Non-Gallup/Rasmussen tracking polls thread

[Zarn]

Maybe before each Zogby and Kos update, we put in the post "Warning: Poll is for humor only."

At least, they figured out about the nifty invention of the telephone.

[J. J.]

Maybe before each Zogby and Kos update, we put in the post "Warning: Poll is for humor only."

At least, they figured out about the nifty invention of the telephone.

I've actually suggested something like that (which is ironic because both have shown tending to McCain).  Both are still however bad.

[Wall St. Wiz]

I feel great now, the only poll showing this a close race is from the guy who actually changed his prediction at 5pm on election day in 2004 (and completely missed the mark).

I hope there are no McCain supporters allowing this to get their hopes up.

[Sam Spade]

Diageo Hotline Poll - 10/19
Obama 48% (-1)
McCain 41% (-1)

[Tender Branson]

The consensus is Obama by 5-8 now.

If we re-weight Zogby, where Obama leads by 8 among Independents, they would also show Obama ahead by 5/6 points.

[J. J.]

The consensus is Obama by 5-8 now.

If we re-weight Zogby, where Obama leads by 8 among Independents, they would also show Obama ahead by 5/6 points.

I'd say Obama +6.

[Zarn]

RCP (for what its worth) says -5.0

[TheGlobalizer]

RCP (for what its worth) says -5.0

I'd say the truth is around O+4, based on LV sampling skews.  Trending ever-so-slowly downward.

[Tender Branson]

RCP (for what its worth) says -5.0

Yes, because they have thrown out all of the (older) non-tracking polls from their average.

With today's Gallup numbers it will increase to about 6%, especially if we include R2000 as well.

[Zarn]

RCP (for what its worth) says -5.0

Yes, because they have thrown out all of the (older) non-tracking polls from their average.

With today's Gallup numbers it will increase to about 6%, especially if we include R2000 as well.

Say NO to Kos!

[Rowan]

IBD
Obama 46.6%(-.60)
McCain 41.5%(+1.7)

http://www.realclearpolitics.com/articles/docs/2008-IBD-TIPP-DAY7.htm

[J. J.]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos?

[J. J.]

IBD
Obama 46.6%(-.60)
McCain 41.5%(+1.7)

http://www.realclearpolitics.com/articles/docs/2008-IBD-TIPP-DAY7.htm

Probably just a very pro-Obama sample dropping off.

[Verily]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Yes to Zogby, no to R2000. But I bet if R2000 was more favorable to McCain and Zogby less so, they'd been amenable to switching. It's their policy with ARG to only include McCain-friendly polls, after all.

[Sam Spade]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Zogby, not Kos.

FWIW -

Rasmussen O+6 (O+1)
Gallup O+3/O+7 (O+1/O+3)
Hotline O+7 (nc)
IBD/TIPP O+5 (M+2)
Zogby O+3 (M+1)
Kos O+7 (nc)
GW/Battleground O+4 (old)

Average:  O+5.29

[Wall St. Wiz]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Zogby, not Kos.

FWIW -

Rasmussen O+6 (O+1)
Gallup O+3/O+7 (O+1/O+3)
Hotline O+7 (nc)
IBD/TIPP O+5 (M+2)
Zogby O+3 (M+1)
Kos O+7 (nc)
GW/Battleground O+4 (old)

Average:  O+5.29

You would think with the huge Obama money advantage, the market woes, and the general environment for Republicans, Obama would be up by a lot more then 5.

If McCain could pick up another 2 points in the average, election day could be interesting.  I'm not betting on it or expecting it though.

[J. J.]

Out of these, watch Rasmussen, TIPP, and Gallup.

My guess is that it's someplace between the turnout models on Gallup.  A pro-Obama sample probably dropped out of TIPP, accounting for the move.  I'll be pessimistic and say Obama +6.

The trend, if any seems to be for Obama.

[Lunar]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Zogby, not Kos.

FWIW -

Rasmussen O+6 (O+1)
Gallup O+3/O+7 (O+1/O+3)
Hotline O+7 (nc)
IBD/TIPP O+5 (M+2)
Zogby O+3 (M+1)
Kos O+7 (nc)
GW/Battleground O+4 (old)

Average:  O+5.29

You would think with the huge Obama money advantage, the market woes, and the general environment for Republicans, Obama would be up by a lot more then 5.

If McCain could pick up another 2 points in the average, election day could be interesting.  I'm not betting on it or expecting it though.

The should be argument is ridiculous.  I could say that running against a liberal black guy during a time of war, a moderate war hero should be up too.  Usually these arguments are made by ridiculous partisans spinning the facts

[Gustaf]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Zogby, not Kos.

FWIW -

Rasmussen O+6 (O+1)
Gallup O+3/O+7 (O+1/O+3)
Hotline O+7 (nc)
IBD/TIPP O+5 (M+2)
Zogby O+3 (M+1)
Kos O+7 (nc)
GW/Battleground O+4 (old)

Average:  O+5.29

You would think with the huge Obama money advantage, the market woes, and the general environment for Republicans, Obama would be up by a lot more then 5.

