Speedkill

Over at &c. Noam Scheiber is discussing the odd exit polling from the election and disagrees with Mark Blumenthal’s suggestion that Republicans are less likely to talk to exit pollsters because they hate the media more than Democrats:

One problem I see with his theory: As evidence, Blumenthal cites polls showing Republicans more inclined to believe the media is biased against them than Democrats are, and also polls showing more and more people (and therefore more and more Republicans) believing in this kind of media bias as time goes by. That would suggest that exit polls should be getting less accurate over time–since fewer and fewer Republicans would be responding. But, according to Kevin Drum’s chart, the raw numbers have actually been getting more, not less, accurate with each election cycle. (The raw numbers were off by 8.3 percentage points in 1988, 7.2 points in 1992, 6.2 in 1996, 1.8 in 2000.) This makes me skeptical that a loss of faith in the media is behind the screwy exit poll data.

I don’t have the data on hand, but wouldn’t it be fair to say that more and more Democrats have been seeing the media as biased against them? It seems like that position has been increasing, since the Clinton years at least. If that’s so, we could see Democratic responses to exit polling dropping, evening things out. I think if someone wanted to prove Blumenthal wrong on that point, you would need to look at Democratic opinions on media bias and the gap between them and Republicans.