Gallup/NBC - Quickie Anaylsis

[The Vorlon]

Taking a tad deeper look at the new numbers shows far less change that appears on the surface.



6 months out, it is worth looking at both sets of numbers - the Likely Voter and the Registered voter pools.

As we get within a few weeks of the actual vote, the "Likely" number obviously becomes a lot more useful, but with half a year, 2 conventions, summer vacations, the Olympics and (most importantly) The World Series between then and now, both sets are worth a look.

Registered Voters

Among registered voters, Gallup Currently has the race tied at 47/47 in a 2 way race, with Bush holding a 1% lead if Nader is included in a 3 way ballot.

Two weeks ago, Bush had a 4 point (2 way race) and 3 point (3 way race) lead.

One Month ago, Kerry held a 2 point lead (2 way race) and a 1 point (3 way race) lead among registered voters.

What has occured is fairly simple.  Things have gotten very ugly in Iraq.  When ever things get really ugly, there is alway a bit of a reflexive rally to the President effect.  This artificial Bush bump is now fading, and things are returning more to baseline.

Over the one month period things are essentially unchanged.  In a 2 way race, 1 month ago Kerry had a 2 point lead, things are currently tied.  The 2 point Bush "improvement" is well within the margin of the error of the poll.

Likely Voters

Today, Kerry has a 1 point lead (2 way Race) or is tied (3 way race) This represents a change right at the edge of the margin or error of the poll.

Two weeks ago Bush had a 5 point advantage (2 way Race) or a 6 point lead (3 way race) versus leads of

4 weeks ago, Bush held a 3 (or 4) point lead in the race.

The Iraq effect was similar to what occured among registered voters - A "Rally to the President" Bump, then a fading of the bump.  The net change of the 4 week period of 4 points in favor of Kerry is again right at the edge of the polls margin of error.

Signifigance...?

Traditionally, a Republican will do "about" 3 % better in a "likely" voter poll than in a "registered" voter poll. - This ratiio is currently reversed with Kerry holding a 1 point lead among "likely" voters while tied among "registered" voters.

The reason for this is again quite simple.  The news in Iraq is very bad, the prisoner abuse matter cannot make anybody feel good. - Those who dislike Mr. Bush will get "outraged" which is more likely to make them "likely" voters, while those who support Mr. Bush cannot help but be a tad depressed about the situation, which makes the less inclined to be "likely" voters as their enthusiasm has no doubt somewhat diminished.

This is clearly a glass half full/half empty situation regardless of your political perspective.

The total mass of voters (registered voters) has not changed - 1 month ago Kerry had a 2 point lead, today it is tied.

If you support Kerry you are happy because the energy balance has marginally tipped you way - The Bush folks were slightly more "energized" and hence Bush had a marginal lead among "likely" voters - this "energy" advantage has evaporated.

If you are a Bush supported you take comfort in the fact that no minds have actually changed versus a month ago (4 weeks ago Bush trailed by 2 points among registered voters, today he is even) and that in the event of some good news, the Bush voters would likely "re-energize" pretty quickly restoring his lead among the "likely" voters.

WSJ Vs Gallup - Why the 5 point Gap?

Among likely voters Gallup has Kerry +1, WSJ has Bush +4 - why the Gap?

Both Gallup and Teeter/Hart represent the very best there is in terms of polling firms.  These are 2 of the 3 best firms anywhere.

The 5 point gap is "barely" within the margin of error of the two polls so it could just be normal statistical variation.

More likely, however, is that it is a product of slightly different "likely"voter models.

The Teeter/Hart model is more heavily weighted to factoring in past voting behavior.  Hence the typical 3% or so GOP "likely voter" advantage is likely to preserved by the Teeter/Hart "likely voter" screening questions.

The Gallup model more heavily weights in current feelings and enthusiasm at this exact moment in time.

I suspect that the dampening effect of Iraq in the Gallup model versus the Teeter/Hart higher weighting of past voting habits likely explains most of this 5 point gap.

[California Dreamer]

But you know that focussing on the head to head now is only half the story, and maybe less. However another way to look at it is that NBC/WSJ says 49% think Bush "Doesnt deserve reelection", with 44% sayign he does.

All the polls show that Bush's approval and right track numbers are weak. His 'trump card' rating on terrorism is also dropping to his lowest levels. And for the first time more americans think Iraq was a mistake.

If Kerry can win over the those who are starting to have doubts about Bush then he can pull ahead....but we wont know that until the end of May after his first month of media spending.

