Fox News/Opinion Dynamics poll: 42-42

[Spin Police]

First, Zogby has a real problem with accuracy.  Of the major polls in 2002, he had far and away the highest margin of error and far and away the highest wrong prediction.

Not sure if you meant 2002 (Congressionals) or 2000.  I'll presume you meant 2000.  His final poll in 2000 was 48% Gore, 46% Bush.  While imperfect, this (Zogby) was one of the BEST polls!!!

TIPP was 48-46 in favor of Bush.  Hotline was 45-42 Bush.  ABC was 48-45 Bush.  Gallup was 48-46 Bush.  Marist was 49-44 Bush.  NBC/WSJ was 47-44 Bush.  Newsweek, the supposedly Democratic bastion, was 45-43 Bush.  Pew was 49-47 Bush.  CBS News was 45-44 Gore.  CBS/NYT was 47-42 Bush.  Fox/OD was 43-43 tie.  Harris was 47-47 tie.  ICR was 46-44 Bush.  Tarrance/Lake was 50-45 Bush.  (These are all the final polls by these groups... taken in early November.)

This is quite obvious... Zogby was one of only four polls showing a tie or a narrow Gore win.  That's out of 15 polls!!!  And of the four that did well, one of which, admittedly, was Fox/OD (the ones I'm harping on), Zogby was probably not the worst.  Fox/OD was, due to its high undecided.  Harris was the best (lowest undecided, with a 47-47 tie)... followed by CBS News, then Zogby, then Fox/OD.  The other 11(!!) may have still caught the MOE to avoid being "wrong" statistically... but they weren't good... and that includes some pretty big names, like Gallup, TIPP, and Tarrance/Lake.

State by state, if I recall, Zogby was also good (though he did mess up the ever-critical Florida).  But clearly, overall, he was one of the best.  Also, check out the comparisons to other polling companies that Survey USA has conducted.  They do reasonably well, and...  One of the pollsters they have trouble beating?  Zogby.

In the 2002 cycle Zogby polled 17 races.  He got 5 wrong.

Zogby is often right, but also often wrong.

[CARLHAYDEN]

First, Zogby has a real problem with accuracy.  Of the major polls in 2002, he had far and away the highest margin of error and far and away the highest wrong prediction.

Not sure if you meant 2002 (Congressionals) or 2000.  I'll presume you meant 2000.  His final poll in 2000 was 48% Gore, 46% Bush.  While imperfect, this (Zogby) was one of the BEST polls!!!

TIPP was 48-46 in favor of Bush.  Hotline was 45-42 Bush.  ABC was 48-45 Bush.  Gallup was 48-46 Bush.  Marist was 49-44 Bush.  NBC/WSJ was 47-44 Bush.  Newsweek, the supposedly Democratic bastion, was 45-43 Bush.  Pew was 49-47 Bush.  CBS News was 45-44 Gore.  CBS/NYT was 47-42 Bush.  Fox/OD was 43-43 tie.  Harris was 47-47 tie.  ICR was 46-44 Bush.  Tarrance/Lake was 50-45 Bush.  (These are all the final polls by these groups... taken in early November.)

This is quite obvious... Zogby was one of only four polls showing a tie or a narrow Gore win.  That's out of 15 polls!!!  And of the four that did well, one of which, admittedly, was Fox/OD (the ones I'm harping on), Zogby was probably not the worst.  Fox/OD was, due to its high undecided.  Harris was the best (lowest undecided, with a 47-47 tie)... followed by CBS News, then Zogby, then Fox/OD.  The other 11(!!) may have still caught the MOE to avoid being "wrong" statistically... but they weren't good... and that includes some pretty big names, like Gallup, TIPP, and Tarrance/Lake.

State by state, if I recall, Zogby was also good (though he did mess up the ever-critical Florida).  But clearly, overall, he was one of the best.  Also, check out the comparisons to other polling companies that Survey USA has conducted.  They do reasonably well, and...  One of the pollsters they have trouble beating?  Zogby.

