Back from a week out, it doesn't look like I missed much by not having email or internet for a few days; from what I saw of the debate in a taco shop nearby the Long Beach airport yesterday, it sounded pretty much just like the first debate.

In the jump, I'll let you know where I think the race stands at nationally, and how I got there.

After doing the calculations, here's where I think the race stands at, for predicting the Nov 4th outcome:

We have a plethora of polls to pick and choose from nationally, and if you want, you can look at Pollster or RCP or 538 or Chris Bowers or others to see how it might look differently. But given we do have a bunch of polls, I've decided to look at the tracking polls of likely voters only. That means, for now, to ignore Gallup's tracking of registered voters-- when they move to an LV polling I'll include them. I also re-weight the R2K/DailyKos poll numbers according to a more likely turnout model. By my calucations, that moves it +/- 3% for McCain/Obama. Others might want to take them at their face value, but that's my take. Now that Zogby has started its telephone tracking, its a good poll; likewise with Hotline switching to a LV model. Battleground remains the gold standard-- I expect that they, along with Rasmussen, will adjust to nail the final outcome.

I have three measurements of the national outcome above (Low, Med, High), which I'll describe below:

Low-- This is basically the Peter Feld model that I blogged about a week ago. It's a conservative estimate which speculates that if Obama is not breaking out above 50 percent, there is a higher base of McCain than what is polled, which he pegs at 46 percent, and that undecideds break 2:1 in McCain's favor.

Med-- This model takes the McCain average at face-value, but still breaks the undecideds toward McCain by a 2:1 advantage.

High-- This takes the polling averages for both candidates at face-value, and breaks the undecideds 50-50.

I expect the 'other' vote to be very low, about the same as 2004. The range I currently have is of Obama winning nationally by a .35 to 3.4 margin. I'll try and take a look at the states by the end of the week, but 5 days off the net while in disneyland and LA beaches has me behind quite a bit of work.

Of course, this means nothing as its the states that matter in the EV count. Right now, the MyDD EV counter looks solid for Obama, with wins in the toss-up states of FL, VA, OH, CO, and NV. Those are the five states that McCain would have to flip to change the outcome of the election.

Tags: 2008 (all tags)

Welcome back!

What do you make of MO and NC?  RCP has a 0.3 lead in both...

It's my feeling that MO will eventually go to McCain, but NC is a true toss-up in my eyes.

It's important to reiterate that is McCain loses a SINGLE one of the states we are talking about, he loses the entire election.  Unless there is a drastic change of events in the next 4 weeks (or even 2 weeks due to early voting) Obama's floor is 264 EVs.

by neko608 2008-10-08 07:17PM | 0 recs

I was just thinking of the 269-269 tie...where McCain wins every state we are talking about except NV (which is entirely possible).

The senate would vote for VP...Lieberman would vote for Palin (of course) in which case Cheney would cast the tie-breaker for Palin, giving here the vice-presidency....but what about Chuck Hagel?  If he votes for Biden (we've seen his comments about her), what kind of firestorm would that start???

by neko608 2008-10-08 07:21PM | 0 recs
##### Blanche Lincoln would be the key

Would she be under pressure from her constituency to vote for Palin as VP since she's up for re-election?  McCain will likely win 55% in Arkansas (though Hillary campaigning there on Friday is curious to say the least); Huckabee and his cabal will push for a bipartisan administration and he would probably beat Lincoln if he threatened to run against her if she didn't vote for Palin as veep.

We've probably got Hagel as a vote for Biden; how would Lugar vote?

by Blazers Edge 2008-10-08 07:34PM | 0 recs
##### Re: Blanche Lincoln would be the key

Um NO... Way No... REALLY REALLY WRONG!!!

You guys REALLY need to learn your Electoral processes... You are overlooking a very key thing here...

THE NEW SENATE WILL BE IN SESSION!!!!

Cheney is ONLY a tie breaker in a 50-50 scenario... The Senate will be a minimum 5 seat Dem advantage.

by yitbos96bb 2008-10-08 07:38PM | 0 recs
##### Open constitutional question

I think up to interpretation.

by Blazers Edge 2008-10-08 07:40PM | 0 recs
##### Re: Open constitutional question

Um not really... The incoming Senate votes.  That's it.   There is no constitutional question, its simply FACT.  The incoming House votes for President and the incoming Senate votes for the VP.

Nothing with the Electoral College is official until Congress Certifies it.   This takes place January 6 at 1pm, as prescribed by Federal Law.

http://www.law.cornell.edu/uscode/html/u scode03/usc_sec_03_00000015----000-.html

The 111th Congress is sworn in on January third.

THEREFORE... THE NEW SENATE IS IN SESSION...

