Anecdotes and Evidence

Paul Rosenberg over at OpenLeft has some deeply wonky stuff up about framing the debate between progressives and conservatives and some theories promoted by Gramsci. In essence Rosenberg frames the "media bias" debate in startling, though accurate terms. The Left sees media bias as inaccurate or incomplete information, the Right sees media bias as any coverage that is not inherently supportive of their positions.

While the left is reality-based, and defines bias in terms of things like (a) spreading falsehoods, (b) ignoring truths, and (c) presenting biased pictures by excluding some topics, stories, sources and points of view while dramatically over-representing others, the right is war-of-position-based, and defines bias in terms of "are you for us, or against us"?- OpenLeft

As Steven Colbert has said, reality has a well-known liberal bias, and hence does not serve the conservative cause well. That is why another note Rosenberg makes is so important. He notes that conservatives depend on anecdote where liberals and progressives depend on tested evidence.

That's the problem the anecdotal approach that conservatives are so fiercely wedded to: there are anecdotes supporting both sides, but there's no sense of what they add up to without using scientific methods, without digging past mere appearances into actual patterns and their causes. - OpenLeft

Rosenberg goes on to demonstrate how statistical analysis shreds all the assertions of liberal media bias which the right provides as even marginally testable.

It's a long post, but if you want a definitive takedown of the "liberal media bias" canard, it is excellent.