Agreed. Numbers aren't biased. If the numbers are used based on solid statistical and theoretical methods developed prior to the observation, you really limit the ability to bias things. Any bias could easily be found through systematic analysis of the predictions. Any systematic analysis of Nate Silver's methods (or the ones used by a couple of University Profs at places like Emory and Princeton) would show very limited bias. Conservatives are so jumpy about somebody telling them something they don't want to hear, that a person must be "biased" simply by telling them what they don't want to hear, even if it is true. After the election, when I pointed this out, I was told that if the situation was reversed, you would see liberals acting in a similar manner. However, this thread is evidence against that point. Most responses seem to be that the numbers make a decent amount of sense considering what we know and don't know. While I think much of partisan behavior is really a mirror image, this is one of the instances where that is not true. It is the difference between the awful methods and highly questionable assumptions of "unskewed polls" versus the solid methods and limited assumptions of Nate Silver (or other statistical prognosticators). And as would be expected, better methods equal far better results.