The tightening in PA comes from his own voters coming back to him.  And he still can't seem to top 45 or 46 percent.  Obama's numbers have held steady in 50s.  I don't know what the worry is about.

As for Mason-Dixon, they did so well in 2004 they apparently are using the likely voter model from 2004 for 2008.  I'd like to see what the numbers are for registered voters before they lop off a significant number of Obama voters for being "unlikely to vote."

Nevertheless, what's important is look at the trends within the polling, and Mason-Dixon is showing Obama increasing his lead in every one of their polls.

Mason-Dixon does have a slight GOP lean.  The one danger right now for Obama is losing the Electoral College while winning the popular vote.

So many undecideds in their PA and VA polls--if most of the undecideds break to McCain. doesn't he win them both?

Jerome, I don't think that the 'more sensitive' setting offered by Pollster makes any sense for at state level because there aren't enough polls.
(Even at national level it doesn't make much sense, as the curve jerks up and down constantly.)

The output here is a curve which goes backward: clearly a meaningless artefact. The average sensitivity (default setting) makes much more sense, and does show a slight narrowing - not at all of the same kind as the one suggested on the graph you gave.

I live in Ohio and early voting is way up from 2004, and it's overwhelmingly democratic.  I stood in line for 2 hours to vote early this weekend, in a line full of "sporadic" voters.  There is no way that Obama is only at 45% in Ohio.  

I'm not sure that Mason-Dixon's supposed accuracy in 2004 means anything in 2008, and I suspect that we'll have a new "most accurate" pollster this time around.  I much prefer looking at the aggregate of polls than following any one particular pollster.

Unless the polling this year across the various polling groups is identical (which it is not), we can't look at the numbers from four years ago.  It's obvious that some pollsters are basing their results on 2004 data, not 2008 data.

M-D has kept it close all this cycle.  They were accurate on several states in 2004.  Yet look at a pollster such as SurveyUSA.  They too were accurate in several of the same states, but they are showing better results for Obama this time around.

M-D is not pushing the leaners and is not using the current data.

Regardless, all I want is a win, and the easiest way to get to 270 is Kerry states + Iowa + Colorado + New Mexico or Nevada.

Anything after that--and I still feel Obama will win with 350+ electoral votes--is just icing on the cake.

One difficulty the Democrats are having in the state is they're undrperfroming downballot.  Pennsylvania could be one of just a few states the Dems lose House seats.

McCain doesn't break 200.

M-D in 2004 got EVERY state winner correct and was off on the margin of victory by as many as 4 points in only one poll.  Applying that standard and forgetting that they are 4 points more pro-McCain than everyone else, Obama still has NV, CO, and PA in the bag and is likely to win VA and FL as well.  That's more than enough.

The five to watch are:

NV: the No. 1 back-up plan for a freak loss in PA:

Kerry- PA +IA +NM +CO +VA +NV; we are killing them on the ground in NV, maybe only exceeded by the gap in NC and IN.  GOTV will win Nevada for us.

PA: they could have squeezed a last trip to PA in the schedule and they have not - as of now.

VA: go to Manassas on Monday night and do a live blog, be part of history.

CO: the early vote gap in CO is small.  I'm watching for the Jesus people from Colorado Springs going up to the big south Denver suburbs- Aurora, Highlands Ranch, for big time last minute GOTV.  We have to win CO.  

OH: back-up plan No. 2 for a freak PA loss.  Three stops for the campaign today in Ohio, plus HRC.  We got burned in OH in the primary, thought it would be closer.  M-D Ohio poll made me wet myself.

FL is too close, large and complicated to figure out.  NC is icing on the cake.

I've long believed the undecideds would break heavily for McSame because of their demographic make-up - mostly white with lots of "Reagan" Democrats. They are undecided because they don't like McSame, but they'll never vote for Obama.

I believe most will come out to vote for McCain and the rest will stay home or not cast a vote for Prez. It's going to get a lot closer.

I also think that a lot of undecides will break McCain just like they broke for Gore in 2000.  However, there is just no way that BO loses PA.  I cant remember the last time the dems lost PA but it must have been reagan???