Meaning that unless things go really bad, in the case of a tie, Joe Biden is the next VP of the US.

by yitbos96bb 2008-10-08 07:59PM | 0 recs
##### huzzah n/t

by xeju 2008-10-08 11:48PM | 0 recs
##### Re: Blanche Lincoln would be the key
by yitbos96bb 2008-10-08 07:46PM | 0 recs
##### Would Murkowski turn on Palin?

and get her "revenge" by voting for Biden?  Specter could face some pressure to vote for Biden as well if Chris Matthews can put the heat on him.

by Blazers Edge 2008-10-08 07:43PM | 0 recs
##### Re: Would Murkowski turn on Palin?

No because the Dems are poised for a 5-8 seat Gain...

by yitbos96bb 2008-10-08 08:00PM | 0 recs

Its a NEW Senate.  The electoral votes are not read until after Jan 3.

by gavoter 2008-10-08 09:16PM | 0 recs

Being a Missourian, I have to say that the current odds are ever so slight in McCain's way. However, this is all about turn out. The vast majority of voters in MO are in 4 areas.  St. Louis and Kansas City are the largest and then Springfield and the Jefferson City area.  If the turnout in those areas see significant increase, then Obama wins.  St Louis City and St Louis County are likely going to go Obama as it has trended Democratic fairly well recently.  KC is also an area that Obama can do well in.  Springfield is very republican.  The Jeff. City area has a lot of college students and government workers...that can be close. If we call those areas the urbam areas, then Obama has a pretty good edge.  The rest of the state will go republican, but they simply lack the numbers to overcome high urban turnout.  So what can you all do to help us get the vote out in the urbam areas?  All we need is about 8% better than average turnout in those areas.

by tominstl 2008-10-08 07:40PM | 0 recs
##### Do you know why Plouffe excluded

MO, IN, and NV from his list of states that he highlighted as red states in which they were making great progress?  He may have been playing chicken but the video is available at Ambinder's website.  He mentioned every other state (including NC) but Ambinder noted the omission of IN and MO.

by Blazers Edge 2008-10-08 07:46PM | 0 recs

I tend to agree with your conservative assessment, but still, not bad news by any measure.  Who'd have thunk it?  Are there any game-changers left?  That's the question.

by Shaun Appleby 2008-10-08 07:18PM | 0 recs

game changers:

an attack of some kind
major scandal on either side
major illness on either side
the unspeakable on either side
Um...

yeah thats the list... all in all I would state the odds of a major game changer at about 5%

by JDF 2008-10-09 12:07AM | 0 recs

meant that to be .5%

by JDF 2008-10-09 12:08AM | 0 recs

Not to mention the sudden discovery of a NEO on a collision course with our planet, now that would be interesting.

by Shaun Appleby 2008-10-09 12:46AM | 0 recs

I agree with the assessment also, although I'm sure the majority of comments will assail it as overly pessimistic.

Many ways to tinker with this, but 2-4 points is always what my Excel model spits out, generally in the 3 range.

I'm unimpressed with high end margins like Obama's current national lead, similar to dismissing McCain's theoretical advantage of a few weeks ago. I like to default to a national situational terrain which amends the natural tendencies of individual states. I can't get to Obama winning by 5+ without slanting many demographics beyond any tendency they've demonstrated since '96, specifically white women voting our way instead of narrowly for the GOP.

by Gary Kilbride 2008-10-08 07:26PM | 0 recs

Then again, historically the close divide from the last two cycles is atypical.  Usually, periods like that end and the pendulum swings one way or the other.

by HSTruman 2008-10-08 07:31PM | 0 recs

You COULD be right Jerome, but I sure hope you'll admit you were wrong if Obama's folks are right in their turnout estimates.  I'm all for being conservative now, since nothing counts till it counts, but your predictions strike me as pessimistic.

by HSTruman 2008-10-08 07:29PM | 0 recs

Interesting article, but it is pretty arrogant to claim that someone else's weighting of Party ID is less correct than yours.  Either use THEIR numbers or don't use them at all... otherwise, by reweighting is pretty much just taking an educated guess, one that could easily be wrong (or right to be fair).  Either way, it comes off pretty arrogant.

by yitbos96bb 2008-10-08 07:36PM | 0 recs

Honestly, if it was possible to re-weight all the polls with the same LV model I'd say it makes an awful lot of sense to re-weight - that way you're comparing apples to apples. And I wouldn't know what to do if Jerome was optimistic, anyway.

by Exhausted Pennsylvanian 2008-10-08 08:52PM | 0 recs

Except that assumes that it's possible to determine one "correct" way to weight.  What if the assumptions used to determine that weight were off?  Then every one of your polls will be off, abd the erroneous assumption magnified.  The point of averaging polls is that those errors (hopefully) cancel each other out.

by sneakers563 2008-10-09 06:01AM | 0 recs

I think the vote is taken by the new House/Senate.

by InigoMontoya 2008-10-08 07:37PM | 0 recs
##### That's an open constitutional

question I believe.

by Blazers Edge 2008-10-08 07:38PM | 0 recs
##### Re: That's an open constitutional

Hmm.

