NationStates Jolt Archive

http://online.wsj.com/public/resources/documents/info-battleground04.html

Kerry is now ahead in AR (0.1), FL (0.6), IA (3.0), MI (6.0), MN (9.7), NH (3.6), NM (12.7), OR (12.0), PA (3.1), WA (8.7) and WI (2.4), while Bush leads in MO (5.4), NV (2.2), OH (3.3), TN (5.5) and WV (12.4), according to Zogby

Well then, what does that make it? 282 Kerry, 256 Bush?

I have wildly different numbers: http://www.electionprojection.com/elections2004.html

Anyone with an opinion on accuracy? (I'm praying for your numbers btw)

depends on if anyone will listen to me when i tell them that on election day states where the margin of victory is smaller than the margin of error have no winner. and if no one wins then they can't just go handing out electoral votes. not without looking utterly silly anyway.

depends on if anyone will listen to me when i tell them that on election day states where the margin of victory is smaller than the margin of error have no winner. and if no one wins then they can't just go handing out electoral votes. not without looking utterly silly anyway.

How do they get a margin of error in counting votes? As in human error? I hope they're not doing some sort of sampling data to find the winner .... that would be daft.

Polling has gotten sloppy, and so have I. You have to track down methodolgy in sampling and questions and how the the sums are accumulated (since it's not national polls). Or, you know, we could wait for November. Kinda glass house, as I can't resist looking everytime someone posts one of these, but-polls have become less and less accurate over the last few years.

I really wish I hadn't lost my statistics text book. There are some formulas in there that would really be helpful to me right now.

I worked for a polling place a long time ago and learned the importance of knowing what was asked (and that I'm glad I live very far from New Jersey...sorry guys.)

How do they get a margin of error in counting votes? As in human error? I hope they're not doing some sort of sampling data to find the winner .... that would be daft.

Well, most of the states are going to electronic voting without a paper trail. So, basically, we *assume* that no one has messed with the code and we *assume* that none of the cards got hit with ESD and we *assume* that no one tampers with the results.

As a note on the ESD thing, I would like to point out that my boyfriend, while working for my state senate this past year - found out that shocking the card with ESD not only erased all the information already on it - but also gave him full administrative rights to the computer! How easy it would be for him to rig the results (at least on that computer) so that someone would *think* they were voting for Mickey Mouse, but their vote would actually go to Goofy.

Without a paper trail, I would say the margin of error is larger than any actual differences we will see between votes.

That electronic voting sounds like modern Bonapartist totalitarianism to me.


[EDIT: Yeah they get in a big fuss over sampling data being unconstitutional for the census, but an unsafe, alterable, hackable vote, with the Reichmarshall Bush in the White House?...no that's fine, nothing wrong with that. :headbang: :mad: :sniper: ]

I have wildly different numbers: http://www.electionprojection.com/elections2004.html

Anyone with an opinion on accuracy? (I'm praying for your numbers btw)
Enjoy it while it lasts. I'm sure these numbers will inflame the liberal base to work much, much harder, tipping the balance their way.

I have wildly different numbers: http://www.electionprojection.com/elections2004.html

Anyone with an opinion on accuracy? (I'm praying for your numbers btw)
New poll. Election Projection is updated on the weekends, these were released today.

No EP site has those numbers up yet.

How do they get a margin of error in counting votes? As in human error? I hope they're not doing some sort of sampling data to find the winner .... that would be daft.

all measuring/counting procedures have unavoidable error inherent in them, even when used properly (and not by a bunch of crazy old people, like run the polling places in the us). basically, it is impossible to measure something with 100% accuracy. that's why all scientific instruments list a measurement of error on them; volumetric flasks saying + or - 3%, scales that are + or - .1 g, etc. vote counting machines are no different, except that nobody bothers to account for the error there when reporting results.

iirc, the margin of error on the scantron and punch card ballot readers is something like 3%. or maybe it was 1.5%ish - something like that anyway. hand counting is even worse.

New poll. Election Projection is updated on the weekends, these were released today.

