John, seeing the relative balance of complaints as evidence that there is no bias is faulty. It utterly disregards the self-selection of the sample.

assume a paper is on some (indefiniet) point on the liberal-conservative scale. Those readers who are at about the same point will of course not complain about it. Those readers who are very far from the papers position - whatever that position may be - will not complain, since they will be just about non-existent (and those few who are readers know full well they are reading "the enemy" as it were, and will not bother).

What you have left is a cloud of moderately dissatisfied readers extending in both directions - no matter where on the scale the paper happens to be. Even a very conservative publication likely receives accusations of "pandering to the pinko commie left" by people somewhere to the right of Attila the Hun. And very leftist publications are surely regularily still accused of "betraying the proletariat" by people who think Bakunin was a right-wing wimp.

Having complaints from both sides in thus normal and expected no matter where you are or in which direction you are moving (if any). It is just a lousy predictor, pure and simple.

John, I recommend that you read this article instead.
http://rhetorica.net/bias.htm

What timing. I was just reading this article from the
Guardian and thinking "Wonder whether John has
seen this yet?" :)

More on perception of reporting quality and bias in
the US:

http://media.guardian.co.uk/site/story/0,14173,1224127,00.html

The stuff about political bias is towards the bottom.
I found this quote and the details below it interesting:

"Just 8% of journalists believe the media have been too
critical of the US government compared with 34% of the
general public."

Telsa