Bush Election Improvement Came in the Cities

CNN has some pretty fun exit poll data here.  As both an engineer and an analyst at heart, I enjoy plowing around in data to find new conclusions.

One interesting thing I found was where the vote for Bush and Kerry came from in terms of urban vs. rural locations:

VOTE BY SIZE OF COMMUNITY
BUSH
KERRY NADER
TOTAL
2004
vs.2000
2004
2004
Urban (30%)
45%
+10
54% 0%
Suburban (46%)
52%
+3
47% 0%
Rural (25%)
57%
-2
42% 0%

Because people are thinking of the red/blue state map, and even more the red/blue county map, they want to portray the red/blue split in part as an urban-rural split.  This is reinforced by the perception of Bush as the NASCAR loving gun toting country redneck and Kerry as the overeducated stiff urban intellectual. 

The problem is that the exit poll results don't necessarily support this.  Note two things:

  1. The suburbs drive the result.  Rural and urban basically cancel each other out, with a larger urban than rural population but a larger relative Bush lead in rural.  In both '00 and '04, the suburban vote split closely mirrors the total tally.
  2. Bush won the popular vote because he improved about 3 points and added 4 million votes to his differential.  This table says that nearly all of that came in the urban vote, with a whopping 10 point improvement that drove a 3 point improvement in the total.

Because a lot of this urban improvement occurred in Blue states, the electoral college margin was still close because he ended up closing the gap in blue states rather than flipping many.

Assymetrical Information mines the same data to demonstrate the surge in Latino support for Bush.  Interestingly, it also shows that Bush led among both high school and college graduates (defeating some of his stereotype).  Kerry, on the other hand, carried both high school dropouts and post-grads (ie, the under and over-educated)

UPDATE

Welcome to Professor Bainbridge readers.  If you are not burned out on election news, here is my winner for the most over the top post-election article, and my response.

UPDATE #2

Confirming data from New York City via The Galvin Opinion