First, I meant EXACTLY what I said.  

Second, read the review from the National Council on Public Polls for the 2002 elections.

Third, Zogby is a highly partisan Democrat who had a brother working in the Clinton administration.  He had inside information that the Democrats were running a very effective drive in 2000 to get their cattle to the polls.  If that cattle drive had not been so effective Bush would have won by a couple of points in the popular vote.

[classical liberal]

Everyone is talking about how some poll shows one candidate up in the "battleground states".  But who is defining "battleground"?  What states are included in the "battleground state" numbers?

[Spin Police]

Here are some actual facts from Congressional/Governor races in 2002.. As provided by the National council on polling.



Highest average margin of error = Zogby
Highest % of races incorrectly predicted = Zogby

Looks like last place for Zogby to me.. anybody wanna disagree...?

Lowest Average Error = Gallup
Best "winning" % = Mason Dixon
(Excluding Quinnipiac which only did 4 races)

Either Gallup or Mason Dixon in first place depending what you look at.

Anybody wanna argue?

These are the facts.

Read 'em and weep as they say...

Vorlon wrote about a 3 pager post on Zogby about 2 months ago and described him as "brilliant but erratic" - Even gave him his own "category" of clasification.

Seems about right to me.

Gave Gallup and Mason Dixon his highest ratings.

Seems about right to me.

We can argue "bias", the number of angels that can dance on the head of a pin, the meaning of life, all we want.  

At the end of the day it's what you predict, and what the actual numbers turn out to be.

Mason Dixon = Good
Gallup = Good
Zogby = Unreliable.

Anybody wanna disagree based on the FACTS...?

You must be logged in to read this quote.

[CARLHAYDEN]

Good work, SpinPolice.

Two points which most polls fail to report is the 'hardness' of the support for the candidate, and the turnout prediction.

Both of these factors are interrelated.

People are reluctant to say "undecided" when they in fact are undecided as they believe that this would be interpreted as meaning they aren't following the campaign (which many of them are not) so they just say "yes" to one of the names offered.  Not wishing to totally lie, if asked, they admit that they are "leaning" to whatever candidate.

In addition, a much higher percentage of respondents say they will be voting in the upcoming election than will actually be the case.   Interesingly enough, far more respondents in post election polls say they voted in the election than actually did.  Again they don't want to admit they didn't do what they have been told is their 'civic duty.'

Every poll I have seen over the past couple of months has indicated that the 'hard' Bush support is greater than the 'hard' Kerry support, and that most of the hard 'Kerry' support is in fact merely anti-Bush support (anybody but bush).

Also, Kerry's seeming ambivlance on the issues is in part based on the fact that the non-bush vote (i.e. not 'hard' for bush) is almost evenly split between hard lefties and moderates who agree more with Bush on the issues than they do with Kerry's pre-2004 stance on the issues.  

Kerry cann't afford to enrage the moderates but cann't really bring himself to renounce his real (long term) policy preferences, hence the stupid sounding double-talk.

[Spin Police]

Fox/OD hit the 2000 election perfectly.  48-48.

Looking at the Fox/OD numbers from 2000 right now.  They still had 10% undecided in the last poll.  They showed 43-43.  Still, not pro-Bush, I will grant you.  But I wasn't commenting on that (too difficult to assess w/ the large undecideds).  That's why I used the low-undecided, easy to compare approval ratings since 2000.  There's simply no way around it... Fox/OD is almost always a fringe or outlier pro-Bush.  Just as Newsweek is pro-Dem.  I notice the only folks debating this are Republicans.  If I come on and rip Newsweek when they show a pro-Kerry bias (which, actually, they've not been too bad of late... but that's only a recent trend), I bet the Dems will chime in.  I'm NOT trying to be partisan here... just factual.

In fact, let's just jump back a ways and randomly select a longer-term period (so you can see I'm not cherry picking) and average their polls.  For all of 2003 (averaging the polls per month, so multi-polled months don't skew the data too much), the average approval rating for Bush...

