The Official Obama Approval Ratings Thread

[Alcon]

Just for the record, the City of Omaha isn't contiguous with NE-2.  It's only about 2/3 of the district.

[Alcon]

Obama's approval rating is +26 among San Francisco Republicans, and Schwarzenegger's is -7.  Funny.  Tiny sample but weird.

[Alcon]

Mormons don't like Huckabee too much, "hate blinding" or otherwise.  We had numerous approval polls and head-vs.-heads where a lot of Utahans claimed they'd vote Obama vs. Huckabee.  Obama even led Huckabee pretty solidly in one, albeit one that showed the Presidential race improbably close.

Would Utah vote Obama against Huckabee?  I sincerely doubt it, but Huckabee is not especially popular in Utah.  Voters would "come home" gradually, especially because I doubt a Huckabee candidacy would concentrate much on evangelical Christianity.  But otherwise I see it as analogous to Obama's GE performance in West Virginia -- not a tanking, but swimming against an added current.

[Alcon]

Well, if we're going to be cross-comparing pollsters, indiscriminately suborning uni polls, and juxtaposing polls conducted months apart as if it were a current snapshot...what the hell, Artur Davis's polling team can come on board too!

[Alcon]

A system that identifies Nebraska as likely Republican.  You've clearly found the Holy Grail of analysis here

[Alcon]

I'm sorry, px, I wasn't meaning to be as critical as you're making the comment out to be.  I just find the idea that this map is an especially good predictive, or even cross-sectional analysis, to be misfounded.  Even if we were assuming that approval ratings are strongly correlated (which they are sorta), this ignores:

1. That some outfits treat "fair" as disapproval, while people often think of "fair" as "OK."

2. Time.

3. Differences in pollster quality.

4. Differences in push levels.

It's not a bad effort, it's just not anything to base an electoral narrative on.  pbrowser deserves credit for compiling this all, but the extrapolations being made are unfounded.

[Alcon]

I also much prefer polls to "extrapolations" (I'm not sure what we'd be extrapolating if not polls?).  My point is that you're comparing un-alike polls to each other, and then making a theoretical extrapolation of electoral returns based on that.  It's a perfectly fair effort if you don't want to go to the trouble of developing an algorithm to adjust for the passage of time, or a way of managing polls that ask different questions than "strongly approve"/"approve"/"somewhat approve"/whatever.  But, as it stands, this information is poisoned by a number of variables.

[Alcon]

Approval is actually a very good tracker of how well an incumbent performs, generally.  Approval tends to decline among members of the opposite party and ideology as election-time increases tensions.  Other than that, it's genuinely the best projection method we have.

It's Generic GOPer vs. Generic Dem.  I mean, it's useless, but would an accurate presentation of who's leading as of now be any more useful?  It's fun.

[Alcon]

If anything, in order to get an accurate picture of results you had to overweight Democrats and Independents substantially to make up for the extreme conservatism of the rump of the GOP, whose supporters are a lot more cohesive since most Republicans who are likely to vote against the party or support Obama have switched identifications.

 This is a issue with polls that use identification rather than registration as their baseline. It also seemed mostly to apply to state polls more than national ones. A lot of this is because most national polls were by CBS/ABC/NBC/NYT whose polls suck generally.

I don't really understand why this would show up in polls but not at the polls.  I'm very tired, but could you explain in greater depth?

Or just, like, say "read it tomorrow you jerk."  I can do that, too, although I think I'll still probably be confused.

[Alcon]

I say his approvals will be between 0% and 100%...

I think you are too quick to discount the possibility that every single person in America approves/disapproves.

[Alcon]

Plus: A new Mason-Dickson Nevada poll of 400 RV will be out in the next few days ...

YAY for joke polling.

God, the days when that would have been blasphemy...

[Alcon]

Latest newsweek poll!!!

Obama approval:
approve- 69 pc
dissaprove-18 pc

Latest rass poll!!!

Obama approval:
Approve - 21%
Disapprove - 79%

lol
but rasmussen polls are more fair than newsweek's

Of course, only the right leaning pollsters are ever fair...

OH GOD FORBID, OBAMA GETTING LOW RATINGS FROM ANY POLLS MEANS THAT THOSE POLLS ARE REPUBLICAN POLLS!

GROW UP!

