Polls

[opebo]

I was surprised to see Pennsylvania light blue for slight Bush when the last poll posted shows a 48% to 48% tie.

[rockhound]

I just noticed that Survey USA is showing MD in a tie.   I find that hard to believe.  Anyone know anything about their methodology?

[ElectionAtlas]

I added a new capability for 50-state polls.  The recent ARG 50-state poll from September is on-line.
Dave

[jdmoore]

There's a new national poll from the AP/Ipsos.

944 likely voters
50-46, Kerry over Bush, 3% M.o.E.

Unknown number of registered voters
47-47, 2.5% M.o.E.



http://www.cnn.com/2004/ALLPOLITICS/10/07/presidential.poll.ap/index.html

[classical liberal]

There have been no post-debate polls posted for IL.  I've seen 2:

a SUSA released 10/7 showing Kerry +16 (http://www.surveyusa.com/2004_Elections/IL041007pressen.pdf) and a R2k released 10/5 showing Kerry +17 (http://week.com/election/election-read.asp?id=5843).

[classical liberal]

I haven't seen this posted:

http://actforvictory.org/act.php/truth/articles/new_swing_state_polling_from_america_coming_together/

polls from FL, NM, MN, IA, and WI released 10/7

[opebo]

Why is it that Pennsylvania has been moved to tossup because of a tied poll, while Colorado has a tied poll in one of the last three but remains 'Leans Bush'?

[ATFFL]

Why is it that Pennsylvania has been moved to tossup because of a tied poll, while Colorado has a tied poll in one of the last three but remains 'Leans Bush'?

Colorado still averages over 5%.  If the average is over 5% it will show lean despite ties, or even the candidate losing in one poll.

[Siege40]

I just checked the poll section of the forum, looks like a huge Kerry surge... maybe. According to my excell chart (ya, I have my own chart to determine the winner, leave me alone) there's 17 EC Tossups (Minnesota + Iowa), 112 ECs within reasonable margin of error ( equal to or less than 4%). And for states with over 5% support in the latest poll the Dems have 213 ECs and the Reps have 196. I'm checking my numbers but it looks pretty good.

Siege

[CARLHAYDEN]

Recently there has been a rather significant divergence between polls from different pollsters as to the present intent of the voters.

While there are a number of possible explanations for this divergence, one which seems to me to explain much of the variation is in the way the particularly polling firm handles the 'likely voter.'

Lets take a concrete example of two radically different approaches.

Potential voter A is 50 years old, and has voted in the 2000, 1996, 1992, 1988.1984. 1980, and 1976 presidential elections.  He says he's leaning to vote for Bush.

Potential voter B is 23, has never voted in any election before in his life (was just registered) and declares that he hates Bush and will vote for "the other guy."

Some firms will look at the history of the voter to determine whether they are likely to vote this year, others will look at the 'intensity' expressed by the respondent to determine whether they will vote this year, and some consider both factors.

Those firms that use (at least primarily) historical voting to project who is more likely to vote, are more likely to give weight to voter A. than voter B.

Contrawise, those firms that use (at least primarily) 'intensity" to project likely voters, are more likely to give weight to voter B.

[Colin]

I find Rasmussen polls to be the most fair and the most accurate. They also do EV calculations and their site is very informative.

[CARLHAYDEN]

Rasmussen generally does a good job for the money, but he does have two signficant problems.

First, he weights his polls with a four point Democrat edge.  My data suggests that a one point Democrat edge would be more accurate.

Second, his state polls are often just subsets of his national polls, such that the data for his state polls can be from a two week period (not too reliable).

[Kevinstat]

Here's an interesting Maine poll from SurveyUSA.com:

http://www.surveyusa.com/2004_Elections/ME041020pres2q.pdf

In this poll, Kerry leads Bush by 6 points statewide (up from a 2-point lead in the last SurveyUSA poll), but by 9 points in the 2nd district (reversing a 9-point deficit in the last poll) and only 3 points in the 1st district (down from a 10-point lead in the last poll).  The second district is the one Bush came closer to carrying in 2000, and is the one pundits seem to expect he has a better chance of carrying this year.  So the results of this poll are a bit of a surprise.  This poll didn't seem to include candidates other Bush and Kerry in the question, but did leave an "other" option open.

The statewide result of the poll was Kerry leading Bush 51% to 45%.  In the first district, Kerry led Bush 49% to 46%.  In the second district, Kerry led Bush 53% to 44%.  Kerry polled best among those voters most likely to vote, although that bloc of voters was a far larger sample of the vote than the other bloc so that isn't saying a whole lot and may indicate that that bloc is inflated.

At the conservative web site AsMaineGoes, one poster commented that while he generally doesn't take much stock in polls, he found this polling organization particularly unreliable.  Another poster agreed, and it was discussed how the same polling firm overestimated support for a school funding initative by 19 points shortly before the election.  That was a primary election, however, and there weren't any top-ticket congressional primaries this year, and one poster pointed out that the low turnout in that election made it a very difficult one to poll.

Sincerely,

Kevin Lamoreau

[alcaeus]

   Kerry moved ahead in the Rasmussen poll today 48- 46 percent.   Rasmussen uses a five percent margin of error for states, and therefore has more tossup states than other polls.

     Most national polls are tracking even with greater frequency.   The national polls look identical to the '00 election at this stage.  However Bush is getting more votes in Bush states and Kerry is holding onto more states than Gore.

     

[A18]

Bush is goin to lose the election by a landslide.  Trust me. hehe.  I'm a republican, but I cant stand bush.  He grew up with a silver spoon and had an easy life.  He was a deserter to.  Bush has gotten away with everything easy.  Bush is a moron also. 

So Bush is going to lose by what again?

[Alcon]

Bush is goin to lose the election by a landslide.  Trust me. hehe.  I'm a republican, but I cant stand bush.  He grew up with a silver spoon and had an easy life.  He was a deserter to.  Bush has gotten away with everything easy.  Bush is a moron also. 

So Bush is going to lose by what again?

Wow...May 28th. Reaching a bit there, Philip?