bo:
Raw state counts are absolutely meaningless. Popular vote was pretty close, but still meaningless. Electoral college matters, and it came down to 50.8% to 48.7% in Ohio, which is quite close.
And the problem with your media narrative is that you're diagnosing the wrong cause. Yes, voters thought the media was biased for Kerry in '04, but according to Pew the media's coverage also favored Obama in '08:
http://www.journalism.org/2008/10/22/winning-media-campaign/
And in '12:
http://www.journalism.org/2012/11/02/winning-media-campaign-2012/
NB: the above links take care to point out that they're not necessarily talking about *bias*, because differences in coverage may reflect real differences between candidates. But we can at least say the differences probably don't reflect a bias for McCain and Romney, which are what you'd need to support your "media is biased for close races" claim.
Lastly, w.r.t. 2012 polling: No, it was really just Romney's internal polls and the 'unskewers' who hilariously screwed up. (Every pollster and every Democratic operative has probably jerked off at least once to the video of Karl Rove desperately spinning on election night.) Hell, 538's notoriety comes largely from them using polling to predict the 2012 results, which wouldn't be possible if the polls were hilarious screwups.
)
Help! I'm new!
)
)