I just read the 12th Amendment.  It says the House shall vote immediately, if necessary, after the Electoral votes are counted.

But both House and Senate have a 2/3 quorum requirement and language suggests that the issue could drag out into the new Congress.

So the House could proclaim Obama...and then all sorts of hinky tricks could occur in the Senate with the Democrats foiling the 2/3 quorum requirement if necessary.

by InigoMontoya 2008-10-08 07:57PM | 0 recs
##### Where's Steve M. to provide his insight

into this constitutional interpretation question?  Don't have the pocket constitution in front of me but "counted" is distinguishable from "certified" so the House could be forced to vote right after all the votes have been "counted" whatever that means.

Good idea on what the Senate dems would do, though Palin could try some shenanigans through the SC.  Political question doctrine?

by Blazers Edge 2008-10-08 08:03PM | 0 recs
##### Re: Where's Steve M. to provide his insight

Steve M doesn't know WTH he is talking about.   Just because one guy on a blog says it, doesn't make it so...   Instead, READ FEDERAL LAW...

http://www.law.cornell.edu/uscode/html/u scode03/usc_sec_03_00000015----000-.html

And then READ every other opinion on it, that pretty much tells you HOW IT IS DONE.

by yitbos96bb 2008-10-08 08:06PM | 0 recs
##### Why don't you just calm the hell down

by Blazers Edge 2008-10-08 08:08PM | 0 recs
##### Re: Why don't you just calm the hell down

Because you've spouted absolute IGNORANT STUPIDITY multiple times on the board despite people telling you and PROVING that you were wrong... yet you just kept on posting it OVER AND OVER (much like a Fox News Cable host keeps spouting ridiculous crap that everyone knows is incorrect but them in an insane attempt that the 20th time it might become true) which frankly tends to piss me off... especially after you were shown to be wrong.  Had you GOOGLED it before you post it repeatedly and you would have seen you were wrong and not posted that stupid crap.

by yitbos96bb 2008-10-08 08:13PM | 0 recs
##### Re: Where's Steve M. to provide his insight

Hey, be nice.  As Arlen Specter once shouted at Russ Feingold, "You are no more a protector of the Constitution than am I."

by Steve M 2008-10-08 08:21PM | 0 recs
##### Re: Where's Steve M. to provide his insight

Our friend yitbos is exactly right.  The report of the Senate Judiciary Committee on the 20th Amendment is instructive:

If it should happen that in the general election in November in presidential years no candidate for President had received a majority of all the electoral votes, the election of a President would then be thrown into the House of Representatives and the memberships of the House of Representatives called upon to elect a President would be the old Congress and not the new one just elected by the people. It might easily happen that the Members of the House of Representative, upon whom devolved the solemn duty of electing a Chief Magistrate for 4 years, had themselves been repudiated at the election that had just occurred, and the country would be confronted with the fact that a repudiated House, defeated by the people themselves at the general election, would still have the power to elect a President who would be in control of the country for the next 4 years. It is quite apparent that such a power ought not to exist, and that the people having expressed themselves at the ballot box should through the Representatives then selected, be able to select the President for the ensuing term....

By design, the new Congress selects the President and VP if no one gets 270 electoral votes.  Looks open and shut to me.

by Steve M 2008-10-08 08:17PM | 0 recs
##### Well, that's good news

if it was a sitting Congress, that would be more interesting though heart-aching.  The only word in the 12th Amendment that worried me was "immediately" but even Scalia will look at legislative history if the plain language is unclear (though would that apply with Constitutional amendments?)

by Blazers Edge 2008-10-08 08:22PM | 0 recs
##### Re: Well, that's good news

It's kind of confusing.  A lot of the 12th Amendment is superseded by the 20th Amendment, and the statute yitbos cited (3 USC 15) is obviously an Act of Congress rather than part of the Constitution.  But the bottom line is that Congress anticipated this exact scenario and said hey, it makes no sense to let a lame-duck Congress select the President and VP for the next 4 years (let alone allow a lame-duck VP to break the tie!).

by Steve M 2008-10-08 08:27PM | 0 recs
##### Re: Well, that's good news

Thanks, Steve.  I thought it was the new Congress.  And then I read the 12th Amendment.  Without reading the 20th.

by InigoMontoya 2008-10-08 08:39PM | 0 recs
##### Re: Well, that's good news

Until the votes are certified on Jan 6, technically there would be no tie.   Do not forget about faithless electors.  They have never been a factor in the past, but they could be a factor in a tie vote.

by gavoter 2008-10-08 09:21PM | 0 recs
##### Re: Where's Steve M. to provide his insight

Your ears must have been ringing pretty loudly...

by lori 2008-10-08 08:22PM | 0 recs
##### Re: Where's Steve M. to provide his insight

