Welcome, Guest. Please login or register.
Did you miss your activation email?
July 23, 2019, 02:53:11 am
News: 2020 Presidential Predictions (General) are now active.

  Atlas Forum
  Election Archive
  Election Archive
  2012 Elections
  The Official Obama Approval Ratings Thread (search mode)
Pages: [1] 2 3 4 Print
Author Topic: The Official Obama Approval Ratings Thread  (Read 1000216 times)
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« on: January 22, 2009, 02:08:49 pm »

Not pushing those leaners are we
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #1 on: January 22, 2009, 10:30:11 pm »

What was the question they were asking? Perhaps "Do you approve of Obama swearing the oath?" I don't think that's about the only thing on which to judge him one way or the other at the moment...

Well, cabinet appointments and actions over the first two days, however few of them are.

Cabinet appointments is 90% what I'm going on, currently.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #2 on: February 21, 2009, 03:11:13 am »

Doesn't Rasmussen count "fair" as disapprove?

That makes a considerable difference and is annoying and inane.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #3 on: February 21, 2009, 09:01:03 pm »

Likely voters for what? The election is 4 years away.

And Rasmussen was far from the most accurate pollster in 2008. They were pretty mediocre overall, and had a bit of a Republican bias.

On the national poll they were essentially dead on. They had 52-46, the final was 53-46. I'd call that pretty damn accurate. Did I say they got the states right?

When you're determining who is a good pollster, ignoring state polling in favor of the final national result is a bad idea.  It's essentially taking one poll and ignoring dozens of others that use a similar methodology.

Either way, an LV screen right now does very little.  The only criteria that can be employed, is past electoral participation.  I'd much prefer RV for now.  Adult does tend to have a Democratic bias.

But you're still doing that annoying mode of analysis you did last time you were here.  You do realize that you were, how to say this, completely wrong?
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #4 on: February 21, 2009, 09:11:56 pm »

Likely voters for what? The election is 4 years away.

And Rasmussen was far from the most accurate pollster in 2008. They were pretty mediocre overall, and had a bit of a Republican bias.

On the national poll they were essentially dead on. They had 52-46, the final was 53-46. I'd call that pretty damn accurate. Did I say they got the states right?

When you're determining who is a good pollster, ignoring state polling in favor of the final national result is a bad idea.  It's essentially taking one poll and ignoring dozens of others that use a similar methodology.

Either way, an LV screen right now does very little.  The only criteria that can be employed, is past electoral participation.  I'd much prefer RV for now.  Adult does tend to have a Democratic bias.

But you're still doing that annoying mode of analysis you did last time you were here.  You do realize that you were, how to say this, completely wrong?

Apparently I haven't. And the only RV poll out there right now is FOX and that is at 60%.

It's equally unnecessary to just disregard adult polls, though.  If you ask an average group of adults, "are you registered to vote?," many will lie in the affirmative.  Therefore, the % RV is not really hugely lower than the % adult.  It does not account for a nine-point swing like that.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #5 on: February 21, 2009, 09:18:03 pm »

You also have to take into account the quality of the pollster. CNN and R2K are known to have a liberal bias.

1. "Are known to"?  Is this a feeling, or do you have 2008 analytical numbers?  It may be true, but I doubt you know why other than because you want it to be.

2. You're known to have a sharp implicit conservative bias, but yet you never ever ever ever account for that.  You account for other biases rigorously, but only in order to support your own.  What the hell is up with that?
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #6 on: February 21, 2009, 09:27:19 pm »

You also have to take into account the quality of the pollster. CNN and R2K are known to have a liberal bias.

1. "Are known to"?  Is this a feeling, or do you have 2008 analytical numbers?  It may be true, but I doubt you know why other than because you want it to be.

2. You're known to have a sharp implicit conservative bias, but yet you never ever ever ever account for that.  You account for other biases rigorously, but only in order to support your own.  What the hell is up with that?

1. In some states they did, but since they never released the crosstabs its hard to tell.

Why would the crosstabs be relevant to the topline?  We're not adjusting for crosstabs with these polls, either.

2. I am a moderate.

I don't care if you're a full-fledged communist.  Your "adjustment" techniques render a sharp pro-Republican bias to your analysis.  Your political affiliation does not make the analysis any less bad.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #7 on: February 21, 2009, 09:36:32 pm »

They are relevant because if they oversampled Democrats and just happened to get lucky with the topline, that doesn't make them a good pollster.

