Some months ago I was working on a paper on media myths and the narrative imperative, particularly as relates to disaster response (for instance, the prevalent media myths of disaster panic and looting when far more common responses to disasters involve increases in pro-social and altruistic behavior), and in the course of my research I came across a fair amount of information about the ways in which people select and pay attention to news media. It will come as no surprise that most people seek out information that confirms their own beliefs, while avoiding dissonance -- a tendency that has been reinforced by increasingly partisan television news offerings and new media such as the internet (with the political ramification that it's now easier than ever for politicians to "preach to the choir" but more difficult than ever to "convert the flock," making it quite difficult anymore to achieve bipartisan support for anything).
As I recall, the numbers were something like this: roughly 20% of people surveyed reported preferring news that reinforced rather than challenged their beliefs, while roughly another 20% prefer to check their news against other news sources with the opposite bias. (No information that I remember on the preferences of the remaining 60%, except for the indication that some number of them have opted away from the news altogether, preferring entertainment to politics.) (Interestingly, while media bias apparently increases with more of those latter "conscientious" consumers giving business to purveyors of either extreme, there is some evidence that those consumers are able to get a more accurate picture from increasingly biased news sources on both sides than from a single, "balanced" news provider.)
I wonder how the numbers would stack up in a survey of readers rather than news consumers: how many would prefer to avoid dissonance and stick to escapist or formulaic or "comfort" fiction, or to nonfiction that confirms their biases and beliefs; and how many would report preferring fiction that challenges them, perhaps makes them uncomfortable, or nonfiction on topics about which they know little, or by authors with biases different to their own. And of those who read both the comforting and the challenging, what percentage of their reading falls in which category. (I'd also interested in seeing a comparison of self-reports to actual reading habits.) And, for that matter, how many people report not reading at all.
While I don't have access to my sources right now I believe the numbers I mentioned came from one or both of these:
Baum, M. A. and Groeling, T. J. (2010). War Stories: The Causes and Consequences of Public Views of War. Princeton, NJ: Princeton University Press.
Xiang, Y. and Sarvary, M. (2007). News Consumption and Media Bias. Marketing Science, 26(5), 611-628.