Well, it is so rare that anyone actually ASKS for my opinion...

by Steve M 2008-10-08 08:28PM | 0 recs
##### Re: That's an open constitutional

Its written into Federal Law.   The votes are NOT COUNTED by the electors.  THey just vote.  THe votes are counted by congress.   THis happens on Jan 6... it cannot happen before UNLESS federal law is changed.  At that point, the 12th amendment kicks in if there is a tie.

http://www.law.cornell.edu/uscode/html/u scode03/usc_sec_03_00000015----000-.html

by yitbos96bb 2008-10-08 08:08PM | 0 recs
##### Re: That's an open constitutional

No its not.   A five second Google search will show you that you DON'T KNOW WHAT YOU ARE TALKING ABOUT.  Please stop writing that, because it really just shows that you are not taking the time to learn you are dead wrong.

The new congress COUNTS the ELECTORAL VOTES to certify the election.  It happens on Jan 6... WHICH IS PRESCRIBED BY LAW.   If there is a tie, then immediately it is supposed to go into session to vote.  At that point, the NEW House votes for President with each State delegation casting 1 vote and the Senate votes.  Biden is a shoo-in in this case.

http://www.law.cornell.edu/uscode/html/u scode03/usc_sec_03_00000015----000-.html

The new congress is sworn in on Jan 3.  T

by yitbos96bb 2008-10-08 08:05PM | 0 recs

I found Jerome's analysis this election cycle to be pretty much wrong.
He has mostly under estimated Obama's support.

Turn out will be much higher this cycle than 2004 and young people and AA;S will vote in record number's.

Democratic registration is up all over the country while Republican registration is down. Obama's ground game is worth 2-21/2 pts to his total. I PREDICT a 5.5 national margin for Obama and a 350 EV plus total.

by BDM 2008-10-08 07:49PM | 0 recs

I see the California trip has done nothing to improve your maturity levels, Jerome. I would really like to be a fly on the wall inside your head when Obama is called the comfortable winner on Election Night.

by UMassforObama 2008-10-08 07:49PM | 0 recs

" I see the California trip has done nothing to improve your maturity levels, Jerome...."

- Spend some time in front of the mirror...

by lori 2008-10-08 07:57PM | 0 recs

There's no such thing as a lead that is big enough in the polls.

They know where our turnout will be high, and they'll do everything they can to keep them from voting.  That's why we need to keep fighting and not be complacent!

by neko608 2008-10-08 07:57PM | 0 recs

You know, if Obama is called the comfortable winner on Election Night, I'll be pretty ecstatic to be anywhere.

by Exhausted Pennsylvanian 2008-10-08 08:56PM | 0 recs
##### Question....

What about this election is anything like the ones you are comparing the models to from previous years?  This seems to be shaping up to have record turnouts, so comparing it to that from the last 3 elections seems to be pretty naive.  Or, at the very least, silly.

by Seeking Cincinnatus 2008-10-08 07:51PM | 0 recs
##### Jerome, what do you make of McCain

not giving up in Iowa?  He's going to be back there this Saturday.  I read in the Fix today that his people are claiming that their internals show a low single-digit lead.

That he has spent more time in Iowa than let's say North Carolina even though polls show a close race there suggest to me that he's going more off on demographics than anything else.  The only saving grace for him in Iowa is the fact that 30% of the population is over sixty and that Obama would need to get 35% of the white vote in North Carolina to win the state; I don't see any other reason why he would stay on offense in Iowa while not even considering playing defense in North Carolina with just one Palin visit other than playing the "demographics are destiny" game.

by Blazers Edge 2008-10-08 07:55PM | 0 recs
##### Re: Jerome, what do you make of McCain

Internal polling has the race within single digits in Iowa according to the Mccain campaign....

- Bogus in my view..

by lori 2008-10-08 07:58PM | 0 recs
##### Re: Jerome, what do you make of McCain

It ain't gonna happen for him in Iowa.

- McCain is anti-ethanol

• Fairly large student population with respect to the rest of the state, plus Illinois students get reciprocety with Iowa schools.
• Racism isn't nearly the issue in Iowa as it is in states like OH, MO, or MI
• Barack spent a lot of time there in 2007, they just really like him there

by neko608 2008-10-08 08:02PM | 0 recs
##### Re: Jerome, what do you make of McCain

I think the reason is that if McCain loses NC, he's already lost in the other more pivotal states...  In other words, NC would not be a tipping point state...

As for his obsession with Iowa, I think it's some sort of stubborn, prideful arrogance... It's like he believes he is somehow entitled to Iowa, and he's going to make damn sure he proves it!

by LordMike 2008-10-08 08:36PM | 0 recs
##### Pennsylvania

McCain was in Pennsylvania Wednesday, so he's not giving up on that state either.  Maybe he's seeing something in the dempgraphics.