Fox/OD = 59%
Newsweek = 56%
Time/CNN = 56%
Zogby = 54%
CBS = 59%
Gallup = 59%
Pew = 58%

So... Fox/OD an outlier?  No.  But, they ARE right on the edge.  Coincidental that they happen to be on the high edge?  I doubt it, since I earlier showed that they've been on the high end all this year as well.  If it was mere chance, it'd balance out.  For example, in all three most recent, CBS, one of the other 59%, was lower than Fox/OD.  In two of the most three recent Gallup was also lower.

I haven't gone through a "statistical significance" on this, but since we're talking about the entirety of 2003, and up through the first five months of 2004, this is a LONG trend.  I seriously doubt this can be written off to chance.  This is especially true when you dig into the data even more... CBS and Gallup (the only other two at 59%) had more late-March '03 polls, after a jump in Bush's approval, pushing their March averages up more than the Fox/OD.

Plain and simple... Fox/OD leans pro-Bush.  Its trends are trustworthy; its raw numbers are not.

Or....

(stating the incredibly f**king obvious)

All of the polls have very slightly different wordings of the questions and/or ask them at different stages of the interview.

Might not these factors easily explain a difference that is, after all, only 3%? (excluding Zogby) without resorting to any grand theories of poll manipulation?

Also, some are likely votes, some are all adults and/or registered voters.

Typically "Likely" voters are about 3% more Republican than Registered voters.

MIght this also not explain the 3% differences ?

The Zogby question is also TOTALLY different, which explains the different result for him.?

[Rococo4]

First, Zogby has a real problem with accuracy.  Of the major polls in 2002, he had far and away the highest margin of error and far and away the highest wrong prediction.

That's true, I think he predicted that Mondale would defeat Coleman, Johnson would level Thune, Dole/Bowles would be a nail-biter and Jeb Bush was in serious trouble in Florida.

Im no Zogby fan, but he had Jeb up by like 15% the weekend before the race on "Meet The Press"....one of his only correct calls. But people thought he was crazy on this one, and he ended up being right

[Smash255]

This is when most other polls showed Gore as up by 2%-3% margins.

really?  Most polls had Bush up.

http://www.ncpp.org/poll_perform.htm
http://www.pollingreport.com/2000.htm

[Smash255]

Not to "rag" on Fox, since I'm independent and try to play nice and non-partisan on this board, but they are generally thought to have a conservative leaning.  And I have noted in the past three years that they USUALLY (not always, but usually) show Bush with a better approval rating than most other polls.  So, I not only wonder about their methodologies, but I wonder if they also threw Bush-leaner states into the "Battleground", like Arizona, Nevada, Arkansas, Missouri, Colorado, Louisiana, etc... some of which (like LA and MO) are more than just slight leans.  Those are states Kerry COULD pick off, but if he does, the race has probably been long since over, since it means he'd have probably already wrapped up FL, PA, and OH.

I'm not saying the poll is certainly wrong.  I'm just always skeptical of the Fox/OD polls.  They have notoriously leaned, oddly, the same direction many accuse their broadcasts of leaning.

In case you think I've hit my head...

Fox/OD Bush approval 5/19 = 48%
Newsweek 5/14 = 42%
Time/CNN 5/13 = 46%
Zogby 5/13 = 42%
CBS 5/11 = 44%
CNN/USA Today/Gallup 5/9 = 46%
Pew 5/9 = 44%

7 polls in the last 10 days.  Fox/OD shows the highest approval rating.  Their early April poll was similar, though Newsweek was even and Gallup actually higher.  Early February (not as many March polls to compare)?  Time/CNN was the only one higher, and barely.  Fox/OD always leans pro-Republican... whether intentional or not, I won't debate that... but the fact is, they do.

As I provided in a link to my post above Newsweek in 2000 last poll had Bush up by 2

First, Zogby has a real problem with accuracy.  Of the major polls in 2002, he had far and away the highest margin of error and far and away the highest wrong prediction.

