Stephen, I would only warn you that the “Project for Excellence in Journalism” is not always an excellent source of journalism studies. To be precise, they produce what we’ve called Swiss-cheese studies. Their studies are not comprehensive, but a series of little snapshots making random selections of certain hours of TV content and not others. Look at their methodology page.  For MSNBC, they coded two out of these four programs per night: Tucker, Hardball, Countdown, and Scarborough Country. Obviously, if you only code frenzied Keith Olbermann and Chris Matthews, you’d get a much different result than if you analyzed Tucker Carlson and Joe Scarborough. (Or at least we can hope.) Then they only watch the first half-hour of each program, not the whole thing.

Then notice they didn’t actually watch TV coverage, but just read transcripts, and squeezed down the sample further: “For both newspapers and the television transcripts, searches were conducted using the last names of the candidates that appeared in either the headline or lead paragraphs of the story. The reason for searching for the names in the headline or lead paragraphs was to determine the number of stories that focused on the candidates. Another option would have been to search for mentions anywhere in the article or transcript which would have yielded more results, but would have also included many stories where the candidates were not central to the story.”

Studies like these might give the public a vague sense of TV coverage trends. (For example, I think it’s right that the Democrat presidential race is getting more publicity than the Republicans. Finding the reason for that is a debatable point: is it a natural liberal media attraction? Are those candidates just more doggone interesting? Or is there simply a belief that the Democrats are the early slam-dunk winners in the media’s conventional wisdom factory?) But these PEJ numbers also suggest the pitfalls of relying on pollsters who say they polled across the United States but in fact only polled in 13 states. As Brent Bozell put it, PEJ studies can be “about as comprehensive with statistics as James Carville has hair.”