I think it's a mistake; McCain needs to be in OH and FL more than he needs PA or IA.

As for the horserace, Diageo switched to a less favorable LV model.  I do wish Gallup would do the same.  My big question, are Republicans at least as motivated as Democrats?  "Traditionally,"
LV polls tend to be more favorable to Republicans.

by esconded 2008-10-08 09:31PM | 0 recs

jerome is back to pretend this is a horse race.  Michelle Malkin can breathe a sigh of relief.

Obama landslide.  I'll be here all week, folks...much to your chagrine.

by lojasmo 2008-10-08 08:01PM | 0 recs
##### Useful as a baseline

I think the "conservative" scenario is unlikely given the ground advantage Obama has this cycle compared to last election.  I just don't believe any poll showing a major fall-off from RV to LV. I can see a point or two in McCain's favor but nothing more.

I do agree, however, that the numbers out of Gallup (Obama +11) are unrealistic and even dangerous. We cannot get complacent over a few great poll numbers.

My guess is that if the election were actually held today Obama would win 51-47-1, with an electoral count of 338.

by elrod 2008-10-08 08:12PM | 0 recs
##### Re: Useful as a baseline

I know there is the perception out there , that the Obama campaign has a ground advantage but you really can't know until election day.

I heard this in PA , OH , , CA , Texas , Indiana  and some other states when I volunteered for Clinton , we were supposed to be overwhelmed by the overwhelming ground game but it just ddn't happen ...

I am not disputing that Obama has built up a great ground operation and he might have a better one depending on the state , but its not like the other side is asleep at the switch and the other side has been winning on the ground operation in past election cycles infact much of what the Obama camp is doing is at least from watching some of what the republicans have done in past election cycles...

by lori 2008-10-08 08:20PM | 0 recs
##### Re: Useful as a baseline

In PA, OH, Texas, and Indiana, Hillary had the help of Limbaugh Republicans trying to muck up the works.  In CA, Hillary had the advantage of very early voting...  Obama's near win in IN, a state that was decidedly against him, proves that his ground game can be effective....

I agree with you that we don't really know how well Obama's ground game will work int he general, but what we can be sure of is that it will be the best we've ever had!

Contrast that to the Republican ground game, which had its finest moment 4 years ago... there's very little room to grow from there...

by LordMike 2008-10-08 08:33PM | 0 recs
##### The spreadsheet predicted a six-point

win in Indiana.

by Blazers Edge 2008-10-08 08:34PM | 0 recs
##### Re: The spreadsheet predicted a six-point

That was before the kitchen sink... months before... by the time the actual primary rolled around, everyone expected Hillary to rock the state...  His close finish shocked everybody!

by LordMike 2008-10-08 08:38PM | 0 recs
##### Re: The spreadsheet predicted a six-point

The spreadsheet was created in February, three months before the primaries in Indiana.

by GermanAmericanDem 2008-10-08 08:39PM | 0 recs

raised their expectations: Obama "Indiana is the tie-breaker" (he tried to clarify the statement on MTP but it pretty suggests that he thought the that it was an equal battle in Indiana) and Clinton "North Carolina is going to be a game-changer" when she had no chance to win the state.

by Blazers Edge 2008-10-08 08:43PM | 0 recs
##### Notwithstanding the spreasheet prediction

I did think Obama did well given the context of the situation with Wright reemerging a week earlier; she could have won by 6-8 points and Obama played the expectations game to perfection by keeping it close.  She probably should have spent more time in IN and just let the chips fall where they might in IN.  It was probably worth it more to win by 8 in Indiana, rather than trying to get a loss of less than 10 in North Carolina where she had no chance given the demographics.

by Blazers Edge 2008-10-08 08:38PM | 0 recs
##### Re: Useful as a baseline

Come on she won by a pretty decent margin in a few of those states about 10 points , at least in PA , OH and CA...

You can't pin that on a few Rush republicans , how many folks even have time for that ..

The subtle point I am trying to make is that a lot of the prevailing perception is a function of front page stories in major newspapers about the ground operation of the Obama campaign , they have been quite a few stories about that and very little about the ground opration of the other side whether it is Clinton in the primary or Mccain in the general..

This feeds the perception that the opposing camp is asleep at the switch...

Take New Hampshire for Clinton , she was down about 10 points the day before the election in most polls and she still pulled out a 3 point win , that is one heck of a ground operation , kudos to Kathy Sullivan, yet I don't think any newspaper wrote about that at that time , the perception was Obama had a better operation ...

I am not saying he hasn't built one of the best , but it would be tested on election day , there is really no way to tell if it is better until then and I don't expect state by state there won't be differences...

