Bush gains 4% in Newest Gallup (registered) Loses 3% (Likely) Approval up 1%

[Spin Police]

Gallup Poll

May 21-23 (Registered Voters - 2 way race)

Kerry 48
Bush 46

May 7-9 (Registered Voters - 2 way race)

Kerry 50
Bush 44

Likely Voters

May 21-23

Kerry 49
Bush 47

May 7-9

Bush 48
Kerry 47

http://www.usatoday.com/news/politicselections/nation/polls/usatodaypolls.htm

[© tweed]

Likely voters is all that matters.  And don't say I'm just saying that because Kerry is doing best in the likely, I said that on the last Gallup when LV was Kerry's worst group.

[California Dreamer]

nice way to 'spin' the fact that Bush has gone from +1 with likely voters to -2

So with likely voters he has gone from winner to loser.


you sould arrest yourself

[© tweed]

Approval went to 47%

[Spin Police]

nice way to 'spin' the fact that Bush has gone from +1 with likely voters to -2

So with likely voters he has gone from winner to loser.


you sould arrest yourself

No spin at all.  

A statement of BOTH facts, up 4 in registered, down 3 in likely.

Where is the bias in providing both relevant facts?

[California Dreamer]

No spin at all.  

A statement of BOTH facts, up 4 in registered, down 3 in likely.

Where is the bias in providing both relevant facts?

you changed your headline...the original headline only included the registered voter gain and the approval gain.

[© tweed]

According to  today's CBS approval poll his rating has dropped from 44% to 41% in the last two weeks.

CBS leans left, Vorlon already explained that

[Giant Saguaro]

Good, the approval has hopefully bottomed out already. 49-47 and with the "modest," as Vorlon says, undersampling of Republicans and with Republicans tuning out to the election during the summer months, Bush is hanging in there.

Now this is the CNN/USA Today/Gallup. Wonder if this is different from the actual Gallup poll that they post on their page? Bush sometimes does slightly better in that one than the CNN/USA Today/Gallup tracking poll.

[Spin Police]

No spin at all.  

A statement of BOTH facts, up 4 in registered, down 3 in likely.

Where is the bias in providing both relevant facts?

you changed your headline...the original headline only included the registered voter gain and the approval gain.

Wrong.  You are lying or confused (A lot like your candidate)

Your post retains EXACTLY what the current heading said at the time you posted your reply.  

It is still on the top of your post, and every other post except the top of the thread.

That is why evry single post (except the lead post) says EXACTLY what the original said. - ie Both the registered and likely result.

The "Approval up 1%" was added.

Your apology is accepted.

[??????????]

No spin at all.  

A statement of BOTH facts, up 4 in registered, down 3 in likely.

Where is the bias in providing both relevant facts?

you changed your headline...the original headline only included the registered voter gain and the approval gain.

Wrong.  You are lying or confused (A lot like your candidate)

Your post retains EXACTLY what the current heading said at the time you posted your reply.  

It is still on the top of your post, and every other post except the top of the thread.

That is why evry single post (except the lead post) says EXACTLY what the original said. - ie Both the registered and likely result.

The "Approval up 1%" was added.

Your apology is accepted.

Spin don't mind him. Sometimes he says some really great things but a lot of the time his name describes him to a "T". "California Dreaming".

[Smash255]

According to  today's CBS approval poll his rating has dropped from 44% to 41% in the last two weeks.

CBS leans left, Vorlon already explained that

With all of how bias CBS has, they had the 2000 election almost right on.  They had Gore winning the popular vote by 1% he won by .5%.  With all their so called biases they were pretty damn close.

[Reds4]

I don't understand why you guys think likely voters is the only one that matters in May. Likely voters could easily change by November. Granted Likely voters is all that matters in November, but in May I believe registered voters definitely matters as well.

[Smash255]

I don't understand why you guys think likely voters is the only one that matters in May. Likely voters could easily change by November. Granted Likely voters is all that matters in November, but in May I believe registered voters definitely matters as well.

In the latest Gallup Poll Kerry leads by 2 in both anyway.  The last time their was a 5 point difference in registered and likely, and I think that was an abstraction.  The registered and likley voters may see some difference, but I doubt that is more than 1 or 2%.  Bush going up in registered and down in likley was probably just the poll correcting itself out because of the difference the poll had last time

[millwx]

No spin at all.  

