such as editors offsetting liberal editorial pages with opposing conservative editorial views, as was the original notion of the “op-ed,” rather than the suggested, ominous conspiracy of right-wing media owners.
But even though the finding is not inherently implausible, this particular study appears to be highly questionable. The result is to exaggerate quite a bit. It relies not on clip counts or database searches but on surveys of daily-newspaper editors. Many respondents, it appears, did not provide sufficient information about which individual columnists the papers run regularly (roughly each week) or occasionally (a couple of times a month). That left Media Matters to guess based on syndicates subscribed to, web searches, and such, which ain’t pretty. Moreover, some of their decisions in coding columnists are just odd. Mort Kondracke is not a conservative. Dan Rather, Cokie Roberts, Neal Peirce, and Tom Friedman are not centrists.
How did I come to the conclusion that the methodology might have gone haywire? Because a NC editor forwarded me today’s release of the state-specific results of the Media Matters report, which listed me as in the top-10 of syndicated columnists in North Carolina. That would be kind of neat, except for three things: 1) their count of the number of dailies that run my column is way too low, about a fourth of the real total; 2) they list me as a centrist, not a conservative, which obviously hurts their own agenda and is just a boneheaded error (though I am, of course, highly reasonable and centered); and 3) Media Matters specifically notes that it excluded columnists unless their work was syndicated to newspapers in more than one state. My column is about North Carolina and is only syndicated here (running in about 50 papers a week, on average, some dailies and some not).
In other words, they got my newspaper count wrong, my political leaning wrong, and I wasn’t even supposed to be included in the survey in the first place. But other than that, a penetrating analysis.