The beautiful Commonwealth of Virginia , my neighbouring state is an example , the GOP has perfected GOTV in that state precint to precinct , voter to voter for some years now and yet there is the perception that they are being wiped out by obama in terms of gotv..

by lori 2008-10-08 08:55PM | 0 recs
##### Re: Useful as a baseline

Just because HRC won those states does not mean his ground game in those states didn't help in lessening the margin.  Hell in IN he was supposed to lose by double digits, in PA his ground game in the big cities held the electoral gains by Clinton to single digits, he tied TX.

No campaign in our political history has put together a ground game like the Obama campaign.  Without his ground game he stood no chance at beating HRC.

Don't forget his 1.5 million new donors in the past month.  He is outspending McBush 3-1 in all swing states.

by hootie4170 2008-10-08 08:43PM | 0 recs
##### Re: Useful as a baseline

My point is you can't tell if he has a ground advantage over the other side until the votes are cast .

Perception is not always reality in this case , especially when you don't know the tricks the other side has up its sleeves .

The gotv effort would vary from state to state for both campaigns in any case..

by lori 2008-10-08 09:00PM | 0 recs
##### Re: Useful as a baseline

Obama seems to have more campaign offices (that's where all the money went during the summer), more trained staffers, more trained volunteers, more outreach (more undecideds contacted) and more coordination in more of the swing states.  More undecided voters are getting contacted by the Obama campaign than the McCain campaign.

We won't know about the GOTV effort until people start voting, but it looks like the organization is there.

by the mystical vortexes of sedona 2008-10-08 09:15PM | 0 recs
##### Re: Useful as a baseline

Delegates are the prize in the primary.  By that measure, Obama WON Texas.

by fogiv 2008-10-09 09:40AM | 0 recs
##### Re: Useful as a baseline

i live in nyc in charlie rangel's district.  this should have been locked down for hillary big time.  however, obama's ground game was excellent and we fought hillary to a draw here.  the obama campaign literally had volunteers at every single subway entrance in the district and every polling place (there are approximately one every other block here) from the crack of dawn until polls closed.  the only on-the-ground troops hillary was able to muster came from the teachers union (who all left early) and a relatively small number of paid canvassers.

obama had the biggest GOTV effort i've ever seen in this neighborhood, and that's saying something.  they know what they're doing, at least with regards to maximizing urban turnout.

by bluedavid 2008-10-08 09:07PM | 0 recs
##### Re: Useful as a baseline

PA and OH were never supposed to go to Obama.   In fact, he effectively chipped away at 20-30 pt leads.   PA is also a total machine state and that machine was completely behind Clinton.  Additionally, these two states are also among the most hostile to Black candidates at a statewide level.

Indiana is a much better example to use in talking about the ground game.

I can also tell you that all the Republicans I know have already conceded the election, they have all stated that Obama is going to win.  These same people never had any doubt that Bush was going to win either time.  I do not think that McCain has any goo ground game this time.

by gavoter 2008-10-08 09:28PM | 0 recs
##### Re: Useful as a baseline

Note that I'm comparing it to Kerry as much as to McCain. Kerry's ground game was disorganized and overlapped with outside groups. Obama's has been built from the ground up over two years, including a long and vigorous primary (thank you Hillary).

McCain's ground game isn't worthless. In NC Mike Duhaime knows what he's doing. But until the Palin pick there were virtually no volunteers at all. That's changed since the RNC but the McCain camp clearly feels they are at a disadvantage. They've also wasted a lot of time and money in states they have no chance of winning (PA, MI) and have let their ground game wither in FL, VA and CO.

We won't know until election day how it turns out. But to assume that LVs wff

by elrod 2008-10-08 09:38PM | 0 recs
##### Re: Useful as a baseline

You can choose to ignore the ground advantage or say it doesn't exist until election day but it is simply not true. Obama has more volunteers, more staff and more offices than McCain and he is battling on the ground in more states.

Does that mean the ground advantage will translate into a voter advantage? No, although it seems likely. However, to say that he does not have an advantage in that department would be akin to saying Sarah Palin does not have trouble forming proper sentences.

by JDF 2008-10-09 12:04AM | 0 recs

Why are the results that actually reflect the polls "the optimistic scenario?"  Wouldn't that be the baseline?

by Jess81 2008-10-08 08:17PM | 0 recs

You can have your math. Jerome has the math.

by johnny longtorso 2008-10-09 04:15AM | 0 recs

we all know the answers to?

by sneakers563 2008-10-09 06:22AM | 0 recs
##### I predict an Obama landslide.

As for McCain spending time in Iowa?  That's perfectly understandable considering how he wow'ed them in that Des Moines Register interview, eh?

Jerome, I agree with yitbos96bb that it is really strange to rework R2K's polling results to suit yourself and then use them for predictions.  Either use a poll's results as is or don't use them.  While I've seen people discuss the demographic model used in a particular poll, I have never ever seen anyone change the results.  Why would you do such a thing?

by GFORD 2008-10-08 08:19PM | 0 recs
##### Re: I predict an Obama landslide.