[pieman]

Taking a tad deeper look at the new numbers shows far less change that appears on the surface.



6 months out, it is worth looking at both sets of numbers - the Likely Voter and the Registered voter pools.

As we get within a few weeks of the actual vote, the "Likely" number obviously becomes a lot more useful, but with half a year, 2 conventions, summer vacations, the Olympics and (most importantly) The World Series between then and now, both sets are worth a look.

Registered Voters

Among registered voters, Gallup Currently has the race tied at 47/47 in a 2 way race, with Bush holding a 1% lead if Nader is included in a 3 way ballot.

Two weeks ago, Bush had a 4 point (2 way race) and 3 point (3 way race) lead.

One Month ago, Kerry held a 2 point lead (2 way race) and a 1 point (3 way race) lead among registered voters.

What has occured is fairly simple.  Things have gotten very ugly in Iraq.  When ever things get really ugly, there is alway a bit of a reflexive rally to the President effect.  This artificial Bush bump is now fading, and things are returning more to baseline.

Over the one month period things are essentially unchanged.  In a 2 way race, 1 month ago Kerry had a 2 point lead, things are currently tied.  The 2 point Bush "improvement" is well within the margin of the error of the poll.

Likely Voters

Today, Kerry has a 1 point lead (2 way Race) or is tied (3 way race) This represents a change right at the edge of the margin or error of the poll.

Two weeks ago Bush had a 5 point advantage (2 way Race) or a 6 point lead (3 way race) versus leads of

4 weeks ago, Bush held a 3 (or 4) point lead in the race.

The Iraq effect was similar to what occured among registered voters - A "Rally to the President" Bump, then a fading of the bump.  The net change of the 4 week period of 4 points in favor of Kerry is again right at the edge of the polls margin of error.

Signifigance...?

Traditionally, a Republican will do "about" 3 % better in a "likely" voter poll than in a "registered" voter poll. - This ratiio is currently reversed with Kerry holding a 1 point lead among "likely" voters while tied among "registered" voters.

The reason for this is again quite simple.  The news in Iraq is very bad, the prisoner abuse matter cannot make anybody feel good. - Those who dislike Mr. Bush will get "outraged" which is more likely to make them "likely" voters, while those who support Mr. Bush cannot help but be a tad depressed about the situation, which makes the less inclined to be "likely" voters as their enthusiasm has no doubt somewhat diminished.

This is clearly a glass half full/half empty situation regardless of your political perspective.

The total mass of voters (registered voters) has not changed - 1 month ago Kerry had a 2 point lead, today it is tied.

If you support Kerry you are happy because the energy balance has marginally tipped you way - The Bush folks were slightly more "energized" and hence Bush had a marginal lead among "likely" voters - this "energy" advantage has evaporated.

If you are a Bush supported you take comfort in the fact that no minds have actually changed versus a month ago (4 weeks ago Bush trailed by 2 points among registered voters, today he is even) and that in the event of some good news, the Bush voters would likely "re-energize" pretty quickly restoring his lead among the "likely" voters.

WSJ Vs Gallup - Why the 5 point Gap?

Among likely voters Gallup has Kerry +1, WSJ has Bush +4 - why the Gap?

Both Gallup and Teeter/Hart represent the very best there is in terms of polling firms.  These are 2 of the 3 best firms anywhere.

The 5 point gap is "barely" within the margin of error of the two polls so it could just be normal statistical variation.

More likely, however, is that it is a product of slightly different "likely"voter models.

The Teeter/Hart model is more heavily weighted to factoring in past voting behavior.  Hence the typical 3% or so GOP "likely voter" advantage is likely to preserved by the Teeter/Hart "likely voter" screening questions.

The Gallup model more heavily weights in current feelings and enthusiasm at this exact moment in time.

I suspect that the dampening effect of Iraq in the Gallup model versus the Teeter/Hart higher weighting of past voting habits likely explains most of this 5 point gap.

Excellent analysis Vorlon. Your insight is greatly appreciated.

As far as campaign strategy at this point, it would seem that maintaining the base is the most important since IND don't decide until Oct/Nov anyway. What would you look for or what do you see so far with regard to each candidate maintaining their base?
I know you mentioned married woman before as an indicator, anything else?

[The Vorlon]

But you know that focussing on the head to head now is only half the story, and maybe less. However another way to look at it is that NBC/WSJ says 49% think Bush "Doesnt deserve reelection", with 44% sayign he does.

All the polls show that Bush's approval and right track numbers are weak. His 'trump card' rating on terrorism is also dropping to his lowest levels. And for the first time more americans think Iraq was a mistake.

