VeggieBoards banner

Status
Not open for further replies.
1 - 20 of 32 Posts

·
Registered
Joined
·
1,785 Posts
Discussion Starter #1
WASHINGTON (CNN) -- The percentage of registered voters who say they would support President Bush in 2004 fell below 50 percent for the first time, according to a new CNN/USA TODAY/Gallup poll, which finds more Americans concerned about the economy.<br><br><br><br>
Two-thirds of those who responded to the poll, released Thursday, describe current economic conditions as poor, a 10-point increase since December. Optimism about the future of the economy also dropped 10 points during that time.<br><br><br><br>
Asked their choice for president, 47 percent of the registered voters polled said they would support Bush in 2004 -- compared with 51 percent in December. About 39 percent said they would support the Democratic candidate, compared with 37 percent in December.<br><br><br><br>
Still, a majority of those polled, 57 percent, said they approved of the way Bush is handling the job of president. That Bush approval rating is the lowest since the terrorist attacks of September 11, 2001.<br><br><br><br>
The poll -- based on telephone interviews with 1,004 adult Americans between February 24 and 26 -- also found that support for sending U.S. troops to Iraq remains steady at 59 percent. Public attitudes, however, are likely to be shaped by the events of the next week or so as indicated by the respondents' answers to other questions. Nearly half of all Americans say they may change their minds on Iraq; about a third said they are committed to war.<br><br><br><br>
The poll comes as Bush continues to lobby the U.N. Security Council to pass another resolution declaring that Iraqi President Saddam Hussein has missed his last chance to disarm. And the president has made several speeches in recent weeks, emphasizing his concern about the economy and his administration's determination to strengthen it.<br><br><br><br>
The poll numbers suggest Bush has further to go in convincing Americans that he can turn the economy around. About 45 percent of those polled said they favor Bush's economic plan, while 40 percent said they oppose it, and 15 percent described themselves as unsure.<br><br><br><br>
On Iraq, the support for invading that country seemed to hinge on several factors. One example: Forty percent of those polled said they would support an invasion of Iraq with U.S. forces only if the United Nations approves another U.S. resolution against Iraq. And support for an invasion drops significantly if Saddam destroys missiles cited by U.N. weapons inspectors, falling from 71 percent to 33 percent.<br><br><br><br>
As for Saddam's recent challenge to Bush to join him in a debate, poll respondents left no doubt about who they thought would win. Three-quarters of respondents said Bush would win a debate.<br><br><br><br>
The poll has a sampling error of plus or minus 3 percentage points.
 

·
Banned
Joined
·
1,045 Posts
these statistics are silly.. if bush is at 47% now, with +/- 3%, and was at 51% before, presumably with +/- 3% as well, then the statistics overlap. you can't say that his support did indeed go down, or if it was due to sampling error. besides, i can't see how 1,000 people can be very representative of a country where about 100 million people vote.
 

·
Registered
Joined
·
1,785 Posts
Discussion Starter #3
I disagree...if it was four points lower last time with a plus or minus rate of 3%, then I think one can safely assume his approval rating is dropping...but thanks for raining on my lil' parade... :)
 

·
Banned
Joined
·
4,276 Posts
<div class="quote-container"><span>Quote:</span>
<div class="quote-block"><i>Originally posted by stonecrest</i><br><br><b>these statistics are silly.. if bush is at 47% now, with +/- 3%, and was at 51% before, presumably with +/- 3% as well, then the statistics overlap. you can't say that his support did indeed go down, or if it was due to sampling error. besides, i can't see how 1,000 people can be very representative of a country where about 100 million people vote.</b></div>
</div>
<br><br><br><br><br>
Aaaargh! Stupid statistical fallacy!<br><br><br><br>
If you only learn one thing about statistics it should be this: THE QUALITY OF THE SAMPLE SIZE IS UNRELATED TO THE SIZE OF THE POPULATION!<br><br><br><br>
It does not matter if the population is 100, 100 thousand, 100 million or 100 billion, it has no effect on the error produced at a given sample size.
 

·
Registered
Joined
·
1,785 Posts
Discussion Starter #6
Yarnblossom wrote: "Well if this board is any indication, his approval rating is way lower than that."<br><br><br><br>
Unfortunately, I don't think it is...
 

·
Registered
Joined
·
682 Posts
When Bush started his campaign on Iraq I did wonder when he was up for re-election. It wouldn't be the first time an unpopular politician used military aggression to swell the number of voters at the polls that really approve of this type of offensive action.
 

·
Banned
Joined
·
1,045 Posts
<div class="quote-container"><span>Quote:</span>
<div class="quote-block"><i>Originally posted by kirkjobsluder</i><br><br><b>Aaaargh! Stupid statistical fallacy!<br><br><br><br>
If you only learn one thing about statistics it should be this: THE QUALITY OF THE SAMPLE SIZE IS UNRELATED TO THE SIZE OF THE POPULATION!<br><br><br><br>
It does not matter if the population is 100, 100 thousand, 100 million or 100 billion, it has no effect on the error produced at a given sample size.</b></div>
</div>
<br><br><br>
the key phrase being "at a given sample size". i was merely pointing out that as a population becomes larger, so too should the sample size. the larger a sample size becomes, the less "random error" that is produced because unpredictability is reduced. so while you are correct in saying that population size does not affect error, the size of the sample DOES. and when your +/- 3% error equates to a span of about 6 million people, anything that can reduce the error is pretty important.
 

·
Banned
Joined
·
13,022 Posts
<div class="quote-container"><span>Quote:</span>
<div class="quote-block"><i>Originally posted by mushroom</i><br><br><b>I disagree...if it was four points lower last time with a plus or minus rate of 3%, then I think one can safely assume his approval rating is dropping...but thanks for raining on my lil' parade... :)</b></div>
</div>
<br><br><br>
Um, no. You can't safely assume that...<br><br>
I would suggest investing some time and money into a statistics class.
 