Well considering Scott Rasmussen worked on Bush Jr's re-election campaign, i'm inclined to detect a slight right lean.

Don't you think you should probably use numbers, not just assume that a previous affiliation has caused a pollster to introduce a subtle bias into their sampling?

I mean, he'd (generally) have to go out of his way to do that intentionally.

[Alcon]

The problem is, he actually discloses all of his methods.

He discloses his phone list and his sampling method?  No pollster discloses all of their methods.

[Alcon]

Well statistically speaking, they are exactly the same considering Rasmussen's poll has a 3.5% margin of error. But I get your point.

not quite.  Statistically speaking, we cannot be 95% sure that they are not the same result.  They are not "exactly the same" until we're 95% sure they aren't the same and then we magically decide that they are definitely different.  That doesn't make any sense

Just because a poll is within the MoE does not mean the result should be assumed to be the same, and 95% is a fundamentally arbitrary cut-off anyway

[Alcon]

At this point, I think Obama's approval is somewhere in between Rasmussen and Gallup, tilting a little bit to Gallup....say, around Obama+7

Obama +7 (52-45)  is very close to the result of Election  2008. That's about 370 EV with the regional polarization of voting patterns that we now have.

Bush Approval Rating in 2004 Exit Poll:

Approve - 53%
Disapprove - 46%

Election Result:

Bush - 51% (286 ECV's)
Kerry - 48% (251 ECV's)

Well within the margin of error. So was a near 50-50 split.

1. This was pretty consistent across polls and exit polls -- Bush was slightly more popular than he was strong electorally.  Hard to imagine, but there was a time where Bush was personally popular but lukewarm approvals -- just like Obama is now, except it's 2009 and not 2004.

2. The fact that it's within MoE does not automatically mean we can reject it.  That just means that we haven't reached a scientific level of reasonable certainty (somewhat arbitrarily, 95%).  Eighty percent certainty is still something.

3. I don't like using MoE with weighted exit polls because they are not a representative random sample.  Otherwise, they would not have to be re-weighted so frequently.

4. No, it wouldn't be within the Margin of Error.  The 2004 exit poll's MoE would have been ±0.84%.

[Alcon]

Low Obama approvals in Summer '09 will unquestionably lead to the Autumn '12 election of Mitt Romney, who will institute a fundamentalist Christian theocratic state.  With no elections.  Duh!

[Alcon]

By the "Rove-like figures," I presume you mean Mormons?

Your ideas intrigue me and I wish to subscribe to your newsletter.

[Alcon]

"Beginning to"?  Like this was based on some sort of substantial upgrade in America's understanding, that will go out exponentially?  haha, if only!

[Alcon]

Make NJ yellow.

Why would you use 2009 LV's instead of 2008 electorate?  I can see an argument for 2008 electorate measure not being included on the map, but why use the number from an election that's obviously going to be less reflective of any Presidential electorate?

[Alcon]

We now have two maps that are misinterpreting poll results in unique ways

[Alcon]

VOTER REGISTRATION-RELATED FIGHT!!!!  VOTER REGISTRATION-RELATED FIGHT!!!!

[Alcon]

Hahaha, so you include the VA poll when it's positive but dismiss the NJ poll when it's negative. You sir are not just a moron, but a HACK.

For the third time, why are you advocating for the inclusion of a poll of likely 2009 voters?

Is the Virginia poll likely 2009 voters, too?  If not, why shouldn't it be inclusion?

[Alcon]

Brandon is right.  The WaPo poll is of likely 2009 voters.  You should be consistent if you're going to dismiss the NJ poll for the same reason.

[Alcon]

Arizona(Rasmussen)

Approve 44%
Disapprove 56%

http://www.rasmussenreports.com/public_content/politics/elections2/election_2010/election_2010_governor_elections/arizona/toplines/toplines_2010_arizona_governor_race_september_27_2009

Likely 2010 voters wtf?

[Alcon]

Arkansas:

Approve: 37%

Disapprove: 62%

http://www.rasmussenreports.com/public_content/politics/elections2/election_2010/election_2010_senate_elections/arkansas/toplines/toplines_2010_arkansas_senate_race_september_28_2009

Strongly Disapprove well over 50% too.  Doesn't seem to be all that much substance to the idea of white southern voters swinging Democratic more than any group in '12.