I just figured out what's out of kilter about that CNN poll

[diplomats]

"The poll was based on interviews with 1,004 adult Americans, including 870 registered voters and 456 registered Democrats, and was conducted by telephone between October 10, 2003 and October 12, 2003."

If 456 of the 870 registered voters polled were Dems, the rest must've been Repukes, because othewise the sampling of Repubs would be too small. Apparently, this poll did not include independents, and since independents are much closer to Dems in how they view Bush than they are to Repukes, Bush's rating is artificially high.

[joeybee12]

1. Naw, the problem is they asked 56% who don't have brains!

He's been quite visible in his defense of Iraq--the blip is just the great wishy-washy middle who go either way depending on that day's news coverage.

[Flying_Pig]

2. Still, the numbers would indicate that a number of Dems expressed

support for shithead. What is it with Dem support for this guy anyway? It only proves how lousy the Party has been doing at getting the word out about the lies and crimes of the regime. Of course, the press has been carrying the ball for Junior, but one would like to think Dems could see through the b.s. ....

:wtf:

[diplomats]

8. Bush's approval rating among Dems is in the low 20s, I think

It's higher than Bush's disapproval rating among Repukes, hence the reason Bush's overall rating is in the mid-50s.

[Caution]

3. your reasoning is incorrect

456 dems = 45.4% of the total sample

that is of the 870 registered voters. In fact, these numbers suggest that independents are leaning more heavily towards bush than away from him. You cannot assume that the independents will lean towards the dem camp, nor can you assume that ALL dems will give bush a no in a re-elect.

[diplomats]

7. There's no evidence ANY independents were polled

There were 870 registered voters in the poll; the rest were not registered - they weren't independents. 456 is more than half of 870, so I can't believe the rest of the registered voters polled were split evenly between Repukes and independents. They had to be Repukes to make the poll somewhat representative. You need to check out http://www.pollingreport.com and see the polls I'm talking about where Bush's support among independents is below 50 percent.

[Caution]

16. but there is no evidence that there *weren't*

any independents polled. You cannot say "well they didnt mention independents so they must have not polled them" They didnt mention republicans either, maybe the remained 50+% were in fact the american nazi party and they just think bush is the new fuhrer?

Your argument is the same as "Hussein says he has no WMD, we know Saddam lies so there must actually BE WMD." One does not follow from the other.

[w4rma]

10. 456 / 1004 * 100% = ~45.4% of the total sample are registered Dems (n/t)

[ant]

4. on lou dobbs, was it?

Don't remember if that's his name or how you spell it, but I'm talking about the guy who does the money show on CNN.

Anyway, they were talking about this poll earlier - I'm pretty sure it was this one - and one of the analysts (he's always on talking about polls, glasses, grey hair, sort of a funny way of talking...) mentioned that part of the bounce came from higher income people, people who tend to watch stocks and were maybe reacting to the recent bounce. I'm not sure how he knew that, but it seemed interesting. I'm guessing they asked people what their income is?

Everything the analyst said seemed to suggest this wasn't really the great news it appeared to be on the surface, but Dobbs kept trying to counter his points. His hard-on for W was so obvious; I was really quite stunned. At the beginning of the segment one of the analysts (there were two, but I don't remember which one said this) commented that Dobbs' use of the word "dramatic" to describe the numbers was an exaggeration. It was pretty funny.

[ant]

21. transcript available

I just thought I'd post this in case anyone was interested:

http://www.cnn.com/TRANSCRIPTS/0310/13/ldt.00.html

This is the relevant part (emphasis mine):

DOBBS: Well, to all of us, Frank, because these poll numbers suggest which way the country is looking. Bill Schneider, give us your best assessment as to why the rebound.

WILLIAM SCHNEIDER, CNN SR. POLITICAL ANALYST: Well, I think a lot has to do with hopeful economic signs. What we're finding is that the rebound was concentrated among higher income Americans, those earning more than $75,000 a year. Those are people who keep a close eye on the stock market. And as you report every night, the stock market has been doing well. Look at that, the over $75,000 earners, it went up 19 points just in the last couple of weeks. Again, those people look at the stock market. And I think a lot of that is response to the fact that they feel better off; they're very optimistic about the economy; they think things are going to be better a year from now. And the gains are really concentrated in that income level.

DOBBS: Frank, as you look at the numbers, as you have been analyzing them, what is driving this? Are Democrats, are Republicans being surveyed in your poll? Is it men? Is it women? Where is the strength?

