Finally, A National Poll That DOES Matter

It amazes me how often t.v. pundits and newspapers report national head-to-head polls, especially this far out. Give t.v. pundits credit for usually reporting that these national polls are irrelevant because they ignore the electoral math.

Finally, a state-by-state poll by a reputable firm that DOES matter, atleast as far as polls eight months out matter.

The lede:

SurveyUSA's 30,000 sample poll (600 voters per state) gives a slight edge in the Electoral Vote race to Obama at first glance, and a significant electability edge upon further analysis.

It shows Obama beating McCain, 280 - 258, while Clinton beats McCain, 276-262.

Both match-ups show either Democrat winning California, Connecticut, Delaware, Hawaii, Illinois, Maine, Maryland, Massachusetts, Minnesota, New Mexico, New York, Ohio, Rhode Island, Vermont, Wisconsin and the District of Columbia, for a total of 202 electoral votes.

How they add to those totals is very interesting. Clinton adds Arkansas, Florida, New Jersey, Pennsylvania and West Virginia, for a total of 74 more. Winning the Big States, just as she says.

Obama adds Colorado, Iowa, Michigan, Nebraska (split), Nevada, New Hampshire, North Dakota, Oregon, Virginia, and Washington, for a total of 78 more. Doing better in smaller states, Midwestern states and Western states, Purple States just as he says.

What's MOST interesting about these results for Obama is that Florida (-2 points), New Jersey (virtually tied) and Pennsylvania (-5) are states that he hasn't campaigned in. Yet. Head-to-heads also show Obama down only 2 points in North Carolina, another state where he hasn't campaigned. That shows 78 additional swing states that Obama could compete in.

To be fair, Clinton is competitive in Michigan (virtually tied) and Washington (-2), but much less so in Colorado (-6), Iowa (-5) and Oregon (-5), and don't expect her to compete at all in Nebraska (-27), Nevada (-8), New Hampshire (-8), North Dakota (-19) or Virginia (-10). That's 51 potential swing votes for Clinton, but one would expect that being so well known already, there's only modest gains for her to be made.

Aside from Florida (and Arkansas for Clinton), neither candidate really competes in the South. Which makes the whole "Will the South elect a black President?" question the Clinton's hint at moot.

With the exception of Arkansas, it also blows a pretty big hole in Clinton's "I can win his voters, but he can't win mine."

Quite to the contrary, Obama does significantly better in Western states and the Midwest, where the other big question, "Do voters want the Clinton's back in the White House?" looms apparent.

(Note: I've been staying away from Clinton-Obama stories, but I thought this might be a good story on poll analysis. Feel free to disagree with me, but before anyone who doesn't like what they read lashes out at thepollster or their methodology, SurveyUSA was only 1 point off on the Foster-Oberweis race, and out of 26 presidential primaries they've polled so far, they've had a median error of +/- 2.0 points). Compared to 5.0 for Zogby (17 polls) and 7.0 for Rasmussen and Mason-Dixon. These guys are to national polling what Dave Fako is to Illinois. No offense, Dave, I know you do business elsewhere)

X-posted at Yellow Dog Blog.