From the thread on polling:
I think this would be the part of his post that I would like to see you address specifically before you dismiss him as a “klutz.”
The only poll I really pay attention to is Zogby. Not that I think Zogby is necessarily correct, only that it seems to be highly-regarded as to polling science. Supposedly, it maintains its independence well, controls for factors, and asks “neutral” questions better. Less opportunity for bias.
I wouldn’t trust a poll from USA Today for anything. CNN might be a little better. I’m not sure what Gallup is doing slumming with these two (especially USA Today. USA Today???–Oh, yeah, that’s good news. :rolleyes: ). Gallup used to be trustworthy.
Other polls, including those done directly for one campaign or another, or those done by PACs or interest groups (conservative or liberal) should be taken with a grain of salt.
Al Gore himself, From Earth in the Balance:
I am not trying to debate the right or wrong of his positions, only that I am sure auto workers would have a hard time getting excited about bashing cars.
I think much of the current polling is “soft polling” measuring peoples reactions to specific events (Bush’s malapropisms or Gore’s exaggerations)…and not so much who the pollee really wants to vote for…
Someone in an earlier thread pointed to an interesting Washington Post piece about political scientists who predict presidential election outcomes based entirely on economic conditions, current events and peoples attitudes about the economy. They al predicted a Gore win…and at least in recent memory have been quite accurate.
One of the more interesting "polls’ is the Iowa Electronic Markets, where investors purchase “shares” in a candidate based on projected final popular vote percentage
http://www.biz.uiowa.edu/iem/
In the Winner Take All category , Gore is trading at 58.6 and Bush at 41.7
In the Vote Share category, Gore was trading at 49.7 and Bush at 48.7
These markets tend to be pretty good predictors of final outcomes, because people invest real money in what they THINK the outcome will be…
I can’t get the link to work, so I cannot comment upon the site as a whole, but this:
Can be criticized on a couple of levels.
Mathematically: If the shift in question is supposed to be acounted for by the shift of male republican respondents to female democratic respondents, then the magnitude can be no greater than the magnitude of teh smallest element of change. In this case, 4.5%. Even if we assume that every additional democrat polled was a female and the entire loss of republicans polled were male, we can get no higher than 4.5%. Of course, I think it unlikely that either of those assumptions is correct, meaning the actual shift in this area was smaller thatn 4.5%
Practically: As someone once worked for a market research company, I can tell you that the assumption that a company begins a poll by selecting a specific number of responses required from specific demographic pools and then calling folks until each pool is filled is extremely unlikely/nonstandard.
A pool of potential numbers/adrresses/whatever is selected and then contacted. Whoever from among that pool chooses to respond/answers the phone/etc. and is not disqualified from the study due to their responses gets tabulated. If poll2 has more democrats in teh breakdown than poll1 it is almost certainly attributable to either a demographic spike in the pool of potential respondents or (more likely) a factor of who happened to answer the phone that day.
And I will bring up the same counter-argument I brought up in that thread.
First, the article stated that these simulators ‘accurately predicted the results of elections since 1952 in which the winner received 53% or more of the vote.’ (I don’t have a link right off hand; I’ll try to find one.)
Since 1952, there have been 12 elections. Of those, only 6 have seen the winner receive 53% or more of the vote. In other words, a coin flip is as accurate as these simulators. In fact, were I to add a simple statement to a coin flip (“In cases where a President is running for re-election, he wins; otherwise, flip a coin”) I can on average accurately predict 7 out of 12 races, meaning I’m more accurate than these simulations, but I’m still only as accurate as “A Republican will always win”, which also predicts 7 out of 12 of these races.
Second, on that same note, think about it. Take that ‘in which the winner received 53% or more of the vote’ and look at it. In other words, the simulation accurately predicts the outcomes of all races which are not close.
Third, these simulations generally follow the same rule- throw in economic numbers, and get an output. Adjust by a Mystic X factor which turns that output into the correct number. What’s the Mystic X factor? Cynically, whatever you need to turn the output into the correct number. Non-cynically, an unpredictable factor. Even the study the Post cites depends upon the “time for a change factor”. So the economic numbers ensure a Gore election, unless the scientists underestimate the value of an unknowable factor. This is science?
So, in toto, what you have is this: “Our formula, based solely upon economic data except for a “time for a change” variable which we decide upon after the fact in order to get our data to match the election result, has computed that Gore will be the winner (assuming we assumed the variable correctly). Our formula is completely correct in predicting winners except in close races like this one, but we stand by it even though extrapolating figures into the past at tweaking the formula to make it correct is no assurance that future results will be correct.”
Yeah, I’d definitely put my faith in that.
The Washington Post has a bit on the “wacky” swings of the Gallup Poll lately
http://washingtonpost.com/wp-dyn/articles/A51737-2000Oct11.html
As far as those ‘wacky swings’ go, you have to remember that these polls have an error margin of around ±4%. That means that if the first poll showed Gore leading Bush by 51% to 43%, it could actually have been as low as 49% to 45%.
Likewise, the current poll showed Bush leading 49% to 41%. With the same margin of error, that could be as low as 47% to 43%.
If the errors in each poll turned out to be on the opposite ends of the spectrum, then what we’d really have is Bush gaining 2 points and Gore dropping 8 points. That’s not a huge swing for a three week period before an election with the first debate in the middle of it.
Freedom2’s link just confirms that this election is going to be won or lost in the states of Florida and Pennsylvania. Both are still toss-ups, both carry enough electoral votes to swing the election one way or the other.