NEWPORT: Well, those are good questions. Bill just pointed out higher income Americans more likely to have gone up, Democrats more likely than Republicans or independents. It's been quite significant, when we look back across the last several polls, Democrats aren't overwhelmingly positive about Bush by any means, but clearly, they've moved up more than have Republicans or independents. And also men, back when Bush's overall approval, Lou, was 50/50, 50 percent men, 50 percent women, when it was overall 50 percent.

(snip)

...I think this is significant out on the West Coast. People who live in the west part of the United States a little more likely to have jumped up. And that could be -- could be -- a Schwarzenegger effect of some sort, although it is very hard for us to try to disentangle the impact of what happened in California.


I'm not sure what that Dems being more positive about Bush than Reps or Inds is about, but I thought the breakdown here was interesting.

[Snellius]

5. You can be a registered independent.

45% That's about average.

[diplomats]

9. My point is, what percentage of the 870 were independents?

Since just over half were Dems, I suspect zero.

[Snellius]

11. We don't know. The data is incomplete.

I came to the same conclusion at first but realized we have no way of knowing from the data what percentage of the registered voters are Republican and how many Independent. 45% is about a national average for Democrats. Trends for Dems have not been encouraging. Of new voters in CA recently before the recall only 9% were Democratic, with the rest evenly split between Rs and Is.

What we do know that's suspect is that there are 13% unregistered who are not only unlikely to vote but are usually the poorest informed.

[diplomats]

13. I think we still have a large voter registration edge in Calif.

but the recall was a bad situation for us all the way around.

[childslibrarian]

6. I agree the other 45.5% were morons

They didn't ask for their income levels, they asked for their IQs...

[diplomats]

12. OK, consider this

That Newsweek poll with Bush at 51 percent also was based on 1,004 people. However, the sample included 377 registered Dems and Dem-leaning independents. Compare 456 to 377. The Newsweek poll obviously included fewer Dems and Repukes but a decent number of independents, which better represents the electorate at large.

[Frodo]

20. OK. Here's what you're missing.

The numbers you give are actually bad news, not good. If 456 out of 870 registered voters were Democrats, it isn't that you are under-sampling independents - you're over-sampling Democrats. We just don't make up 53% of registered voters. You assume that the others "must be Republicans" or they would be under-sampled... but you only know the Democratic percentage. If shrub can improve his numbers in a majority Democratic poll? (and who cares what people say who aren't registerd to vote?) Bad news.

There is zero chance that a pollster would have dropped independents yet included 134 people who weren't even registerd to vote.

[Snellius]

14. CNN and USA Today are hyping this poll differently than Gallup

Gallup's own website calls their results much more mixed than what the media is trying to herald:

The most recent Gallup Poll shows a mixed set of results on ratings of George W. Bush. The president's job approval rating is up slightly from his previous rating, which was the low point of his presidency. While Bush's overall job rating remains in positive territory, his ratings on the economy, foreign affairs, and Iraq show decline, and now as many Americans disapprove as approve of his handling of these. In fact, Bush's ratings on all three issues are as negative as they have been at any time in his presidency. The public continues to hold a favorable impression of the president, although Bush's 60% favorable rating is his lowest since he took office.

http://www.gallup.com/

[diplomats]

15. That article refers to the Oct. 6-8 CNN/Gallup poll

not the Oct. 10-12 one that I'm addressing. Although Bush's ratings (55 and 56) are about the same in both, so that analysis could possibly apply to this one as well.

[zubeneshamali]

17. It's not so much that Bush is more popular per se

What it means is that he's enjoying a bounce thanks to Caliban voters who hired Conan.

65% :bounce: :bounce:
35% :puke:

[newyawker99]

18. Hi zubeneshamali!!

Welcome to DU!! :toast:

[arwalden]

19. This Is Why *ALL* Polls Are Bullshit. Not Just The Online Instant-Polls...

but ALL POLLS.

I don't trust them. Never have. Never will.

Even when they appear to be in our favor, the results are probably skewed to fit the opinions of the pollsters or the those who report the poll... or those who commissioned the poll.

Valid and accurate polls fall into the same category as the "liberal media" myth.

Rant over. Thanks for listening.

-- Allen

[Oilwellian]

22. Approval numbers make no difference

The real number to look at are those who will vote for Bush in 2004. They have been consistently running between 38-42% for the past several months. People will say they approve of Bush out of loyalty to our troops in Iraq. They somehow think if they support Bush, it shows support for them, but it doesn't mean they'll vote for him again.