Two traditionally noncompetitive states are polling closer than usual today. In Connecticut, Rasmussen has John McCain just 3 points behind Barack Obama (McCain trails Hillary Clinton by 6 points). From the internals of the poll, Obama is having a little bit of trouble with the middle-class vote, while winning both working-class and wealthy voters. A lot of Hillary Clinton's support is concentrated among older, middle-class suburbanites, particularly in the Eastern half of the country, and it's possible that we're seeing some defections among that group here. Still, since Obama has been polling strongly elsewhere along the East Coast, this is probably also a case where we're observing some random noise.
Meanwhile, in North Carolina, Public Policy Polling has Barack Obama within 3 points of John McCain; Hillary Clinton trails McCain by 5. There are several states where Obama is in the 15-30 percent win probability range right now: North Carolina, Missouri, Florida, and a handful of Western states like Montana and North Dakota. If we think about the Western states as a group, Obama might need to pick and choose two or perhaps three of those states in which he wants to make a serious effort to compete (Missouri will almost certainly be one of them).
Also a correction: yesterday, we listed John McCain's margin over Obama in SurveyUSA's Nebraska poll as 12 points. In fact, the margin is 9 points. It looks like we had caught an earlier, unweighted version of that poll.
Monday, June 2, 2008
Today's Polls, 6/2
-- Nate Silver at 3:05 PM
Labels: connecticut, north carolina, today's polls
Subscribe to:
Post Comments (Atom)
17 comments
So you are treating polls with 3rd party candidates the same as the ones with just the two candidates from the major parties?
What about polls that include only one of the two "major" third party candidate? (Like the PPP poll which included Barr but not Nader)?
SurveyUSA has a new Minnesota poll out too.
North Carolina has 15 electoral votes. I'd like to see Obama focus there. Dems control the state apparatus and a majority of their Congressional Representatives (7 of 13) are Dems.
Anon #1,
I consider a poll valid so long as it consists entirely of declared candidates for the race. So I'll include a poll with Barr, or Nader -- it's up to the pollster whether they want to include them specifically. But I wouldn't include one with, say, Jesse Ventura running as a third party candidate.
Anon #2,
Yeah, that missed my deadline by about 5 minutes. I usually won't re-run the numbers twice in a day unless there's something earthshatteringly interesting going on, and that SurveyUSA result (O +5) was pretty consistent with their other polling of the state.
Poblano,
Don't you think it would be more fair to acknowledge that Rasmussen is polling an anti-Obama electorate?
Since February, Rasmussen has been polling a group with >44% unfavorables for Obama. That's in contrast to every other pollster who releases fav. ratings.
Nate, but wouldn't that skew the results, seeing as how the third party candidate draws support from one candidate, and it is usually overstated? For instance, this poll gives Barr 6% in both cases, which, almost assuredly, is not going to happen. Most of those that stated for Barr will eventually end up voting for McCain and as such, the 3rd party candidate skews the poll toward Obama in this case.
Jeez, you guys didn't complain this much at the SurveyUSA poll that had Obama +21 in Connecticut.
The PPP poll is a great one for Obama, even if Barr's support is overstated (and I suspect it is).
Exit polling from 2004 in NC showed that black voters made up 26% of the electorate and Kerry won 85%. This PPP poll has only 20% African-American and Obama is only winning 72%.
Let's make more realistic estimates. Black voters are 28% of the electorate and Obama wins 90% and white voters are 69% of the electorate and Obama wins 30% (3 points higher than Kerry). And let's give Obama 50% of the people of other races. That gives Obama more than 47% of the vote.
North Carolina is certainly in play. And if NC is in play, then VA certainly is, and so are SC and GA.
Bill in Georgia:
Do you think Edwards made absolutely no difference in NC 2004? If he did, a +3 W for Obama relative to Kerry is more difficult.
Perhaps Edwards didn't make any difference, I'm just wondering.
Bill in Georgia-
PPP addressed that issue. They had been using ~26% AA turnout based on those exit polls.
However, they emailed the North Carolina Board of Elections and it turns out the actual black electorate in the state in 2004 was a little under 19%. Which makes sense since about 20% of the population is AA in NC.
Obviously the NCBE data is far more accurate then the exit poll data.
http://publicpolicypolling.blogspot.com/2008/05/erroneous-exit-poll.html
As a North Carolinian and a former consultant, I'm glad that so much time and attention is being devoted here. I tend to agree that NC is in play, but caution against anytone thinking an Edwards VP selection turns this state blue automatically. Edwards is a much more popular figure nationwide than he is here. He very well may have lost his reelection campaign for senate, and he brought nothing to the table here. That having been said, the two biggest issues of this campaign (the Economy and Iraq) make for an interesting dynamic here, since both play extremely large roles among lower-income rural folk and an abundance of military personnel and their families.
Anon 17:55,
I'm just curious, how would the NC Board of Elections knows the percent of the electorate that is black? Also, is this 19% number supposedly (1) the percent of registered voters in 2004 who are black or (2) the percent of actual 2004 voters who are black?
The 19% is the percentage of those who actually voted in the 2004 election who were black. The BOE knows because they know everyone who voted in the election, and their demographic data as filled out on their registration forms.
This particular example is a very good caveat about exit polls- they can be as wrong as any other polls.
Tom Jensen
PPP
Also to the above commenter asking about us including Barr but not Nader- North Carolina has pretty restrictive ballot access laws and while Barr will be on the ballot in the state, Nader will not.
Thanks for the info Tom!
Besides poll numbers it would be interesting to see how many field workers Obama has in some states. I think NC is a state that obama might focus on.
In Connecticut, Obama will win by over ten, but right now he is probably having real trouble with the white ethnic democratic vote in the working class suburbs around Hartford and New Haven. In the primary Obama lost big in places like East Haven and West Haven, both of which are middle/working class and are about one-half Italian, and suffer from extreme ghettoization inherited from New Haven.
I don't want to call Italian Connecticuters racists (mostly because I am one), but I have noticed a tendency among these voters even in my own family (group the Irish in with them too) to be hostile to African American candidates. There is also a "sheep" vote here, which is how Lieberman won in 2006. If the media goes sour on McCain, however, these voters may finally rush to Obama.
Obama can overcome this as long as he can debunk those Muslim rumors and such.
Post a Comment