If McCain could pick up another 2 points in the average, election day could be interesting.  I'm not betting on it or expecting it though.

The should be argument is ridiculous.  I could say that running against a liberal black guy during a time of war, a moderate war hero should be up too.  Usually these arguments are made by ridiculous partisans spinning the facts

Whether it is ridiculous or not depends on how it's framed.

[Sam Spade]

RCP (for what its worth) says -5.0

Does that include Zogby and Kos.

Zogby, not Kos.

FWIW -

Rasmussen O+6 (O+1)
Gallup O+3/O+7 (O+1/O+3)
Hotline O+7 (nc)
IBD/TIPP O+5 (M+2)
Zogby O+3 (M+1)
Kos O+7 (nc)
GW/Battleground O+4 (old)

Average:  O+5.29

We have the seven tracking polls.  Last time around, the ABC/WP also did a tracking poll, but I guess that's not going to be the case this year.  Anyway, each of these polls have issues with me and two of them have no record (which is another problem), so just a few words:

Gallup
Purest LV model (traditional).  Has a tendency, historically, to underestimate one side's number if that side has weak enthusiasm.  Also can overestimate last minute movement. 

The new expanded model is a curiosity to me, mainly because I can't figure out which of the 13 questions they're removing.  It also seems to be a bit jumpy, which is what makes me a little leery of it, but who knows.

Gallup performed quite well in 2004, so long as they don't break down the undecideds the way they did (wtf?)

Rasmussen
In 2000, he had too much of a Republican weighting, plus an outlier sample at the end = disaster.  2004, he switched to a hard-weight of 39D, 35R, 26I and nailed it.  This year, he's using a dynamic weighting method that pretty much no one else does.

The dynamic weighting model is my only really big question concerning his polling methodology, frankly.  Otherwise, he looks to be polling using similar methods to 2004 when he performed quite well.

IBD/TIPP
Uses a hard weight of 39D, 35R, 26I.  Their methodology has always been pretty good in my mind.  My main concern right now is the smallness of their samples.  Even with the weighting, we're getting a lot of bumping and some silly internals.  Truthfully, as I might sound silly saying this, the Zogby/Hotline method of 300/400 voters would probably add a bit more stability to their poll.

Nailed 2004, as we all know.  Wrong winner in 2000 (as many polls did), but well within MOE, which is the key.

Zogby
Uses a hard weight of 38D, 36R, 26I (I'm pretty sure - it's the 2006 exit poll number).  Uses a whole lot of other weights and is well-known to occasionally massage out certain numbers if he thinks the poll should really read that result.

Zogby is the wildcard.  I don't particularly like his methodology (whatever one there is) and the polls often more represent his opinion than his polling data.  He's been a mile off before.  He's also nailed races that almost everyone else got wrong.  In particular, in 1996 and 2000, he was saying that the polls were way too Democratic (1996) or way too Republican (2000) than what the reality would be on election day.  Of course, in 2004, he changed his (actually, it would have been pretty accurate) prediction based on exit poll rumors (lol).

GWBattleground
Dead-on accurate in 1992 and 1996.  Off (within MOE) in 2000.  Slight Republican bias in 2004.

Of course, with the elimination of the party weighting, this poll really isn't the same as it was in the past - whether good or bad.  The rest of their methodology has always been fairly strong.

Hotline
No record.  I think their methodology is a little odd, actually kind of similar in its oddities to the Kos poll (but without the political issue).  We'll see how they perform.

R2000/Kos
See Hotline, except add political issues.  Also add the standard R2000 slight Dem bias.

[Lief 🗽]

R2000/Kos
See Hotline, except add political issues.  Also add the standard R2000 slight Dem bias.

You people are annoying.

[ChrisFromNJ]

R2000/Kos
See Hotline, except add political issues.  Also add the standard R2000 slight Dem bias.

You people are annoying.

No mention from Sam about Zogby push polling for the McCain campaign.

[Alcon]

And Zogby is a Democrat, and I'd though R2K's founder was actually a GOP-leaning independent.  I think we're putting too much stock into the R2K/Kos-as-political-propaganda idea.

[J. J.]

And Zogby is a Democrat, and I'd though R2K's founder was actually a GOP-leaning independent.  I think we're putting too much stock into the R2K/Kos-as-political-propaganda idea.

The only problem I have with it is the weighting.

[Alcon]

And Zogby is a Democrat, and I'd though R2K's founder was actually a GOP-leaning independent.  I think we're putting too much stock into the R2K/Kos-as-political-propaganda idea.

The only problem I have with it is the weighting.

Yea, OK, cool, so basically it's like Hotline, but with objectionable weighting.  I actually agree with Spade that Hotline has some strange methodology quirks too, so I put them basically in the same caliber (question, maybe tinker, but don't ignore)