Zogby: Bush Approval=52%/Disapproval=48%; Re-elect=45%/Someone new=48%!

43% Bush / 43% Un-named Democrat!
(http://www.zogby.com/news/ReadNews.dbm?ID=728)(Polling done August 16th-19th)

The slide continues.

...sliding into the toilet.

What a drunk asshole in chief!:beer: :puke: :wtf:

He's just trying to get out of the pResidential limo. He's a fucking degenerate.

Someone needs to photoshop it real bad.

as the unnamed Democrat.

You can check the older files right here on DU2. I said 45% (+/- 2%) by Labor Day.

I'm standing by it. He's in free fall right now.
The Professor

45% for re-elect or approval?
which polling organization?

45% reeclect

52% approval


I like Zogby .... He had 00 right but he f--ked up in 02 big time.

I will await other polls to confirm or deny....


btw

go to www.zogby.com

I use an exponentially weighted average combination (based upon sample size) of Pew Research, Zogby, ORC, and Gallup numbers.

Then i run a log-log (Weibull) prediction curve over time and estimate the points of probability into the future. I cut off at the 75% probability point. (Alpha = 0.125)

I run these curves iteratively assuming a series of increasing rates of decline. (Also exponential). At some point these curves converge. If the convergence happens at a point that is not possible to confirm (like one that says 45% by yesterday, when that's already untrue), it's rejected from the curve set.

The equations of all these curves are then combined into a 3d model of time, probability, and approval. Point of maximum fit includes all three variables.

The model indicates that by Labor Day there was a 75% confidence that the approval rating would continue to decline to 45% by Labor Day.

Hope that helps.
The Professor

nt

Data is viewed on a subscription basis only. They do polling for specific research, mostly on product introductions for consumer companies. But, they've got a social stats division.

I've got subscriptions to several data sources, since analyzing data with a myriad of mathematical techniques is both what i do and what i teach.

I use their data as part of the whole, because they use somewhat larger samples, like Pew Research does. (I use lots of Pew data because i like their methodology and i know people who work for them.)
The Professor

nt

. . .bewteen 53 and 54%. Their margin of error is 5% relative, or +/- about 2.7% at 95% confidence interval. I think the sample size was 1800 or so. That's pretty standard for them. A little smaller than Pew, but bigger than Gallup usually uses. (I think the former is usually up around 2500 - 3000, and the latter 1000 - 1200.) I'm not sure who commissions their political polls at ORC. (Dems, Repubs, or someone in news media. Not sure.)

All these numbers are from memory, however. I don't have that disk with me right now. It's at home by my other laptop. Sorry

I think you'll remember this group. They were the ones who did the follow-up to the Florida election count. Same company. Different division. I don't know anybody there. I know people who work at most other polling orgs. Lewis, Gallup, Pew. (I know several people at Pew, all around the country.)

This is what i do for a living, so i have access to lots of databases that aren't in the public domain. I don't download results. I download raw data. I need subscriber status for that.

Not that i don't trust the people who draw the conclusions and compile results, it's just... No, i don't trust the people who compile and conclude.
The Professor

...raw polling data from various polling organizations and leaves the viewer to draw the conclusion.

Not to defend Pollkatz or anything. It's just very interesting and revealing to view all of the various polling data together. Viewing it with a critical eye can lead to "hunches" that hint at statistical conclusions. Pollkatz provides the "hypothesis", you provide the "proof".

Yeah, I remember ORC now. Thanks for the info.

It's a good piece of work. No analysis, just a nice picture of the data taken.

Now that i think about it, ORC might be a division of American Research Group, or vice versa. I saw that on your graph and it rang a bell. I could be wrong about that, though.

If you check the older archives, i had a prediction he would have approval of 48% by April 20th, as long as the war in Iraq didn't start. I amended my models 3 weeks after the "Mission Accomplished" stunt. Even including that bump, it's been showing this 45% by Labor Day number since mid-May. I think i'll stick with it.
The Professor

...bear out. I also felt he'd get a slight bump from Iraq, but that it wouldn't last. The startling thing was how quickly it dissipated and that the decline resumed at an even steeper rate. I believe another war or terrorist attack at this point would only serve to continue Bush's decline in popularity (i.e. at this point, I think it's inevitable; he's teetering).

The interesting thing looking at the Pollkatz summary graph is that (at least according to my low-tech "eye-balling" method) no matter WHAT happens (9-11, Iraq bumps), the decline in popularity resumes at an always steeper rate. Each of the three "lines" traced by Bush approval (before and after 9-11, after Iraq) all converge at around 40% by year's end. I realize that before 9-11 there wasn't really a defined "line" as people hadn't really made up their minds about the man; but nevertheless... I think his "true" approval rating is around 35-40%, and the polls will inevitably come to reflect this.