If Kerry can win over the those who are starting to have doubts about Bush then he can pull ahead....but we wont know that until the end of May after his first month of media spending.

Regarding "Kerry's first month of media spending" - This is a substantial misrepresentation of reality.

Between them Kerry/Dean/Edwards/Clark, etc spend $100+ million on the primaries.  Even if we say only 50% of that was "Bash Bush" money (likely a low estimate) thats $50 million.. (agreed?)

Since Super Tuesday Kerry has spent $22 million and friendly to Kerry 527s have spend $31 million = an additional $53 million in Bash Bush money.

Net media spending on both sides has been pretty close actually, with a modest advantage in spending to the Dems if you are objective about it...

Feel free to add up Dean/Kerry/Edwards/Clark + 527 spending versus what Bush has spent and draw you own conclusions...

Regarding Terrorism...

The majority of folks think Iraq was a mistake, but even though they think Bush made a big mistake they still trust Bush more then Kerry on this issue..

I am unsure which candidate is the most damned by this result.

Bush is a screwup - but Kerry is viewed as an even bigger screwup...?   Hard to parse a political advantage on that one for either side.

regarding you "Deserves to be re-elected" numbers...

Yes and no...

At the end of the day, if you vote, you have to vote for somebody.

If you think Bush is an idiot who does not deserve re-election, but the Kerry is an even worse idiot, you hold you nose and vote Bush. - What matters is the approval rating "gap" between the candidates.  Do people dislike Bush more than Kerry or the other way around.

Both sides are deeply negative in their campaigning.  The media ads are negative but the stuff on the ground and person to person is even more so negative.

 This not only drives down you opponents positives, but also drives up your own negatives.

Much of the "decline" in Bush's rating is the result of the negative attacks Kerry and all his 527 friends are running, as well as Bush's own negative adds against Kerry.

Both candidates have numbers that are just brutal.

If they were not running against each other, they would both be un-electable.

[The Vorlon]

What would you look for or what do you see so far with regard to each candidate maintaining their base?
I know you mentioned married woman before as an indicator, anything else?

The number of truly "swing" voters left in American politics is "about" 8%;

Married White Women
Younger Slightly poorer Men (PMCs as they are sometimes called)
Hispanics.

For a very quick and dirty indicator, look at the difference between the "registered" voter result and the "Likely" voter result. -

Historically, the number is "about" a 3% shift to the GOP - ie in the likely voter poll the GOP does "about" 3% better than in the "registered" result.

Right now Kerry does 1 point better in "likely" versus "registered" -Advantage to Kerry, for now, watch this gap

James Carville's poll (Democracy Corps - an excellent poll BTW, Actually done by Greenburg Quinlan Rosner) asks a set of "temperature" questions to measure party enthusiasm, and while not directly useful is a nice "generic" guage.

[ilikeverin]

the rvolon si teh pollnig/polling ana7ysis master!!!!!!111~~~~~~~

[Beet]

the rvolon si teh pollnig/polling ana7ysis master!!!!!!111~~~~~~~

lol

[Lunar]

Shouldn't the likely voter models be weeding out that 5-6% Nader support?

Democracy Corps has this analysis:

Code:

CONSUMER CONFIDENCE BY PARTISANSHIP
   Consumer Comfort Index   Change Since 01/04
Total          -17   -14
Republicans   +35   +6
Independents  -36   -27
Democrats      -36   -6

The true interesting number is hte independent switch.

[classical liberal]

If you think Bush is an idiot who does not deserve re-election, but the Kerry is an even worse idiot, you hold you nose and vote Bush.

Or you stay home.  I'm predicting that after the next bout of fiscal irresponsibility on the part of the supposedly conservative Congressional Leadership and President, enough of Bush's coure will be so disillusioned with him that though they will answer that "if the election were held today" they would vote for Bush over Kerry or Nader, they will in actuality stay home or vote Libertarian nationally (president, congress) and vote regularly for state and local offices.  I'm sure that the act of observing the phenomenon changes the phenomenon, in this case towards a stronger support for a candidate in response to bing pressed for a preference.  I know that until I read Peroutka's spiel, I was one of those people who would answer Bush in a poll but would probably stay home on 11/2/04

[agcatter]

...and if kerry appoints three more libs to the Supreme Court, what the hell.  Right?

[Beet]

...and if kerry appoints three more libs to the Supreme Court, what the hell.  Right?

You think the Dems'll do better than break even in the Senate? I hope you're right, but I sincerely doubt it.