I'm non-partisan, but most of my posts to date have been anti-Bush, as just a matter of chance.  In fact, I think SpinPolice and I have gone back and forth a couple of times before already.  But, this time, I'll defend him...

Not only is it good of him to post both sets of numbers, but many believe that this early in the game registered voters are as good, if not better, than likely voters.  I'm not sure I agree, given the screening methods, but I'm hardly willing to argue.  Like resident poll expert Vorlon, I also have a few connections inside some of the polling organizations.  And by "many" I don't mean Tom, Dick and Harry... I mean some top dogs within the polling firms believe this.  I'd have to go back and review my notes as to which ones.  It may be only those who screen "likely" by voter intent.  THAT would make a sensible argument  (rather than screening "likely" by past voting habits) for favoring "registered" over "likely" right now.  But I'm not sure.

Anyway, I digress.  Point is.  Both sets of numbers are equally valid at this early stage (depending on screening methodology).  There's no "spin" in SpinPolice including both.

[The Vorlon]

Only 55% or so, of the population actually votes so "obviously" the only poll worth anything is a "likely voter" poll...

Unfortunately, like many "obvious" things, when you try to translate what is pretty obvious on paper into something that works in the real world things get a tad complicated.

here is a neat graphic that I hope provides a little bit of ight on the matter.



For turnout purposes, you can break voters into 4 basic groups.  Clearly this in not perfect, and there are certainly grey areas, but it generally works pretty well.

Firstly, there are the "partisans" - people strongly associated with a party or a cause.  They make up "about" 44% of the population, and of the 44% about 34% out of that 44% actually vote.

These people show up in EVERY poll, registered, likely, super likely, hotdog, whatever.. from a polling perspective, these guys are NOT the problem.

The next group of voters is the "Good Citizens" these people are weakly, if at all, associated with a party or cause.  They vote becaue they are, well, simply "Good Citizens" and vote beacuse it is the right thing to do.
They make up about 10% of the population, and 80% of them (ie 8% of the 54% or so who actually vote)

A Registered voter poll will catch them, but a likely voter poll, depending on how it is structured, may or may not.

Because they very regularly vote, a "likely voter" poll that questions about past voting behavior will include them, but a poll which asks about "Are you paying attention" or "How much have you thought about the election" etc will chronically underrepresent this block.

The next group is the Intermittent voters.  These folks vote 40ish % of the time.  

23 weeks from an election a registered voter poll is wrong because it includes all of these voters, but a Likely voter poll is also wrong because it excludes most of these people.

The last group, the Unlikely, rarely vote but about 1/3rd are actually registered.  

A registered voter poll that pushes the leaners will produced gibberish because of these people.  If candidate X had a good news cycle, X will "surge" in this demographic, and the next poll if Y had a good cycle Y will surge back showing huge changes among people who will not vote anyway.

Right now Registered versus Likely voter polls can show us a couple things.

This far out, I really like the Registered Voterpolls that, effectively, have a very mild screen built in.  At this point, polls that show a very high "undecided" are actually good polls to watch simple beacause not pushing the voters thus losing the "unlikely" and leaving behind the 75% or so that makes up the universe of "possible" voters. (Obviously, as we get close to E-day  - Likely is the way to go)

The other way to do it this far out is to actually do a real, proper, legitimate "likely voter" screening.

Perhaps I am showing my alledged Democratic bias here, but Democracy Corps (D) and Zogby (D) both do a real good job here (As does Gallup and WSJ)

Turnout is so important the Zogby, who has an astonishingly long list of methodological transgressions in his polls, actually takes the time (and hence the money) to sort this part of his polling out right.

As an aside, a lot of people have been taking about which way the undecided break in an election.  This is a false question.  Most of the "undecided" have actually made up their mind who they would vote for if they actually voted - the real "break" in the last few days is in deciding if they will actually vote or not...

One very important and useful thing to watch is also the difference between the Registered voter and Likely voter polls.  

For example in the current Gallup Kerry is +2 in both the registered and Likely polls, which indicates roughly equal energy levels between the two parties.

If one party does better in the Likely part of the poll it reflects that they are feeling better than the other guys about things and this higher energy is thus something to track fairly closely.