It's well within the spirit of the Daily Kos poll to re-weight the likely voter model - DKos releases all of R2K's internals precisely to allow this kind of analysis.

And it's well within the spirit of Jerome Armstrong to predict our imminent doom, and I think that's a great thing, somebody's gotta keep us honest.

by Exhausted Pennsylvanian 2008-10-08 09:02PM | 0 recs
##### Jerome, I'm interested in knowing more about

meta-analysis model you're using. Of course if it is a trade secret, you don't have to share the details.

I find Prof. Wang's meta-analysis model of polls very persuasive. Maybe because I'm partial to Matlab scripting, more possibly because Wang clearly enunciated his methodology. He is currently showing Meta-margin of 5.78% in favor of Obama (EV: Obama 353, McCain 185).

Personally I would have preferred to see a Bayesian model for Meta-analysis.

by louisprandtl 2008-10-08 08:22PM | 0 recs
##### Re: Jerome, I'm interested in knowing more about

A Bayesian model wouldn't show anything different (assuming noninformative priors)--it'd just be easier to interpret.

by slynch 2008-10-08 08:45PM | 0 recs
##### As a LIV, anything that is easy to interpret

is preferable...:) I'm assuming your comment is relative to Prof. Wang's methodology.

by louisprandtl 2008-10-08 08:56PM | 0 recs
##### Re: As a LIV, anything that is easy to interpret

I'm ultimately a Bayesian statistician, but I use both classical and Bayesian methods.  The main benefits, as far as I'm concerned, with the Bayesian approach are (1) the ease of modern Bayesian estimation--I'm better at simulation than numerical methods; and (2) the coherence and probabilistic interpretability of Bayesian results is extremely appealing.  What Nate Silver, and Wang (although I'm less familiar with that work--I've just glanced at it) do with their projections is inherently a Bayesian "wannabe" approach.  The results are almost the same as what a Bayesian would come up with, but the interpretability is more difficult.  Based on Wang's electoral college distribution, I'd say there's a greater than 95% chance that Obama will win this thing.  That interpretation can't be made from a classical standpoint.

Anyway, if you have a spare \$60, you might buy my book (go to amazon.com and search on "lynch bayesian").  Hell, I get something like \$4 or \$5 for each copy sold.  I'd like to retire on it, but it ain't exactly making the NYT best sellers list :)

by slynch 2008-10-09 07:55AM | 0 recs
##### Thank you for the great reply...It is always an

honor to talk to a Professor and an expert in the field of statistics (you probably know your colleague Prof. Wang). I've actually seen your book before. Obviously you can tell, I'm no expert in this area. I was staying away from Nate Silver's (538.com) approach because I haven't seen the methodology they implement.

I dabble with Bayesian methods for information fusion of signal data from heterogeneous sensor array elements...we are also looking at Fischer and Mahananolobis distance measures. In a hierarchical system, the fusion study becomes important (hence usage of such meta-analysis). Nothing to do with social population statistics..

In future, please see whether you can attend one of NASA ARC's Data Mining conferences(used to be called DMASES and now called CIDU). Folks in this research area always seek guidance from experts in other fields. Offline please send me an email if interested, I can forward your information to right folks who are in charge of the Data Mining Lab.

by louisprandtl 2008-10-09 08:34AM | 0 recs
##### sorry I got spelling of Mahanalobis wrong..

by louisprandtl 2008-10-09 08:41AM | 0 recs
##### hah..it should be Mahalanobis...now I'm making

an complete fool out of myself...

by louisprandtl 2008-10-09 08:42AM | 0 recs
##### Re: Thank you for the great reply.

You're too kind.  It's no honor to talk to me whatsoever!  In fact, it sounds like you are a wolf in sheep's clothing, so to speak.  Sounds like you're doing much more sophisticated stuff than I am.  My stuff is easy, relative to what it sounds like you are doing!

No sweat with the Mah..WTF distances.  To be honest, I know what Mahalonobis (not sure I have the sp. right) distances are (I can compute them, because they are simply the kernel of an MVN density, but I never use them personally.  Although, we do use them in some of the more recent "advanced" methods that my colleagues and I are playing around with regarding things like religious affiliation and various political perspectives--and change over time therein.  But, again, it sounds like you're doing much more sophisticated stuff.  Frankly, I have no idea what your post means, regarding what you are doing.  And I HAVE worked with physicists in my recent past!

My whole reason for engaging you was that I thought that this entire post (Jeronme's) was oversimple (not to mention the fact that you always make good comments!).  You know this, and responded so, and I was just confirming your "suspicion."  Jerome's--and others' strategies are really ridiculous.  There's really no statistical-theoretic basis for them.  If I had the time and energy, I'd do my own forecasting, but, unfortunately, I don't have the time nor energy.  My colleague Wang does (even though I don't even know her/him).  I trust his/her conclusions over Nate's or anyone else's, even though I don't have the time to verify them.