That's a fair point.  However, in the past you've used that just to claim that Democrats are oversampled, and it turned out you were wrong.

Which brings me to the half of my post you clearly ignored, and I eagerly await an answer to.  This is your chance to admit analytical fault, correct for your errors, and redeem yourself -- the scientific way!

Or, keep living in fantasyland and I'll keep annoying you like this.  What's your choice on this one?
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #8 on: February 21, 2009, 09:42:52 pm »

So you think that you used the information available to the best of your abilities?

And you think that you are not any more zealous in adjusting for things that favor Republicans, than Democrats?

That's your final answer?
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #9 on: February 21, 2009, 11:12:48 pm »
« Edited: February 21, 2009, 11:15:29 pm by Alcon »

Probably not.

...

Of course I am going to try to make things seem better for my candidate than they might appear to be, everyone does that to some degree. But at no point after mid-September, did I ever believe McCain actually had a shot to win.

It's a natural human tendency to overestimate evidence to fit a precept.  Most people over-adjust toward what they want to be true.  Some people over-adjust toward pessimism, because then they're always pleasantly surprised.  Others (studly others with I-WA avatars) attempt to compensate artificially for this, by observing their biases and trying to create "weighted" auto-corrections.  And others use different methods of correction.  Corrections like these aren't perfect (or even close!), but they tend to be closer to reality than what our minds initially tell us.

Outright "try[ing] to make things seem better for my candidate than they might appear to be" is even worse than just working on those precepts.  It's, like, bad ESPN football analysis.  "Everyone" does not do that.  And, what's the point?  I guess, if the point of political analysis for you is to find creative ways to prove that what you want to happen, will happen, that's an option.  But, even ifyour goal is creative thinking, the resulting info is totally poisoned.  You could make it more useful.  As it is, the product of your analysis has nothing to with actuality.  It's useless.

I'm not advocating the way I (over)compensate, but knowing you have a bias and encouraging yourself to propagate it, is even worse than pretending that it doesn't exist.  Which is bad enough.  Things will inevitably end up not corresponding with reality.  If you're going in with the intent of making them correspond with reality as little as possible (that is, "make things seem better than they might appear to be"), why even bother advancing your theories?

So, if you're going for something other than realism...well, mazel tov with getting people interested in that.  Otherwise, I ask again:  What the hell?
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #10 on: February 21, 2009, 11:23:47 pm »

That wasn't a lecture, that was a question with paragraphs!  A question with paragraphs!

I'm not asking you to be likable.  I'm just saying there are more private places to masturbate than Internet forums

Metaphorically speaking.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #11 on: February 21, 2009, 11:34:06 pm »

That wasn't a lecture, that was a question with paragraphs!  A question with paragraphs!

I'm not asking you to be likable.  I'm just saying there are more private places to masturbate than Internet forums

Metaphorically speaking.

Isn't that the whole point of forums? So people can masturbate together? Metaphorically speaking of course.

You're creeping me out.  Unmetaphorically.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #12 on: March 03, 2009, 10:33:39 pm »

RACIST JOKES: “Nearly one in six Tennesseans has told a joke about Barack Obama’s...even though only 15 percent of Tennesseans say they would find such a joke funny.”

lol
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #13 on: March 04, 2009, 01:59:27 am »
« Edited: March 04, 2009, 02:03:24 am by Alcon »

RACIST JOKES: “Nearly one in six Tennesseans has told a joke about Barack Obama’s...even though only 15 percent of Tennesseans say they would find such a joke funny.”

lol

The most hilarious thing is that "nearly one in six" and "15%" are pretty much the same number. Some misleading reporting of the numbers...

That's exactly what I was lol'ing at Tongue  In fact, ostensibly the "not funny" one is likely to be higher, since 15% is closer to one in seven.  But then there's rounding...

Still, funny.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #14 on: March 26, 2009, 01:52:43 pm »

Quote
Iowa:

57% Approve (-6)
40% Disapprove (+8)

Huh? Why has Iowa dropped so much?

Like I said, some of these are just really random, like Missouri went way up. You'd think it would be the opposite since you know, he lost the state. I have no idea why Iowa is dropping that much in a state he won by 10 points.

SUSA really isn't a good enough pollster to put so much stock in movements like those.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #15 on: April 08, 2009, 04:24:16 pm »

one of the biggest critiques of Rasmussen is that he uses far too tight LV screens way out before it's remotely possible to determine who the LV's are.  you should at least acknowledge this criticism if you are going to assert his superiority over other polls for an election four years away

I am no way saying that his poll is perfect right now, don't get me wrong, I'm not saying that at all. My personal preference would be for registered voters polls. But to say he didn't nail the final result in 2008, is just ignoring the facts.

