Somebody is wrong.. :) ABC says Kerry Up, TIPP says Bush Up - Take your pick!

[The Vorlon]

ABC News: June 17th - 20th (Registered Voters)
Sample size = 1200

Kerry 48
Bush 44
Nader 6

http://www.washingtonpost.com/wp-srv/politics/polls/trend_062104.html

TIPP - Registered Voters (June 13th-19th)
Sample Size = 1000

Bush 44
Kerry 41
Nader 6

http://www.investors.com/editorial/issues.asp?v=6/15

Harris Says Bush +10 (unlikely)
Pew says Bush +4
IBD says Bush +3
Rasmussen says Kerry +3
ABC says Kerry +4/+8

Ok.. that's clear... I understand now...

[Wakie]

The important poll is a few months away.

[millwx]

ABC News: June 17th - 20th (Registered Voters)
Sample size = 1200

Kerry 48
Bush 44
Nader 6

TIPP - Registered Voters (June 13th-19th)
Sample Size = 1000

Bush 44
Kerry 41
Nader 6

The MOEs on these two polls are about 3%.  So, I'd just call it a tie.

Also, though I know there's been great debate about this... several polls have shown Kerry getting a much greater percentage of undecideds.  Note that the TIPP poll has far more undecideds (9% left over uncommited compared to only 2% uncommited in the ABC/WP poll).  Obviously, Kerry wouldn't get all the undecideds... not even close.  So, there's still some discrepancy here.  But I don't think they're as contradictory as they look on face value.  Like I said, I'd just call them even.

[The Vorlon]

Gallup should be closer to ABC than TIPP based on methodological similarities.

We'll see tomorrow

[khirkhib]

ABC Polls have generally been very favorable to Bush compared to other polls put out at the same time.  This could be another signal of a major change in the polls.  If Bush and Kerry would tied the whole time during the polls this election wouldn't be interesting, Looks like we have a really fight on our hands.

[millwx]

Nice graphic...


Consider the fact that Reagan's death occurred a bit of the way into the AP/IPSOS poll.  And these polls are charted in chronological order, with the most recent one on top.

Makes me wonder if the Harris poll was right... that Bush's Reagan bounce was greater than expected, but extremely brief.  The rise up and drop back down in this chart is quite smooth and "normal".  Excluding the LA Times poll, three of the four pre-Reagan-death polls show Kerry up by a few percent.  Then the AP/IPSOS poll draw them to a near tie.  Fox/OD (right after the death) holds it there.  Pew brings Bush up by a few.  Harris widens the Bush lead to 10%.  TIPP trims it back to a few percent.  And then ABC/WP puts Kerry back up by a few.  Not coincidental, in my opinion, the Rasmussen daily tracking - though failing to show the Bush lead to the degree Harris does - has the same timing of the swing, putting Kerry back up by 3% today (though, admittedly, that is only one day in the daily track).

I'd bet The Vorlon is right... Gallup will be similar to ABC/WP, but not just because of methodologies... but because this was a real "ebb and flow".

[khirkhib]

Excellent analysis Millwx.  I think you nailed the phenomonon dead on.

[The Vorlon]

Alternative #1

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees TIPP/Harris/Pew

=> ABC is a bad sample

Alternative #2

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees with ABC

=>Brief Reagan bounce, Kerry on Top

[ATFFL]

Alternative #3

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup - dead heat

Its a toss up.

[millwx]

Alternative #1

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees TIPP/Harris/Pew

=> ABC is a bad sample

Alternative #2

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees with ABC

=>Brief Reagan bounce, Kerry on Top

Agreed.  It'll be interesting to see Gallup!  I'd wager it shows a small Kerry lead.  But who knows?  Time will tell.  We'll see.  And the ABC/WP poll is no help on its own.  It fits the up/down trend reasonably and its sample doesn't seem terrible.  But, it is still a fairly good shift from TIPP and its sample doesn't seem "great".

[khirkhib]

The regional Analysis from the ABC Poll.   It seems as though Kerry and Bush are actually beginning to pull away from each other in their perspective regions.  Kerry far better in the East and improving in the Midwest and West and Bush still improving in the south.