It would be very interesting (well, I hope we never have to observe this!) to see if Bush gets blame or glory following another large terrorist attack. Of course all of us here would say he's diverted resources toward Iraq that should have gone to fighting Al Qeada, etc. However, I have no doubt that the fawning pack of Bush butt-licking spaniels that is our mainstream press will have another field day commenting on his impressive leadership and how we all have to rally around our president, blah, blah, blah.

www.pollingreport.com

Approval ratings fluctuate based upon current events, they do not move in deterministic ways. For example if you had used this method to predict Bush's future approval ratings on 9/10/01, you would have been horribly wrong. Predicting Presidential approval numbers in this way is akin to predicting future stock prices based upon past price history--it only works if no significant news occurs.

... to an extent, predictions can be made (with the caveat that events may change the trend).

Plus, in this particular case, the Bush presidency has been marked by such a consistent decline that it begs predictions on where the decline will end up. It has been so consistent that it has taken two world-shaking events to cause it to deviate (and the last one, not so much).

There is the reality that public opinion can change with events. But there is also the reality that (particularly as the president's term nears the end) the people are beginning to solidify their opinion of him. That is, it is not so much a number at the whim of daily events anymore, but is more a real number that is fairly stable. With the exception of the few weeks after 9-11 and after the Iraq invasion, Bush's approval has been remarkably consistent in trending towards the 40s. Adding to the possibility that his "approval" rating is a fairly stable, "real" number, is the fact that he is a very polarizing president. This would have the effect of his approval/disapproval being more or less locked in once enough time has passed. I believe that is what is happening as we near the last quarter of his term. This happened with Clinton (admittedly polarizing, though not necessarily his doing). From about this point in his first term throughout his second term, his overall approval rating hovered around a fairly consistent 60-70% (see the Pollkatz page).

It is, however, mathematically valid. I wouldn't do a prediction based upon extrinsic events in the short term. Rather it is a phenomenon known in social sciences as "The Bandwagon Effect".

As long term conditions cause people to look less intensely at instantaneous events (9/11/01, going to war in Iraq, et al), the overall condition of the gov't starts to look different to them. As a result, they begin understanding the "naysayers" a bit more.

Since many of those polled, even if all registered and likely voters, are not students of current events, the approval ratings are based on impressions of the man and what they hear from others and the media.

As poll numbers fall (or rise), people start thinking maybe they're missing something. They start paying more attention to what the complaints are and find themselves agreeing he's doing a bad (or good) job.

When the numbers fall far enough (the inflection point of this surface was about 55%, some people just change their minds as the jump on the "bandwagon" and become detractors rather than supporters.

This is all fairly established as sociological fact. I didn't invent the concept. My contribution was the selection of the mathematical constructs used in the prediction of the curve based upon over 30 years of polling data and the reactions of the populace to LONG TERM AMBIENT societal issues.

So, would i have been wrong on Sept. 10? Yes. My assumptions have been stated here before that no major extrinsic force would encourage a new "rally 'round the flag" mentality.

But, the principles and the techniques are not flawed in any way. You're free to ignore them if you wish.
The Professor

I assume by this that you have tested the empirical accuracy of this method against historical data? Given that data for approval ratings exists going back to 1938, its should be a simple thing to do...

It said that in my last post. Pay attention

That being said, this conversation is over. I'm bored with it. Believe my model or don't. It doesn't matter that much to me.
The Professor

Your post mentions its use in predicting the "reactions of the populace to LONG TERM AMBIENT societal issues." I was just wondering if it had been tested against Presidential approval rating as well. It was just a question, no need to get all defensive...

... through August 20th (yesterday). I have said repeatedly in forums here lately that I believe the FIRST poll with data up to or including August 20th will have him below 50%. Not necesssarily his "average", but the first one to cross the threshold, as it were. This Zogby poll is likely the last one to be conducted prior to 8-20 (August 16th-19th). I'm cutting it down to the wire!

I think you've been looking at the Pollkatz site too, and drawing an imaginary line for the lower end of the range. That also puts him around 45% by Sept. 1 (so I agree with your prediction, too).

Either way, he is in a mighty steep free-fall that must be keeping Karl up all night.

I don't look at other polling conclusions.

Also, there's no imaginary line drawing. It's more a more sophisticated projection that that.

See my reply to Nederland for a description of my methodology.
The Professor

Sorry if I insulted you. I personally hate doing statistics, but I put great validity in them when done right. Without having the ability to perform the actual stats, I can look at the Pollkatz graphics (or pollingreport.com data) and tell that both of our predictions are on the money (oh, I might be a couple of days early on the "below 50%" part, but he'll get there very soon). I'm glad you're doing the number-crunching.