·
Banned
Joined
·
4,276 Posts
PLEASE don't talk about statistics if you don't know what the heck you are saying.<br><br><br><br><div class="quote-container"><span>Quote:</span>
<div class="quote-block"><i>Originally posted by stonecrest</i><br><br><b>the key phrase being "at a given sample size". i was merely pointing out that as a population becomes larger, so too should the sample size.</b></div>
</div>
<br><br><br>
Wrong, wrong, wrong. Ideal sample size is not a function of population size, but population variance and effect size. So think about it this way, how many people do you need to sample to get a good estimate of the average number of fingers? For human beings the variance for number of fingers is very low (most people have 10.) Because the variance is low one can get a very narrow confidance interval with a small sample size. Likewise, effect size is an important determinant of ideal sample size. With a large effect size, (shooting people in the head causes a change in heart rate of more than 1 standard deviation) you can get good confidence with a small sample. With a small effect size, large samples are needed to have a high degree of confidance in your results.<br><br><br><br>
It is never the case that "as population size grows larger, so should sample size". In fact, large sample sizes can be a disadvantage because very minor effects become significant.<br><br><br><br>
In fact, technically the population size is infinite for most inferential statistics. This is a subtle but important distinction. Confidance intervals are not measured by comparing the sample to a real finite population of people, but to an infinite population of hypothetical observations about people.<br><br><br><br><div class="quote-container"><span>Quote:</span>
<div class="quote-block"><b>the larger a sample size becomes, the less "random error" that is produced because unpredictability is reduced.</b></div>
</div>
<br><br><br>
This is entirely the wrong reason for selecting a high sample size. Large sample sizes do not reduce "random error". They reduce the error of the sample mean. A sample of 1,000 will still have variance, some of which is caused by randomness.<br><br><br><br><div class="quote-container"><span>Quote:</span>
<div class="quote-block"><b>so while you are correct in saying that population size does not affect error, the size of the sample DOES. and when your +/- 3% error equates to a span of about 6 million people, anything that can reduce the error is pretty important.</b></div>
</div>
<br><br><br>
Well, there are a few problems with this statement. It is not always the case that increasing the sample size improves results. According to my references a sample size of 1,000 is overpowered for all but minute effect sizes. Increasing the sample size beyond 1,000 is unlikely to result in smaller error.<br><br><br><br>
There are a number of reasons for this. First, any measuring tool has a certain resolution. It does not matter how many times you try to measure a teacup with a hand ruler, you will always have an error of about +/- 2mm. The popularity of the 3% threshold seems to suggest that in validating their insturments, the surveyors found that +/- 3% is the limit of survey reseach (due to people not understanding the question, selecting the wrong option, or surveyor error.)<br><br><br><br>
Secondly, the additional benefits (in terms of power) of adding more people to a study tapers off dramatically after 100 particpants. Going from 10-50 offers a huge benefit. Going from 50-100 offers a moderate benefit, going from 100-1,000 offers very little benefit unless you are dealing with trivial results. In fact, large sample sizes are likely to produce spurious results (such as the statistically significant relationship between skirt hem length and the stock market.)<br><br><br><br>
Third, there are several sources of error that can't be handled by increasing sample size such as insturment bias, and a skewed population or sample.<br><br><br><br>
And this isn't going into the entire issue of whether it is really possible to get better that +/- 3% on a characteristic that can change by that much in a matter of hours.
 

·
Registered
Joined
·
5,660 Posts
I am not a fan of Shrub. His approval ratings seem incredibly high to me. Also, the voters gave control of both houses of Congress to the Republicans in the last election, so Bush practically has a blank check. I am feeling very alienated from what seems to be a majority of the electorate.<br><br><br><br>
Let me say one thing "favorable" about Bush: I really do believe that he is <b>not</b> having extramarital sex and/or extramarital "blowjobs." To those members of the American electorate for whom this is a very important issue--and there seem to be a lot of you out there--I hope you are very, very happy.
 

·
Banned
Joined
·
13,022 Posts
S'okay, Joe. You are just going through Liberal Angst. The only cure I know of is to overdose on back issues of <b>The Nation</b>.<br><br><br><br>
I have issues with Bush in some areas, as I would with any president. But let's be honest - liberals/leftists will never support a conservative. Conservatives will not support a liberal. The rest of the country floats somewhere in the middle.
 

·
Registered
Joined
·
0 Posts
I don't know about statistics and I don't like statistics (they can be tricksy), but I do know that the majority of people I talk to (on the internet and off the internet) really don't agree with his choices. Even if they liked him before or voted for him, they now do not like him. Does he really think he'll get re-elected after this?
 

·
Banned
Joined
·
13,022 Posts
Herself - but are you interacting with a representative sample of eligible, active voters?<br><br><br><br>
At this point, I would say his chances at re-election are excellent, of course depending on how the war with Iraq pans out.
 

·
Banned
Joined
·
4,276 Posts
I'm not convinced Bush is invulnerable. "Rally 'round the flag" syndrome didn't help Bush Sr. after Gulf War I and Bush Sr. was in a much better situation than Bush Jr.. Of course a lot of it depends on whether we will actually have an opposition party in 2004 that will call Bush out on his contempt for the Constitution, his re-employment of convicted perjurers and the slumping economy.
 

·
Banned
Joined
·
13,022 Posts
The economy was slumping well before Bush was in office. Ah, but facts and reality are pesky things, ain't they.
 
1 - 20 of 32 Posts
Status
Not open for further replies.
Top