SwiftVets.com

[arymann]

--------------------------------------------------------------------------------
The poll result article is posted elsewhere on this site, but this headline is worth looking at on Drudge!

[Nomorelies]

Because so many of the polls are run by democrats and weighted to favor democrats. Do you think that Zogby is impartial?
_________________
Nomorelies Make a donation HERE

[buffman]

I don't get it, I was just reading off Newsmax and the Zogby poll said the race was almost even and Bush was behind in Iowa, Penn, NH, etc. What gives???????
_________________
Never Ever Give Up
America First

[kmmpatriot]

[neverforget]

Zogby has an undeserved reputation for being the most accurate pollster. It is based only on 2000 when he was the one pollster who continued polling after the DUI surprise, and thus caught the loss of support for Bush over that weekend.
_________________
US Army Security Agency
1965-1971

[Kimmymac]

I heard the Harris poll has Kerry ahead by 1 point.

Hate to be the bearer of bad tidings. But! Rasmussen has Bush up by 4.5 points, so who knows.

Not that I believe polls too much. I always get flashbacks of '79 when the news media was saying Reagan-Carter was too close to call, even on election eve, and Reagan blew Carter out of the water.

One of the happiest nights of my life. Other than when my husband asked me to marry him, and the kids were born and that junk. You know.

[kmmpatriot]

[arymann]

I don't live by the polls either, but it was such a delight to see that headline on Drudge's page.

This USA Today article is now linked to the header.

http://www.usatoday.com/news/politicselections/nation/president/2004-09-17-gallup-poll_x.htm?POE=NEWISVA

[Polaris]

()#%*Q!


Sorry, my first reaction to Susan Page was unprintable. If she is going to report polls for USA Today, could she please learn some elementary statistical mathematics! As a physicist (thank you USAF for the GI Bill), it annoys me to no end to see her mislead her readers this badly.

I will explain:

When a poll has a margin of error of four percent, it means that the results are randomly distributed so that 95% of the area of a Guassian curve falls between -4% of the stated figure and +4% of the stated figure.

That means that according to Gallup, Bush is 95% likely to be within 51% to 59%. Likewise Kerry is 95% likely to be within 39% to 47%. However, even if the two curves overlap (like they did with the last Gallup poll), the two curves are still distinct if the curves are not within one standard deviation of each other.

In practical, rule-of-thumb approximations, if the spread of a posted poll is equal to or greater than the MOE of the poll, then the lead is statistically significant. If it is within it, it is not. Anyone that tries to claim twice the MOE (as Susan Page does) is a mathematic illiterate.
_________________
-Polaris

Truth is Beauty

[Hammer2]

From the end of the Gallup poll article on Drudge:

"Sen. Kerry is like Seabiscuit: He runs better from behind," says Donna Brazile, who was Gore's campaign manager. But she acknowledges that "backbenchers" in the Democratic Party "have begun pushing the panic button."

Whistling past the graveyard IMHO.

Back benchers?
You mean like Pat Cadell, Joe Liberman, Zell Miller, Griffin Bell, Dick Morris, and the millions of others who know the Titanic when they see it - and are screaming to change course before its too late?
_________________
"The price of freedom is eternal vigilence" - Thomas Jefferson
"An armed society is a polite society" - Thomas Jefferson
"The beauty of the Second Amendment is that it won't be needed until someone tries to take it away." -- Thomas Jefferson

[Kimmymac]

What an insult to a fine animal like Seabiscuit. Kerry is like (used)Dogbiscuit, and should be put in the compost heap.

Polaris, thank you for the explanation.

My husband is an engineer and mathematician, and the only concept I have been able to grasp is the Law of Large Numbers. I use that to explain just about everything, particularly with regard to our budget. Most recently I used it to explain to him why the costs of my daughter's wedding kept increasing. I don't think he bought the explanation, but it did, at least, provide me something of a fall back position. And then the phone rang, so, you know. Saved by the bell. Then my strategy was to start to talk about taffeta versus organza, silk versus cotton, roses versus orchids--- until he wouldn't even broach the subject of money for fear I would ask him if he thought we should ask the daughter's future in-laws over for dinner. Men are really easy. I think our intelligence capability would improve if the CIA had more women.

So which polls do you think are the most statistically reliable, Polaris?

I use the highly scientific method of going with the one I like the best.

--kimmymac
"I may be a mathematic illiterate, but it hasn't stopped me yet."

[Polaris]

Which polls are most reliable mathematically?

That is hard to say without knowing when the polls are taken, how the questions are put together, and how the pollsters model 'likely' voters, and how they determine their demegraphics.

These are all trade secrets btw for any good pollster.

Given that, the best I can do is look at the raw numbers and their published demographics (often called the 'internals') of the polls, and look for consistancy and correlation between related questions (the president's approval rating has a direct correlation to his current number in a poll....and the difference is often called the 'approval' gap).

As such given so much are trade secrets, the best you can do is look at trends and then see which polls are consistantly out of line with the others...and which modeled the actual election the best.

Given all of these factors, I would say that Gallup is probably the most trustworthy poll overall. I also have to say that I have been impressed this year with Strategic Vision. It may have GOP funding, but their results have been uniformally good and suprisingly non-GOP biased (when compared with other polls).

I would say that Zogby (esp interactive which is garbage) is the least reliable this year along with Rasmussen and ARG (which don't release internals :nonono: )

I hope that helps.
_________________
-Polaris

Truth is Beauty

[Wing Wiper]

[The Balloon Artist]

[thrawn21]

This has been said before, but it needs to be said over and over again.

The only poll that matters is taken on Nov. 2.

Besides, as someone else wisely noted, "good polls" make people complacent while "bad polls" make people depressed. Everyone should always assume that they are 10 points behind so that they'll work harder and with a zeal to win. Of course, this doesn't mean to do what the Kerry zealots do, such as rip up signs and key cars.
_________________
"Anyone can make an error. But that error doesn't become a mistake until you refuse to correct it" - Grand Admiral Thrawn, from the book Heir to the Empire by Timothy Zahn