When people say 'there are lies, damn lies, and statistics", they are not so much talking about statistics as people's misuse of them (i.e. selectively picking numbers that seem to lead to the desired conclusions). Good statistics do not lie. Good polling data does not lie. Of course it is only an extrapolation, but a mighty good one based on scientific analysis. I believe the polls are accurate, and that Bush's approval rating IS actually around 50% now. But you have to factor into that people's reluctance to be "against" their "leader" at a time when the nation is still reeling from 9-11 (although we're recovering, and the increased desire to come out against our "leader" is evidence of that).

I always look at the Pollkatz site because it includes most major polls and, while there is some variation, the compilation shows Bush's downward trend to be remarkably consistent across all polls. The Pollkatz "Bush Approval" graphic gives me great hope every day.

Is the slope of the last bump...notice how much steeper it is then the trend line on the 9/11 bump. His war plan is apparently accelerating the negatives on this slide....good to see.

At this rate, he crash through the floor in about 7-8 months.

Notice how there hasn’t been any warnings since the invasion. I guess we are all supposed to feel safer because of his leadership.

Compare the dots on the left with those on the right.

nt

Unless they pull off a miraculous PR coup...Remember, this poll was done before the UN got hit...

The model works!

I think the number will increase significantly. Call me an optimist but as soon as there's one focused voice out there with a face attached to it, from the democratic side the numbers will get even better. Right now the message is split in 9 or 10 different directions, which is fine for now but -- I predict a landslide for our side in the general election. :toast:

Historically, once a specific candiate is named, the numbers go down.

Out the door in 2004!

Now the scary part -- it's going to be ours to lose.

This motherfucker is gone, the better.

What's there to approve of?
The Iraq debacle?
The alienation of other countries that had always been our friends?
The economy?
The unemployment?
The Afghanistan deterioration?
The deficit?
Health care?

:party:

I'll bet the Rovians are busy trying to overload the site.

I am so, so, so sick of hearing "this popular President."

One look at the Pollkatz "Bush Approval" graphic (see my Post # 20 above) and it is obvious that his entire term has been nothing but one long and VERY steady drop in popularity (but for about a week of post-9-11 shock and a week or post-Iraq "we're at war, let's stick together"). He didn't even have time to do anything to merit the huge bump he got after 9-11 (his approval numbers practically doubled within a few days), and he's been riding that bump ever since. But it's just about gone now.

Think about it. He MUST be doing SOMETHING (wrong) to be losing all these voters hand over foot!

They should have been saying he's a CIRCUMSTANTIALLY-POPULAR president dropping in the polls as the REALITY of his Presidency becomes clear.

It was just based on very little information. Before 9-11, most people hadn't really made up their mind on what kind of president they thought he'd be.

The shock of 9-11 was definitely a big factor (pulling together and all) in his poll numbers skyrocketing. But in the immediate aftermath, Bush also SAID all the right things, and that helped him, too. Then again, it's hard not to say the right things in circumstances like that, and the people are eager to listen.

The decline since, however, has shown that his actions have not measured up to his words. The citizens have gotten wise to him.

Just as Bush has done with the entire country's reputation, he has squandered the good faith and support that the entire country threw his way in the aftermath of 9-11.

Good one!

It was our own Mark SHIELDS being the polite Dem, who said "stabilized at 59%". Kick-a-roo Kick-a-roni for dropping like a rock.

...came out).

I always worry when I see a poll or two slightly bump up or level off across two or three polling periods (as with the last few Gallups).

But when I looked at these polls and the overall trending among all polls (as summarized in the Pollkatz graphic), I saw that these "stabilizations" were smaller than the overall variation and did not really reflect a levelling off of the downward trend. Bush's numbers would have to stay stable in ALL polls for one or two months to start to say that. (that hasn't happened since pre-9-11)

Whenever I worry about Bush's poll numbers stabilizing or rising, I just wait a week or two, and the inevitable decline always reveals itself to be continuing.

Regarding the Gallup numbers appearing to stabilize, I'd BET the next one has him down a few points to the mid-50s (it's now at 60%). It should be due out anyday.

Google News searches for "Zogby Poll Bush" and "Bush Approval poll" sorted by date turn up NOTHING on Zogby reporting his approval numbers down to 52% (much less the re-elect numbers having him well below "Someone new").

That's our "fair and balanced" media for you!

if the numbers went up.

awakening? There will be dancing in the street when tje Bush cabal is out of there.

If the economy continue to slide Bush will likely fall further. Christmas sales this year will likely be the worst in decades which will sink the economy even more.

should be more like 2%!!!!

what s wrong with you peepell!!