I understand both campaigns are currently focusing their efforts on these critical states.
Also, it would not surprise me to see Bush win the popular vote, with Gore still winning the election based on electoral votes.
Let me explain:
When I was in high school, my US History teacher, who was big on analytical methods, spent a whole class building up, quite rationally, defensibly, and step by rigorous step, to the conclusion that…(roll drums)…the South viewed Reconstruction differently than the North did.
Glad I wasn’t drinking a soda or something at the time. 
So this guy says Gallup was trying to manipulate the results to make people think that Gore won the debate. My history teacher’s argument supported a conclusion that didn’t need any support, but at least it supported it. This klutz’ pile of detail had nothing to do with his conclusion; if Gallup manipulated the data, it was to give the impression that Gore’s popular support was increasing before the debate, which if anything magnified the effect of Bush’s post-debate bounce in the polls.
I’ve been relying on the Battleground 2000 poll conducted by voter.com. It’s a continuous rolling sample conducted by (gasp) a Democratic pollster and Republican pollster working together. They haven’t had the wild swings of other polls and claim that their final predictions for the last 2 elections have been accurate within one percentage point of the final result. Portraitofamerica.com is another good site–they have Bush ahead in electoral college votes, for whatever that’s worth. I assume they’re not yet counting Florida and I don’t know if they’re counting California in Gore’s column.
My prediction:
Cheney - 36.5% (If you’re a Republican, this is the only way you can legitimize your vote.)
Liebermann - 32% (Change the above from “Republican” to “Democrat”)
Gore - 12% (There are people in the U.S. who would vote for John Buttwipe if he had a “D” next to his name.)
Bush - 10% (Change the above to read “R” next to his name.)
Nader - 6% (He’s the only presidential candidate who has yet to make an ass of himself.)
Dave Barry - 2% (He’s got my vote!)
Unca Cecil - 1% (So I’m sucking up to the big guy. Sue me!)
Lyndon LaRouche - .5% <—He’s got at least ONE doper’s vote! (Not that I mentioning names)
- I don’t see anything about $4/gal. gasoline.
- I’m also missing the part about eliminating the internal-combustion engine.
- He seems to be suggesting that we should ‘think strategically about transportation’, part of which involves building and using more fuel-efficient vehicles. His attitude seems to be like that of a Martian anthropologist (hey, always possible!), looking at our everyday actions from a perspective that we normally don’t, but in an apparently factual manner. I don’t see how that constitutes ‘bashing cars’.
I heard winners all die young somewhere. Now where the hell is that landslide that is going to take Bush?
HUGS!
Sqrl
I’d have to agree with John Corrado’s critique of the Iowa Electronic Markets, and pile on a bit.
They’re based on the investors’ awareness of what other people are thinking, and we’ve got a society that is, in some ways, fairly divided. So we don’t always have a clear idea of what’s going on with people who don’t see the world similarly to the way we do.
Which brings us to 1994, where the IEM, if I’m reading their tables correctly, didn’t permanently put the GOP ahead in the contest for the Senate majority until October 29, and never did predict the GOP House takeover until it was history.
I think the divisions I mentioned above may have had something to do with that lack of perception. I was teaching at an evangelical Christian college that year, and the students…when they talked politics that fall, they weren’t just optimistic; they were confident to the point of elation. They knew it was their year, and it turned out they were 100% right. But had I been in most other places I might have been, I wouldn’t have seen it coming; I wouldn’t have been tuned in.
So add me to the ranks of skeptics. It’s a nice idea, but it needs to be taken with a grain of salt.
Quoted from Al Gore’s book, “Earth in the Balance” by RTfirefly:
Considering that gas weighs in at about 6.17 pounds per gallon (according to a search using ASK.com) I find it highly likely that unless RTfirefly misquoted the vice-perpetrator, that my information was incorrect, or that the analysis below is flawed, that we have another perfect example of Gore mistating merely to make a political point.
To all you chemists/science types is this possible? It seems to me you would have to assume 100% conversion of the gas to carbon and then each atom of carbon combining with two atoms of oxygen with both carbon and oxygen having the same weight. I didn’t take chemistry but this just seems very unlikely to me, but I am more than willing to retract upon superior info. Thoughts?
John Corrado was criticizing the claims of the political scientists in the Washington Post bit about election predcitions. He did not (as far as I could see) offer an opinion on the IEM.
The success rate “I” discussed was in reference to Presidential elections…which have a different dynamic than congressional elections…I was making no claim about accuracy in other events (they even have investment markets in Microsoft stock value for example)
It is pretty close to 19 pounds of CO2 per gallon of octane. Let’s go to the numbers:
6.17 pounds of octane per gallon * ( 96 AMU of arbon / 114 AMU of octane) =
5.196 pounds of carbon per gallon of octane
Assume all of the carbon is turned into CO2.
Each atom of carbon produces one molecule of CO2.
Atomic carbon weighs 12 AMU per mole and CO2 weighs 44 AMU per mole.
5.196 pounds of carbon per gallon of octane * (44 AMU of CO2 / 12 AMU of carbon) =
19.05 pounds of CO2 per gallon of octane
It didn’t make sense to me at first either, until I crunched the numbers.
This should read as AMU per molecule or atom, not per mole.
Dr.Lao -
My thanks to you for setting the record straight. I am a firm believer in no-spin so I retract my above post in its entirity and offer my apologies to the vice-perpetrator’s followers 
=========seeks cover=========