Megan McArdle suggests that changes in the news business – away from financial structures encouraging “hard news” intended to be objective and toward more explicitly partisan media – might change what she sees as a Democratic advantage in media coverage. This advantage she attributes to “liberal media bias” as measured – predictably – by Tim Groseclose, whose methods are so obscure (and, I would add, grossly tendentious) that Megan doesn’t even try to describe them. What Groseclose does not do is attempt to investigate the actual truth or falsity of various claims; instead he merely assumes, implicitly, that liberal and conservative politicians tell the truth with identical frequency and that if claims made by liberals survive media scrutiny better than claims made by conservatives, that proves that the media (rather than the facts) are biased. The same methods would prove a media bias toward neuroscience compared to phrenology.
I think Megan’s analysis of the changes in progress is plausible: we’re going back to an era where outlets have well-marked ideological, and even partisan, labels. It’s possible that the result will be to benefit the Republicans, though Fox News’s declining share of the youth market raises questions about that, and the specific structure of the conservative media market might well contribute, not to Republican victories, but to the spiraling insanity of the GOP, turning off swing voters and electing more Democrats. (Hurrah!)
But as a matter of pure social science, the construct “liberal media bias” has zero independent explanatory power in Megan’s analysis, compared to the alternative construct “reporting.” Saying, “Changes in the incentives of media outlets leading to less reporting and more dishonest partisan hackery will tend to reduce the advantage the Democrats now enjoy” makes exactly the same prediction, while reducing the number of postulated entities by one.
Megan, may I introduce my friend Brother William?