Statistically Bizarre

This is a real puzzle. I've been thinking over the following contradiction over the last several days. In looking at the polling data and analysis of several posts about the last election, there appear to be some real contradictions. Here's what I mean.

1) Chris Bowers points out that "swing voters are becoming a myth." By this he means that the electorate is becomings so polarized that the middle is unimportant. In his quote from Teixeira's analysis, the size of both parties coalitions are about equal, but the Republicans had higher approval ratings for Bush than the Dems for Kerry, hence the Bush win.

2) New Donkey logically replies that in a close electorate, every constituency really is "swing." All you need to do is shave off a few points of any group's support and it changes the outcome of the election. He also claims that Chris is ignoring independents. I don't think this is fair, because if you recall the exit polls Kerry won among independents and moderates by a healthy margin.

3) Bowers responds by pointing out that a) voters are less likely to change their minds, b) are less likely to turn out to vote, c) the real gap between D's and R's is that the R's really, really loved Bush and the D's just like Kerry, d) probably because of the above, the Republicans strategy appears to be to mobilize their own base rather than going for the middle, therefore e) Democrats should do the same by focusing on a mobilization strategy. What I wonder is how we do this without becoming totalitarian fanatics like the Republicans.

Okay, here is what confuses me:

4) Alan Abramowitz reveals that the Democrats in fact out-mobilized the Republicans in 2004. We turned out our base better.

Okay, let me get this straight. The two partisan bases are of equal size, the D's turn out their base better, they win among independents, and they still lose? I can't think that the Republican loyalty to Bush was that decisive a difference, or am I wrong?

So how in the world did we lose? The only thing I can figure is that either the Democratic base really is smaller than the Republicans, and all we did we close the gap between the two (I'd have to look at the polls to see what their sample was), or that we are losing a substantial part of low SES voters. Either possibility is worrisome, but I'm still scratching my head about this one.