Tuesday, October 21, 2008

Pew Update

Registered Voters
Obama 52, McCain 38

Likely Voters
Obama 53, McCain 39

Most importantly sample size is over 2,000. Very large for a national poll. I believe that's double, even quadruple, a lot of the other nationals. Sample size directly increases accuracy.

As I covered before I consider Pew among the best national pollsters. They've been in the business for a long time and don't seem to have a partisan axe to grind. When you look at the state polls it's hard to imagine Obama only being up 4 or 5% national. How can he be competitive in NC, IN, VA, etc? I can't imagine those previously red states are somehow unique on a national level. Clearly Obama must have very strong support nationwide to be pulling ahead or even in those states. +10 to 14 would make a lot of sense given the local state margins.

Of course then you have OH & FL. Perhaps they just don't follow the national trends. FL especially is a very artifical demographic. Lots of snowbirds, retired folks, etc. Their results just may not be in line with national trends. OH is a bit harder to figure. I would think Obama should be seeing a solid lead there right now given the situation in other battle grounds and national polls. I'm not too familar with the demographics of OH though... perhaps the southern part of the state is really breaking overwhelmingly towards McCain and showing strong signs of likely turnout. That would explain it.

Looking at the racial balance in this poll doesn't seem to make a lot of sense. Obama's lead might, in fact, be bigger:

Pew 10-21:
White 81%
Black 10%
Hispanic/Other 9%

Actual 2004 demographics:
White 77%
Black 12%
Hispanic 8%
Other 3%

No comments: