Bush Approval Rating Slips

[satch919]

From MSNBC.com
http://www.msnbc.msn.com/id/6952383/

Some results:

Adults were evenly divided on Bush’s job performance in January, but now 54 percent disapprove and 45 percent approve.

The number who think the country is headed down the wrong track increased from 51 percent to 58 percent in the past month.

Older Americans, especially those 50 and above, were most responsible for the declining confidence and approval numbers.

Middle-aged people between 30 and 50 were about evenly split on Bush’s job performance.

42 percent approved of the president’s handling of Iraq, while 57 percent disapproved. That compares to 44 percent approval and 54 percent disapproval in January.


Again, why did they vote this guy back in? Geez, it wasn't that hard to see that Bush was going to be a failure.........again. :angry:

 

[GhostFox]

I find this odd because the day before yesterday an official poll said his approval rating had risen to 57%.

 

[No Limit]

I find this odd because the day before yesterday an official poll said his approval rating had risen to 57%.

What official poll? The one he posted is as official as you can get.

 

[GhostFox]

CNN/USA TODAY GALLUP Poll.

http://www.cnn.com/2005/ALLPOLITICS/02/07/poll/

But 57 percent of the 1,010 respondents surveyed Friday through Sunday stated that they approved of how the he does his job.

 

[mortiz]

Why would Bush care? He's out at the end of this term so he doesn't care about approval ratings any more, at least not from the general public.

 

[Korgoth]

CNN/USA TODAY GALLUP Poll.

http://www.cnn.com/2005/ALLPOLITICS/02/07/poll/

But 57 percent of the 1,010 respondents surveyed Friday through Sunday stated that they approved of how the he does his job.

http://www.theleftcoaster.com/archives/003675.php
and there you go...
But wait I thought this was "teh liberal media"

 

[GhostFox]

Maybe those figures represent the political spectrum the best? I know there are more registered republicans then democrats, so more should be polled.

 

[Icarusintel]

every poll is bs, there's only one true way to know and that's to ask every single person, do those polls represent the houseless?

 

[Steve]

Yep. This sounds about right. Old people get pissed off when you talk about messing around with Social Security.

do those polls represent the houseless?

The homeless aren't all that concerned with politics, I'm sure. Much less would they be "in the know".

 

[KoreBolteR]

every poll is bs, there's only one true way to know and that's to ask every single person, do those polls represent the houseless?

what can i say.. its stereotyping at its very best.

 

[GhostFox]

ghostfox they shouldn't be polled more... the best way to do a poll is to go to multiple cities and ask random people

Incorrect. If your country is 40% Republican, 30% Democrat, and 30% 'other", then for every 100 people surveyed, 40 should be R, 30 should be D, and 30 should be O.

By interviewing random people you could potentially survey all R's, or all D's etc. That is why they record your political views ahead of time. They are nessicary to correct polling practices.

 

[satch919]

Why would Bush care? He's out at the end of this term so he doesn't care about approval ratings any more, at least not from the general public.

Thats the point. He doesn't care at all. He does whatever he wants. I don't think he has ever given it a second thought. Now that he does't have to worry about re-election, he has even more freedom to step on toes.

 

[No Limit]

Incorrect. If your country is 40% Republican, 30% Democrat, and 30% 'other", then for every 100 people surveyed, 40 should be R, 30 should be D, and 30 should be O.

By interviewing random people you could potentially survey all R's, or all D's etc. That is why they record your political views ahead of time. They are nessicary to correct polling practices.

You are fully correct; however, I am not a pollster and I suspect you aren't. What Gallup has done; and they got a lot of heat for this during the election, is over sample Republicans by too high of a margin and their polls show this if you compare them to other polls. I personally believe that Zogby is the best pollster out there but all that is based on pure opinion.

 

[GhostFox]

What Gallup has done; and they got a lot of heat for this during the election, is over sample Republicans by too high of a margin

You could be right. I don't know what the margins should be, so I can't say. Which is why you'll notice that I phrased it as a question above. I know there are more registered replubicans then democrats, so more should be polled. I have no idea of the actual percentages.

A lot of people however were taking it that the flaw was surveying more republicans then democrats. That is the correct practice. I don't know if their margins were correct however.

 

[Icarusintel]

The homeless aren't all that concerned with politics, I'm sure. Much less would they be "in the know".

Way to not be PC, they're houseless, some of them do have places they call home, they just don;t have a house in the conventional sense

 

[GhostFox]

Way to not be PC, they're houseless, some of them do have places they call home, they just don;t have a house in the conventional sense

Please tell me you are kidding.

 

[PvtRyan]

Incorrect. If your country is 40% Republican, 30% Democrat, and 30% 'other", then for every 100 people surveyed, 40 should be R, 30 should be D, and 30 should be O.

By interviewing random people you could potentially survey all R's, or all D's etc. That is why they record your political views ahead of time. They are nessicary to correct polling practices.

Not true. If that's the case, then you need to expand your samplegroup (from for example 10 to 1000), and be as random as possible (across different cities and neighbourhoods and cultures).
Whether or not they are democrat or republican is not relevant here, you want to know how the avarage American thinks about it, and you do that by taking random samples within the American population.

 

[GhostFox]

Whether or not they are democrat or republican is not relevant here, you want to know how the avarage American thinks about it, and you do that by taking random samples within the American population.

That is incorrect. You take random samples based on the group composition.

I know you think you are right, and you are entitled to your opinion, but anyone who has ever taken a single statistics course will tell you that you are wrong. It is just a fact. It is not a debateable issue.

To be more accurate, not only would they have to break them down politically, but also geographically, culturally, economically, chronolagically, and societally. That way their sample group would be an identical representation of the US, just on a smaller scale.

 

[Raziaar]

Where do they take the random people? Are they completely random from all parts of the country, or certain parts of the country, or what?

 

[PvtRyan]

I have taken a statistics course

Maybe I'm misinterpreting you, but the point of a sample is that you take a samplegroup, that should be representative of your population (the American people) and if that samplegroup is large enough then you can generalize that samplegroup to the entire theoritcal population. You are not out there to find 40 republicans, 30 democrats and 30 "others", but you simply take 100 people and question them. The more you take, the smaller the chance of errors is, so you better take 1000 or 10000 over 100 to ensure you've covered all groups. And of course, you should spread this out, not in one area, and not among one "kind" of people.
Otherwise you would have to take 40 republicans, 30 democrats and 30 "others" that are white and rich and 40 republicans, 30 democrats and 30 "others" that are black and poor to be representative, etc etc.

I have a book lying in front of me that says that a sample should be of the same composition as the population, what you're saying, but to achieve that, usually you take an aselect group, and the bigger that is, the more chance you have to have a representative view of the population. So as I see it, you and me are both right, you need to have an accurate composition of the population in your samplegroup (30 republicans etc) but to achieve that you take a random samplegroup, as large as possible.

If I had to do an investigation such as this, I would pick at least 25.000 people from the birthregister to assure a representative sample.

Maybe I'm misinterpreting you, or you me.

 

[GhostFox]

Ryan, you are talking about natural representation. I.E. The sample group is so large that statistically speaking you should have a general representation of the whole.

Which is fine and dandy. But for the US you are talking about a sample group that would probably need to be in the hundreds of thousands to be truly accurate, possibly millions.

In situations like this poll, a proportional representation, where only a thousand people were surveyed, it is nessicary to define qualities, because the sample size is much too small to possibly be naturally accurate.

Polls never use natural representation, simply becuase of the costs involved. Can you imagine the price of surveying 1% or even .01% of registered voters?

 

[Icarusintel]

Please tell me you are kidding.

Actually, I'm not, they're houseless, please don;t call them homeless, it's very stereotypical and means the wrong thing

 

[Tr0n]

Actually, I'm not, they're houseless, please don;t call them homeless, it's very stereotypical and means the wrong thing

Icarus does have a very good stand point both politically and morally.

We should all do the same and start saying Houseless.

(I'm not joking or being sarcastic.)

 

[Steve]

Way to not be PC, they're houseless, some of them do have places they call home, they just don;t have a house in the conventional sense

Oh, Jesus ****in' Christ. Give me a break. I was raised hearing the word "homeless" on the TV and from my parents and from everyone around me. You are actually the first person EVER to say that to me. And besides, it all depends on your definition of home whether or not I was being PC.

...on a side note, PC can suck on it. A secretary and an administrative assistant are two very different things.

Oh, and good job. You read my post, but managed to respond to only a single word that I said.

Before everyone gets all pissy, I apologize if this post offends anyone. I felt it was necessary.

 

[GhostFox]

Actually, I'm not, they're houseless, please don;t call them homeless, it's very stereotypical and means the wrong thing

I remember back in the mid-90's when for a little while they weren't homeless, they were "urban outdoorsmen". Being PC just makes you look stupid. Ask a homeless person what they are. If you could get even 10% to say 'houseless' I'd be shocked.

You're probably the kind of person that gets upset when people call black people black.

 

[GhostFox]

A secretary and an administrative assistant are two very different things

That made me laugh. So true Steve.

 

[Kommie]

I dont think a cardboard box constitutes a home, If they had a home why aren't they there? Think about it, to not have ANY friends or family to stay with, means they have no home. Either physical or figurative

But back on topic

I have a quick question for ghostfox, Are you sure there are more republicans than democrats? It makes sense but my not-so-leet google skills cant help me.

I think random sampling is usually the best way to go as it gets a better idea of how the general populace will vote. Controlled samples are well and good (in terms of accuracy), but they are very easy to manipulate. Random sampling is not so accurate but not quite so easy to manipulate.

 

[GhostFox]

Kommie, last time I heard a report on tv, registered Republicans surpassed registered democrats back in 2000, and the disparity increased by 2004. As far as exact numbers, I don't know becuase I have never looked it up. I'm only going on what I have heard reported on all the "election coverage" specials.

 

[The Monkey]

My guess is that they didn't like Bush, but they liked Kerry even less.

 

[PvtRyan]

Ryan, you are talking about natural representation. I.E. The sample group is so large that statistically speaking you should have a general representation of the whole.

Which is fine and dandy. But for the US you are talking about a sample group that would probably need to be in the hundreds of thousands to be truly accurate, possibly millions.

In situations like this poll, a proportional representation, where only a thousand people were surveyed, it is nessicary to define qualities, because the sample size is much too small to possibly be naturally accurate.

Polls never use natural representation, simply becuase of the costs involved. Can you imagine the price of surveying 1% or even .01% of registered voters?

Okay, somewhat agreed. But would it really require a samplegroup of several million? I would think a number of 25.000 from the birthregister of people > 18 years would be enough to cover all groups and to generalize the endresult to the population. Okay, still a large group, and expensive.

But if you start defining qualities, like 40% of my samplegroup should be republican you also do that based on statistics, that 40% is actually the natural representation of the American population right? But doesn't that also color the result because that number may or may not be accurate?

 

[GhostFox]

But would it really require a samplegroup of several million

I was probably exagerrating it, but hey, I like accuracy

Anyway, say you have 50 million voters. You'd probably need a sample in the hundreds of thousands to get any kind of true natural representation. 100k surveyed is only one fifth of a percent of the whole. If you were to survey only 25k as you suggest you could very easily get demographic grouping, causing overrepresentation. So you'd need to use at least one qualiative factor to prevent this. Once you need one you officially have a proportional representation.

Once you have a PR, all of the above described steps need to be taken into account. Which is why most polls are so terribly inaccurate. No one is willing to spend to the money to do one correctly.