  ...                     East   Mwest   South  West     All
Bush                  31%   44%      54%    42%      45%  
Kerry                  66%   53%      44%    56%      53%  
Other                 0%      0%         0%      0%        0%
Neither              1%       2%         1%     1%         1%  
Would not vote 1%       1%         1%      1%        1%  
DK/No opinion   1%       1%         0%      0%         0%

[The Vorlon]

Alternative #1

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees TIPP/Harris/Pew

=> ABC is a bad sample

Alternative #2

TIPP Bush +3
Harris Bush +6 (registered)
Pew Bush +4
ABC Kerry +4/+8
Gallup +> More or less agrees with ABC

=>Brief Reagan bounce, Kerry on Top

Agreed.  It'll be interesting to see Gallup!  I'd wager it shows a small Kerry lead.  But who knows?  Time will tell.  We'll see.  And the ABC/WP poll is no help on its own.  It fits the up/down trend reasonably and its sample doesn't seem terrible.  But, it is still a fairly good shift from TIPP and its sample doesn't seem "great".

8% to the dem side is a bit off in the ABC sample, but that does not explain it all by any means.

TIPP and ABC collect (as oppose to anaylse0 data almost identically

The MOEs intersect at +/- a tie...

If Gallups says tied, I'll take "door #3" please

We will see tomorrow

[ilikeverin]

I strongly object to you having MN be the most swingy state in the Midwest, Vorlon

[John]

i pick the poll with
bush: 44%
kerry: 40%

[muon2]

Great graph to post. The scientist in me likes to look at trends and this one is loaded. The most interesting trend is that before 9/11 the trend was flat even during real recession months. But since and including 9/11, there have been three external events that caused rapid upward spikes. Except for those the trend has been down and with a uniform slope. That's surprising since normally I would expect a sharper fall of initially followed by a leveling of the trend line.

Naive analysis on my part would be that there is a continual "tiring" of the public with Bush that is perked by these unusal events (9/11, the Iraq war, Saddam's capture). Without evidence of a leveling off of the trend, Bush would need one or more singular events to provide the necessary improvement in ratings before November.

ABC Polls have generally been very favorable to Bush compared to other polls put out at the same time.  This could be another signal of a major change in the polls.  If Bush and Kerry would tied the whole time during the polls this election wouldn't be interesting, Looks like we have a really fight on our hands.

[Reds4]

The ABC poll's approval of Bush's handling of War on Terrorism is very strange. It is down to 50% approval. Also, it has Kerry up, albeit barely, on who you would rather see handle the war on terror. This is very strange as other polls have still shown Bush preferred in handling terrorism by 8 to 10 points at least. This abc poll shows Bush gaining on the economy and losing on terrorism. Who would have thought this a few months ago? Anyone with any conclusions on why Kerry would be gaining on handling war on terror (not partisan rants, just objective opinions please).

[Smash255]

The ABC poll's approval of Bush's handling of War on Terrorism is very strange. It is down to 50% approval. Also, it has Kerry up, albeit barely, on who you would rather see handle the war on terror. This is very strange as other polls have still shown Bush preferred in handling terrorism by 8 to 10 points at least. This abc poll shows Bush gaining on the economy and losing on terrorism. Who would have thought this a few months ago? Anyone with any conclusions on why Kerry would be gaining on handling war on terror (not partisan rants, just objective opinions please).

Could the 9/11 comission report statig that their is no evidence to connect Saddam to 9/11 impacted Bush's terrorism handling in the ABC poll?

[Smash255]

Options on why their is a difference in the 2 polls

One is that undecided voters tend to go for the challenger in incumbent elections.  ABC poll pushes the undecidedes more than the TIPP poll, so the undecidedes in this election are stayinh with tradition and going towards Kerry.  That will be shown in the ABC poll, but not in the TIPP poll.

Another explanation could be that Bush was doing better last week and Kerry is doing better in the last couple days.  The TIPP poll started a few days before the ABC poll & the ABC poll ended a few days later.  If Bush was doing better last week than he is doing now that could explain why their is a difference in the two polls.  Also the Harris poll & the recent Rasmussen poll could at least back that up a bit, the older Harris poll which was partially taken last week was in Bush's favor & the Rasmussen poll whichshiows Kerry up 3, was taken within the last 3 nights.

The ABC pol does show a slight oversmapling of the Dems, but not nearly enough to have an impact.  The ABC poll does show Dems with a +8 adv in those in the poll, the 2000 results showed a Dem +5 adv  (39-34), so the oversampling is really only +3 which isn't enough to impact the polls other that 1 or 2 points, and without knowing the Party Breakout  of the TIPP poll its hard to say how accurate their sampling was.

[Françoise]

The details of the ABC/WaPo poll are published on the Washington Post website.  It's fairly useful, allowing one to view a specific demographic group's response to each of the questions.

The trend is availabe, as posted earlier.

Françoise