I do know, because I spent time doing some Bayesian forecasting of the primaries, that a lot of the claims here regarding Obama's inevitable capture of the nomination were statistically bogus.  But, that's old news these days: he DID end up winning.  It just wasn't as predictable as some folks claim.  But, again, that's water under the bridge.

by slynch 2008-10-09 09:24PM | 0 recs
Jerome, McCain will not break 45% nationally.
This will be due to the demoralization of the Republican coalition. So subtract 3-5% from any McCain number.
by Judeling 2008-10-08 08:38PM | 0 recs

I'm not sure I understand your spreadsheet, Jerome.  I'm not sure if it's labeled poorly or I'm just being stupid.

Also, it looks to me like your Dem party ID advantage ranges from 2% (most pessimistic) to ~3.5% (most optimistic).  Isn't this really conservative?

by the mystical vortexes of sedona 2008-10-08 09:02PM | 0 recs

Of course it is.  I suspect if Clinton had won we'd see much more optimistic party ID numbers ;).

by NMMatt 2008-10-08 09:22PM | 0 recs

Wang nailed the last election.  Nate at fivethirtyeight nailed most of the primaries.  They have +5.8 and +4.5 translating into electoral college landslides of well past 300.  Nate has the race tightening built into his model.

Note the similarity to Clinton v. Bush where Clinton got +5.6% and 370 EV.

by NMMatt 2008-10-08 09:18PM | 0 recs
##### Overly pessimistic

Of course the model here is completely overly pessimistic.    Here in Georgia, several county elections officers have already come out and stated that they are expecting 80-90% turnouts for this election. Yes, over 80%.    Even if they are too high by a third, we are still looking at a minimum of 60%.

Early voting here is huge and people are already standing in lines to vote.   I think turnout this year is going to blow the lid off anything in the past.  I honestly do not think that there is ANY valid likely voter model that can be used to predict this year.   Anyone who is attempting that is a fool.

by gavoter 2008-10-08 09:39PM | 0 recs
##### Humorous Digression
Did anybody else notice the large A-frame ladder in front of the eagle logo post-debate?
It looked like an Anheuser-Busch Logo in back of the wrap up shots.
Did the McCains rush off so Cindy could slip the stagehands a few bucks?
by Judeling 2008-10-08 09:47PM | 0 recs

I'm not sure why the R2000 poll qualifies for re-weighting but Hotline (which recently had a Dem advantage of 2%) does not.  Zogby, I suspect, also has suspicious weighting, and Battlefield has already had to re-weight once because their numbers were off.  If you're going to start re-weighting polls yourself, you might as well identify what you believe to be the right weighting and apply it to each poll.

by rfahey22 2008-10-08 10:20PM | 0 recs

But that wouldn't make it look as close...

by JDF 2008-10-09 12:00AM | 0 recs

More simply put: If the poll shows Obama up there must be something wrong with the poll, If the poll shows the race close the poll was done correctly.

by logic is beautiful 2008-10-09 06:41AM | 0 recs

Zogby's a good poll suddenly because they moved to a telephone model?

The same Zogby that said on ELECTION DAY in 2004 that Kerry was going to win Virginia?

Whatever you say, Jerome.

by leshrac55 2008-10-08 11:08PM | 0 recs

eh.  At this point I'll take a .3% win.  I just really want to win.

by Skaje 2008-10-09 12:40AM | 0 recs
##### Hotline and Zogby have Obama up today

Zogby has Obama moving to +4 today (+2 yesterday).

And Hotline has Obama going from +1 to +6 today. They either readjusted their party ID numbers or a bad polling day rolled off.

R2K had a one-night sample of Obama +12 last night.

Rasmussen had a slight tightening from Obama+6 to Obama+5 (50-45 now).

What does this mean for Jerome's spreadsheet?

by elrod 2008-10-09 05:38AM | 0 recs
##### Re: Hotline and Zogby have Obama up today

If you run the numbers and use Armstrong's methodology, the race is now much closer to +3% Obama.

by devilrays 2008-10-09 08:11AM | 0 recs

If you have to preface your analysis with "Zogby...it's a good poll" you are truly grasping at straws. Clinton partisans like Mark Penn and Howard Wolfson don't even buy into this "analysis".

by logic is beautiful 2008-10-09 06:37AM | 0 recs

One final question for Jerome

When you say that you are going to re-adjust the voter turnout model for research 2000, you are presuming that research 2000 is willing to tarnish their reputation to make their client (dailykos) happy.

What is your basis for this presumption?

by logic is beautiful 2008-10-09 06:49AM | 0 recs

You haven't missed much?

While you were away Obama firmly established his lead nationally and in the states.

It's like you wrote this before you left--strange.

by set 2008-10-09 09:51AM | 0 recs