But really my eyes are on VA.  VA is going blue faster than anyone could have guesses and northern VA is really blue.  My understanding is VA closes the polls at 7 pm EST in imho by 8:30 they will call it for BO and that will be that.  No way can McCain win the election without VA.

david

Yes, it appears tighter today in most of the swing states.  I'm concerned about PA due to the composition of rural PA Democrats.  Will they vote in force for Obama?  Meanwhile, the MD polls show worst case scenario in all listed states.  It still looks pretty good, but I guess I'll have to wait 2 and a half days to celebrate.

Most of these state polls have been lagging indicators throughout this election season, taking a few days to catch up to the national trends. Obama's mini surge post infomercial should give him what he needs in all of these states.

2.6 million votes so far.  41% of registered voters.  75% of the total vote from 2004.

51% turnout already amoungst blacks, 26% of the total.  Whites comprise only 69% of the electorate so far.  

Dem/Rep/Un breakdown is 52/30/18.

Based on reports of polling of early voters, Obama is leading early by at least 8 pts (conservatively, PPP had it at 25 pts earlier in the week).  If we assume that turnout Nov4 is massive (2 million more votes), McCain needs to win the election day vote by more than 11%.  I'm struggling to keep my optimism cautious.

Eeking out a win is not good enough for me. I can not see how Obama can fall below 349 EV's .... these goofy last few days polling are just that. Goofy and statistical noise. PA stays with Obama as does OH and FL and VA and NV and MO, read em and weep conservative movement that will last until the rapture!!! One of the remaining suspense items will be to see how long the corporate media can drag out the night. Another would be the huge turnout favoring Obama in GA, MS, AZ.

I don't trust pollsters that don't list internals. Show the assumptions and why result come out the way they do.

Yes, the polls are tighter but there are also fewer undecideds now (at least in the national polls).  Stands to reason that more undecideds would drift to McCain.  Most of them were probably unhappy Republicans and right leaning independents anyway who, in the end, would have to have an extremely compelling and irresistible reason to vote Democratic.  

The RCP average of national polls shows about a 6.5% Obama national lead with around 95% of the vote going to the top two candidates.  If 2% vote for other candidates and McCain gets 2 or the 3% of remaining undecideds, Obama still wins by 5 or 5.5%.

Other day when Mason-Dixon posted a pol with Norm Coleman leading Al Franken, I had an email exchange with Jennifer Duffy because, at the very same time, she predicted Democrats picking up 8-10 seats. Her response to was that "M-D appears to rely on a more normal turn out, which we know won't be the case, so these numbers are a little optimistic for Republicans generally."

Not trying to debunk the polls numbers, but she may know things we don't.

Thankfully only one day left till we find out. I couldn't take another week of this.

has black turnout at 10 percent and 80% for Obama

2004 had black turnout at 13 percent and 84% for Kerry.

If a pollster puts out the information, you can judge the credibility of a particular sample.

obviously, I think blacks will be more than 13 percent of Pa, and be higher than 80% for Obama

they had one out yesterday but now there's another today - perhaps they were hoping the story was going to get even juicier.

Today's poll shows a slight tick up for Obama not just in his polling number, but also favorability. After several days of falling numbers in PA I'll take any uptick for a bit of reassurance.

Details here:

http://www.rasmussenreports.com/public_c ontent/politics/election_20082/2008_pres idential_election/pennsylvania/election_ 2008_pennsylvania_presidential_election2

POV Firewall:

Assuming Obama takes the interior west states of Nevada, New Mexico, and Colorado, all he needs is to take one of these states:

Pennsylvania--No early voting, an aging rural Democratic base outside of the major cities.  Not Obama's base constituency, but PA has gone Blue in every close election.  Reality says that BO leads by 3-7% with fifty hours to go.  He should hold on, but I wish he would visit the state in the next two days.

Ohio--Probably the best Republican organization in a swing state, decided the '04 election for Bush, again it is a state where, outside of the major cities, Obama really has to work hard to get votes.  Ohio was bright blue in '06 and having a Dem gov. and SOS helps.  Polls are close with Obama having the edge.  It should be close on Tuesday.

Virginia--These coastal states are becoming more like Maryland and less like Alabama every year.  Virginia will soon have two Democratic Senators, and a Democratic governor.  Obama has led since August here and I think he wins a close election here.

Anyway, win one of the three and anxiously wait for the blue states to the west to come in.  Win two and relax or win three and start to get greedy, hoping for a landslide and big numbers downticket.

Lose them all=President Grandpa.