No EP site has those numbers up yet.
Today is Monday. Those numbers have changed since I last saw it. But I wouldn't trust any polls. The race is far too much of a dead heat to tell accuratly. The only poll that matters is the one taken on Nov. 3.

Today is Monday. Those numbers have changed since I last saw it. But I wouldn't trust any polls. The race is far too much of a dead heat to tell accuratly. The only poll that matters is the one taken on Nov. 3.
Poll released-Monday.

Poll released-Monday.
Do you mean yours, or the election projection one?

Basically, it looks like a lot of those numbers represent dead heats--nobody's got this thing sewn up yet.

Here's something I read a week or so ago from Zogby, however, that seems to question the very nature of telephone polling. He says telephone polling is dead thanks to the emergence of the cell phone. Cell phones have become so ubiquitous, he says, that telephone polling is inherently flawed now. There are too many people who don't have land lines and who are therefore not polled. Add in the problem of call screening thanks to caller id and there's no way to truly get an accurate representation over the telephone anymore. That's why he's trying out this interactive internet polling service right now--there's no way to know how accurate it is right now because basically he's the only one doing it, but it'll be interesting to see how it comes out in the end.

Do you mean yours, or the election projection one?
My poll. Late monday. His is done early in the morning/late at night on weekends.

all measuring/counting procedures have unavoidable error inherent in them, even when used properly (and not by a bunch of crazy old people, like run the polling places in the us). basically, it is impossible to measure something with 100% accuracy. that's why all scientific instruments list a measurement of error on them; volumetric flasks saying + or - 3%, scales that are + or - .1 g, etc. vote counting machines are no different, except that nobody bothers to account for the error there when reporting results.

iirc, the margin of error on the scantron and punch card ballot readers is something like 3%. or maybe it was 1.5%ish - something like that anyway. hand counting is even worse.

He speaks the truth listen to him. After what happened in 1950, and 2000 I would rather just wait until its all over. Because people inherently make mistakes. Did you know that the order of the names on a poll alone affects results? Things like this make a completely accurate vote impossible.

Basically, it looks like a lot of those numbers represent dead heats--nobody's got this thing sewn up yet.

Here's something I read a week or so ago from Zogby, however, that seems to question the very nature of telephone polling. He says telephone polling is dead thanks to the emergence of the cell phone. Cell phones have become so ubiquitous, he says, that telephone polling is inherently flawed now. There are too many people who don't have land lines and who are therefore not polled. Add in the problem of call screening thanks to caller id and there's no way to truly get an accurate representation over the telephone anymore. That's why he's trying out this interactive internet polling service right now--there's no way to know how accurate it is right now because basically he's the only one doing it, but it'll be interesting to see how it comes out in the end.

The problem is that internet polling will miss anybody that doesn't use a computer, or that just doesn't happen to come across the poll itself. It's about as bad as the missed cell phones and unaccounted-for overseas votes.

Just another electoal vote predictor to help balance the others out:
http://www.electoral-vote.com

The problem is that internet polling will miss anybody that doesn't use a computer, or that just doesn't happen to come across the poll itself. It's about as bad as the missed cell phones and unaccounted-for overseas votes.I know--it's not a perfect solution, but it may be a harbinger of future polling strategies.

you never know when the results may turn around. i still believe there is vote rigging in USA. But some people think otherwise. but still its good news that kerry is in front and that is the way it should stay till the end of the election.

My poll. Late monday. His is done early in the morning/late at night on weekends.
Oh. But that many voters couldn't have changed their minds so quickly. I suspect most of these are conducted by phone. I read somewhere that someone from Zogby was talking about phone polling. He says it's basically dead since the rise of the cell phone. Focus groups are probably more accurate, but it can be hard to find volunteers. My take on this: why can't they do polling through the mail? If we conduct censuses that way, why can't we do polls? No one has a "cellular mailbox".