Second, Newsweek has been a bad joke for years.  It consistently slants its polls to the lefft both by sample and other means.

Third, the other results are all within margin of error.

Fourth, Bush took a real barrage of hits from the liberal media which would have you believe that the only news was the Iraqi prisoners weren't treated well and that Bush was somehow responsible for that ill treatment.  The people who understandably were upset with what happened in the prison have since come to realize that the media is overplaying its hand and appreciates that Bush has not been panicked by the media attacks.

[12th Doctor]

This is when most other polls showed Gore as up by 2%-3% margins.

really?  Most polls had Bush up.

http://www.ncpp.org/poll_perform.htm
http://www.pollingreport.com/2000.htm

Are these the final polls?  The ones the day before the election.  Because I remember the race evened out after the DUI announcement and Gore was up in a lot of the polls.

[millwx]

WRONG.

You folks keep quoting ONE poll in which Fox/OD showed no bias, after I've gone through the trouble of posting 17 MONTHS of polling data, in which, of 7 different polls, Fox/OD was the single most Bush-favorable.  Sure, someone's got to be on top, I suppose, but this hardly seems like a coincidence.

I give up.  Folks delude yourself if you want.  Facts are facts.  One poll in 2000 doesn't make Fox/OD the king of polling.  17 months of persistent bias DOES make them questionable.

[millwx]

Here are some actual facts from Congressional/Governor races in 2002

Finally, someone has finally posted something real... rather than one poll.

Thanks Spin Police.  I'll stop comparing Fox/OD to Zogby as the baseline.

This still, however, says nothing about my original point.  Fox/OD is not in that table (presumably, it would fall under "other" polls, IF they conducted any at all... in which case, it MIGHT be even worse, since the "other" average is even worse than Zogby... but I won't go down that line, we have no idea who else is in other).  And Fox/OD has shown 17 months of bias (probably longer, I simply chose to stop my analysis there randomly, because I didn't feel like going back any further).  Zogby's failings don't make Fox/OD good.  But thanks for at least putting out those 2002 stats.

[millwx]

Are these the final polls?

Supersoulty,

In one of my (way too many) posts on this thread, I actually posted which candidate was ahead in every major poll in their final poll before the 2000 election.  Out of 15 polls, 11 had Bush ahead, two showed a tie, and two showed Gore ahead.

[Spin Police]

WRONG.

You folks keep quoting ONE poll in which Fox/OD showed no bias, after I've gone through the trouble of posting 17 MONTHS of polling data, in which, of 7 different polls, Fox/OD was the single most Bush-favorable.  Sure, someone's got to be on top, I suppose, but this hardly seems like a coincidence.

I give up.  Folks delude yourself if you want.  Facts are facts.  One poll in 2000 doesn't make Fox/OD the king of polling.  17 months of persistent bias DOES make them questionable.

lets look at this again... using your own numbers...

CBS = 59%
Gallup = 59%
Fox/OD = 59%
Pew = 58%
Newsweek = 56%
Time/CNN = 56%

Zogby = 54%

We toss out Zogby - he asks a totally differently worded question, so his results are not comparable.  Ask a different question, get a different answer.

This leaves 4 firms showing the same thing:

CBS = 59%
Gallup = 59%
Fox/OD = 59%
Pew = 58%

Hard to call Fox the "fringe" - Agreed?

and at the bottom we have 2 known Bush/GOP hating polls:

Newsweek = 56%
Time/CNN = 56%

Your argument that Fox is at the fringe is disproved by your own data.. IMHO...

[ilikeverin]

First, Zogby has a real problem with accuracy.  Of the major polls in 2002, he had far and away the highest margin of error and far and away the highest wrong prediction.

That's true, I think he predicted that Mondale would defeat Coleman...

I think he can be excused...

     Coleman (R) 1,091,253 50%
     Mondale (DFL) 1,029,982 47%
     Moore (I) 43,892 2%
     Tricomo (G) 9,723 1%
     Kovatchevich (?) 2,176 0%

Note... I is Independence, not Independent