But you do need to give more concession to LV screens.  One of the major components of a quality LV screen is "how much attention are you paying to this election?"  Such a screen becomes event-sensitive at a time like now, and for instance bailout critics are more likely to identify as "paying close attention" right now than supporters.  If such a screen isn't used, this test falls back on past electoral participation -- a useful variable, but one that inevitably biases Republican at this time in the cycle.

Rasmussen is a reputable pollster.  I'd hardly dismiss his LV screen, even very early, as a measurement of likelihood to be politically educated.  But there is no such thing as a "likely voter" screen three years before an election.  A "likely voter" three years before the election is not hard to screen, but like I said, you're not going to get a representative sample.

By the way, not to criticize results-oriented analysis, but a Zogby could nail final results.  Even if someone nailed three elections in a row, they shouldn't be granted unquestioned credibility unless they can rationally explain their methodology.  Rassy can, but I'm saying, "nailed it" is not the end-all.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #16 on: April 24, 2009, 06:04:57 pm »

Gallup the outlier? Everything is trending down except them.

An outlier in that it's going up and two other polls are going down?  That does not meet any statistical definition of "outlier" I know of.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #17 on: April 26, 2009, 01:03:40 am »

He is willing to listen to different points of view: 90-10

It's amazing how few people in this country are complete hacks.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #18 on: April 28, 2009, 03:44:12 pm »

Quote
Kansas:

44% Approve (-11)
50% Disapprove (+10)
Smiley
I think this will close any assumptions that Kansas could vote for Obama under a good term.


A 21-point swing not exhibited in any other states looks a heck of a lot like an anomaly to me, either in the first poll, the second or both.  Either way, I doubt Obama was +14 in KS last time and -6 now.  I don't think we can learn much from the polls together.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #19 on: April 28, 2009, 06:07:58 pm »

I mostly want to post this SUSA ad graphic because I find it amusing

Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #20 on: April 29, 2009, 02:54:09 am »

Electoralvote.com recognized it (with PPP  on the other side) as a partisan poll.

Weird, on what basis?  They poll mostly for media outlets, and may do polling for the Republicans, but hell Rasmussen did it for the Libertarians.  They've certainly never exhibited any profound GOP bias, so I wouldn't really see it as especially relevant either way.  We have enough of a record to know of one if it were to exist.
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #21 on: April 30, 2009, 06:24:12 pm »

In FL-24, they even showed Kosmas (D) trailing Incumbent Rep. Feeney (R) by 42-43, but Kosmas beat Feeney by 57-41 in the General election.

So, they may not be biased, but... Tongue
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #22 on: May 01, 2009, 04:01:27 pm »

SurveyUSA

Approve 58%
Disapprove 38%

http://www.surveyusa.com/client/PollReport.aspx?g=ac7c1602-9851-41a1-bf10-cfc2a8f69022

Another interesting question, the do-over:

Obama 54%
McCain 39%
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #23 on: May 03, 2009, 01:09:59 am »

SurveyUSA

Approve 58%
Disapprove 38%

http://www.surveyusa.com/client/PollReport.aspx?g=ac7c1602-9851-41a1-bf10-cfc2a8f69022

Another interesting question, the do-over:

Obama 54%
McCain 39%

But it's kinda not relevant since it uses a different electorate composition than what voted in 2008.

1. There were polls of adults in 2008

2. The 2012 election is kinda not relevant since it uses a different electorate composition than what voted in 2008, or can be modeled by the current LV polls you've (IIRC) previously advocated

You're kind of annoying sometimes

An exit poll is just that. Another poll.

Not really -- Different flaws and benefits
Logged
Alcon
Moderators
YaBB God
*****
Posts: 30,923
United States



« Reply #24 on: May 03, 2009, 01:21:51 am »

All polls have different 'flaws and benefits.'

In a very specific sense, yes.  But exit polls have a radically different methodology and sample than phone polls -- they are hence not "just...another poll."  That's what I meant, at least.

It's just not hard enough data.

What do you mean by "hard," and it's not hard enough for what?
Logged
Pages: [1] 2 3 4 Print 
Jump to:  


Login with username, password and session length
Logout

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

© Dave Leip's Atlas of U.S. Elections, LLC