Didn't Gallup have Bob Casey up by double digits in a recent poll?

Zogby interactive polls are all over the place.

Zogby Interactive... the same online polling outfit that showed Kerry within three points of Bush in Tennessee on Nov. 1, 2004.

Right.

Why do people discuss Zogby Interactive without an appropriate disclaimer first?

This polling looks a bit more accurate then the past polling which had cater with the margin of error and now, they have carter down by double digit, which is about right.

I think the webb-allen polling is off by a few points..I dont believe that allen is down by 7..Ill say allen is down, but within the margin of error.

The lieberman-lamont result is just about right..Lieberman is ahead, but i also think that lamont is with the margin of error, and because lamont has the backing of the labor's union and the party, he will win this with good GOTV which lieberman will no have.

The tx-senate is probably off.I dont believe barbara is down by single digit and let me say that i wish it was so because the republicans would have to spend money there instead of not spending a dime on this battle.
I wish barbara the best and pray that a more credible poll shows her within striking distance.

As usual, Zogby entertainment interactive polls are completely the opposite of what pretty much everyone else says.

blehh. I hate this Zogby shit.

I am always somewhat skeptical of Zogby's numbers.  I can't believe that Webb is that far ahead of Allen although it would be great to win that one.  Each cycle there is almost always one upset that nobody really believes is going to occur.  Hopefully this is the one.

And what's up with Casey?  If I were running that campaign, it would quotes from Santorum's book all the time.  He did you oppo research for you.  Use it!

corker-ford: i believe in this battle being close, but because ford is black and a democrat, people will think twice about voting for him.I wish it wasnt so, but you have those kind of folks that just wouldnt trust a black guy to represent them.
Some of them who were polled,probably said they would vote for ford because of corkers campaign blunders, but once they get in the booth to vote for ford,it might be a different story.

Talent-mccaskill: for claire to win, she will have to get out the vote in those heavy democratic cities like st louis and others...Bill clinton came to campaign for her this weekend and this should re energize the base, but she better get to work and make sure that as many people gets to vote...If she gets a strong showing from those huge cities, then talent is toast even if he carries the burbs by a good margin..

pederson-kyl: i wish those numbers were true.Pederson down by single digit is somewhat of a good news in my opinion..I think pederson is down by double digit..Maybe 10-11, no more no less.

Or Mason-Dixon, for that matter? Stick with Ras, SUSA, Research2000, ARG and Quinnipiac for the states they cover for the best results. Even Strategic Vision is better than these Zogby interactives.  

Aren't these polls biased and not valid statistically? Frankly isn't the same poll touted by Pennachio's supporters earlier in the year that their candidate could beat Santorum too?

My stats Prof would freak. This polls are seriously flawed just read this: "Zogby Interactive of Utica, N.Y. has assembled a database of individuals who have registered to take part in online polls through solicitations on the company's Web site as well as other Web sites";in other words, this is NOT random polling at all. It is key to acctually try to find a random sample; this poll does the exact opposite--it makes you come to them. I would imagine that if you took a closer look at that math the margin of errors would be more like 15-20. I am not a whiz but after taking a stats class it is much easier to see flaws in these polls.

I've seen a lot of Zogby polls and I'm still very skeptical.

Has he gotten it right in any race?

phat

For a long time, it was my job to take polls and make sense of them for Dem candidates and create message.  I moved on to another field in 1998, partly out of frustration with the "establishment," a group I was trying to crack but for whom I had little respect because of all the flaws blogs like this one have been busy pointing out over the last few cycles.

But I still have the bug and read sites like this often.  This is my first comment on any political blog.

Zogby's methods are suspect, and I'm most troubled by the CT Sen numbers.  While Zogby admits he "adjusts" his interactive numbers to account for (basically unknowable) on-line sampling error vis a vis other traditional polls, based on ancedotal evidence, I'd expect Lamont to be whooping on Lieberman in a poll that is dependent on internet response.  All the evidence says that the active on-line community is overwhelmingly in favor of Lamont.  No one outside of Zogby's firm really know how he "adjusts," but given the advantage he's finding for Webb in VA, Lamont's people should be worried.

I suspected that Lieberman's own poll released on the 10th were off; the numbers suggest that it's either an outlier or that it was conducted in a way that would give results that could be released for political and fundraising reasons, but now I'm not so sure.

I look forward to Quinnipiac's next survey on this race (despite their woeful pre-primary poll that had Lamont up by 13) and Rasmussen, despite my concerns over their robo calling method.

"Casey has seen his lead over Santorum close to 47-43" I'm not sure if this says more about the incompetence of Casey, Schumer of people who live in PA. Santorum is a nut case. If he wasn't a Senator, he would have been committed to an institution long ago.

We always question the methodology -- that's why you call it candy -- and results of these polls because of it, but they also get the basic info wrong.  Jon Corzine is not the incumbent Senator in New Jersey.

Try these:

http://www.electoral-vote.com:2006/

Does anyone know?  Zogby shows Webb in front of the MOE.

On the Senate side, I just plain don't believe Zogby in AZ, TX, VA, MD, or PA.  

Nobody else shows AZ, TX, and PA as being nearly that close, and I can't think of a reason to believe it in any of those three cases.  Nobody believes Mfume, should he beat Cardin in tomorrow's primary, is 11 points ahead of Steele. And nobody believes Webb's running 7 points ahead of Allen.

That's 5 out of 18 that don't pass even a crude 'smell test.'

And not that it matters much, given the unreliability of the polling, but why does Zogby poll the Senate races in TX, NM, WI, NY, and FL, none of which have a remote chance of changing hands, and not poll MT and RI, which are 2 of the 4 best Dem pickup opportunities?  Makes no sense at all, but fortunately, it doesn't really matter.

He believed early in the cycle that both incumbents were safe, and would rather continue pretending than to admit a mistake (this is the same guy who swears up and down that his polls weren't bad in 2004, remember).

There may also be a methodology problem - ie, he may not have enough potential respondents from those states to form a large enough pool from one month to the next. You'd think NM and NV would have the same problem, though.

Casey is running good ads in Pennsylvania, but he's not yet willing to go for the jugular.  

He needs to remind the voters that Santorum does not live in Pennsylvania.  He needs to ask them if they want to be represented by a Pennsylvanian or a Virginian.  

And then run film of Santorum saying that Pennsylvania should be represented by someone from Pennsylvania, which he said when he made his winning run for the Senate.

I swear that Casey doesn't run like he wants the seat.