Oh. But that many voters couldn't have changed their minds so quickly. I suspect most of these are conducted by phone. I read somewhere that someone from Zogby was talking about phone polling. He says it's basically dead since the rise of the cell phone. Focus groups are probably more accurate, but it can be hard to find volunteers. My take on this: why can't they do polling through the mail? If we conduct censuses that way, why can't we do polls? No one has a "cellular mailbox".
Bounce, and Zog. weighs the results by party alligence.

Bounce, and Zog. weighs the results by party alligence.Doesn't everyone weigh results by party allegiance? You have to in order to keep from having an unbalanced sample group. That's why the Time and Gallup results from earlier in the week were so far off--the Republicans were sampled at far higher rates than previous turnout would dictate.

Bounce, and Zog. weighs the results by party alligence.
That's not very effective. Both sides have their fair share of crossovers. They may be fewer this election, but as this may be a statistical dead heat, even a few swing voters will be enough to swing the election. My prediction is that, whoever wins the popular vote, he will do it with less than 100,000 votes.

That's not very effective. Both sides have their fair share of crossovers. They may be fewer this election, but as this may be a statistical dead heat, even a few swing voters will be enough to swing the election. My prediction is that, whoever wins the popular vote, he will do it with less than 100,000 votes.
No, not when you oversample Republicans to a large amount (they tend not to vote as heavily as Democrats do).

It is a statistical dead heat, and anyone else who says otherwise is too partisan to be helped.

That's not very effective. Both sides have their fair share of crossovers. They may be fewer this election, but as this may be a statistical dead heat, even a few swing voters will be enough to swing the election. My prediction is that, whoever wins the popular vote, he will do it with less than 100,000 votes.
The crossovers get factored in, because they'll generally self-identify late in the interview, after they've already expressed their candidate preference. In the sample sizes these guys work with, it generally averages out pretty well. It would be a mistake to try to estimate what percentage of crossover there will be from a single sample, but over the course of a polling cycle, it'll usually work itself out.

I have wildly different numbers: http://www.electionprojection.com/elections2004.html

Anyone with an opinion on accuracy? (I'm praying for your numbers btw)
George is puttin a whoopin on Kerry in that poll.

That's not very effective. Both sides have their fair share of crossovers. They may be fewer this election, but as this may be a statistical dead heat, even a few swing voters will be enough to swing the election. My prediction is that, whoever wins the popular vote, he will do it with less than 100,000 votes.
I don't know about less than 100,000. Certainly less than 500,000 though.

No, not when you oversample Republicans to a large amount (they tend not to vote as heavily as Democrats do).

It is a statistical dead heat, and anyone else who says otherwise is too partisan to be helped.
Hey, I'm not defending those polls. Right or wrong, I don't believe in any one of them. If I see a poll favoring Bush, I tell myself I'm dreaming. I'm confident that he will loose, especially now that John Kerry has started throwing red meat.

George is puttin a whoopin on Kerry in that poll.
I've long since stopped trusting that electoral counting site though. It has a very heavy republican bias running through it.

http://online.wsj.com/public/resour...leground04.html
I have wildly different numbers: http://www.electionprojection.com/elections2004.html

I wouldn't hold much credence for either of those two polls. The first one is based on on-line polling, "based on people who have registered to take part through solicitations on the company's Web site." In other words only people who bother to sign up are asked. Hardly represenative.
The second one comes from a very pro-Bush blog site. (Quote from the site: "Now that President Bush is clearly ahead in the polls, I'm going to try to temper my enthusiasm as I report this week's Election Projection."). This of course raises the question as to how objective and unbiased the results the blogger is using.

Ahhhh, can you smell that? It smells like another one-term Bush Presidentcy.

Mmmm, That's good lame duck.

Ahhhh, can you smell that? It smells like another one-term Bush Presidentcy.

Mmmm, That's good lame duck.

Smells more like cooked goose....Kerry's goose to be exact. ;)

How do they get a margin of error in counting votes? As in human error? I hope they're not doing some sort of sampling data to find the winner .... that would be daft.

It's a statistical thing.

- Redundancy Master of the Redunant