Pollster.com

Articles and Analysis

 

Morning Status Update for Sunday 11/02

Topics: Status Update

The most important thing to keep in mind about pre-election polls is that they come with random variability built-in. That mostly predictable variation -- known to most of us as "the margin of error' -- comes from interviewing what we hope amounts to a random sample of likely voters rather than everyone that casts a ballot. So, if you look at a batch of new polls every day, the law of random chance guarantees that some polls will show your favorite candidate doing a little better and some will show that candidate doing a little worse. Moreover, one poll in twenty should produce, by chance alone, an odd result that falls outside of the reported margin of error (since that statistic is typically based on a 95% level of statistical confidence).

Given the number of new polls we are now seeing every day, it is all too easy to cherry pick one poll -- or easier still, one subgroup in one poll -- that seems to indicate a sudden, seemingly dramatic change and blow it up with a misleading scare headline.

The best advice I can offer right now, especially to journalists, is to emphasize what most of the surveys are telling us consistently and ignore the odd, contrary result, especially if hyped beyond all recognition by the pollsters themselves. More often than not, that odd "finding" turns out to be meaningless noise. The pollsters that shamelessly hype their results one day, take it back the next, while issuing warnings that important blocks of voters remain "volatile." Hardly. At this point, any such volatility tells us more about shortcomings in the polls than about volatility in the voters.

Taking a broad view of all the new data before us, it is hard to see much change of any significance in the race for president except a steady increase in the percentage that report having cast an early ballot, a steady dwindling in the percentage that are still undecided or persuadable and perhaps some uptick in McCain's support in a state or two (like Pennsylvania) where his campaign is expending resources disproportionately. None of our state classifications has changed in the last 24 hours.

On to the surveys we logged on Saturday. This weekend represents a relative lull before one final polling storm to come over the next 48 hours or so. Yesterday brought 17 new statewide surveys and 10 new national polls. Nine of the state level polls represent updates of tracking polls conducted by the same pollster within the last two weeks. Of these, 5 show nominal gains for McCain, 2 for Obama and two with no change in margin. On the other hand, 6 of the 10 new national releases show nominal shifts to Obama and three show no change in margin since their last non-overlapping sample -- none show any improvement for McCain.

081102 new polls

When we examine how the new polls affect our trend estimates in key states, the results are typically small -- less than two tenths of a percent in 4 of the 7 affected battleground states. The overall pattern is similar to that of the individual polls: 5 states show nominal ticks in McCain's direction, 3 in Obama's direction.

081102 trends

The one place where polls show unambiguous evidence of narrowing margin is Pennsylvania, where the McCain campaign has placed great emphasis (both in candidate visits and television advertising) over the last week. Three new releases yesterday -- including the daily tracking from Muhlenberg University -- show Obama leading by margins varying 4 to 7 points. Obama still leads on our trend estimate by almost 8 points (51.6% to 43.7%) but the margin has narrowed nearly 5 points in the last week.

What is less obvious from the table above is that most of the change in Pennsylvania involves an increase in McCain's support -- from 40.3% to 43.8% -- while Obama has lost just a single point on our estimate (from 52.8% to 51.6%).

Notice also that our trend estimate mostly ignores the surprising 16-point Obama lead on an Ohio University Poll released yesterday. Our trend line tends to ignore outliers, especially one that finished interviewing more than a week ago.

Back to the national trend. There is simply no evidence as of yesterday's releases of a late narrowing of Obama's margin. The margin on our national trend has ticked up about a point in Obama's favor over the last two days. As usual, none of the national tracking surveys showed anything approaching a statistically significant change. Five of those results showed small, nominal shifts in Obama's direction, two showed unchanged margins and only one showed a small shift in McCain's direction.


081102 trackers yesterday

And yes, the table does include one new "tracker." Yesterday's new results from CBS News include the last two nights of interviewing from the last CBS/New York Times poll and CBS plans to release new results today and tomorrow. Of course we should see new national poll results in the next 48 hours from most of the organizations that conduct national surveys. So the tracker/non-tracker distinction is about to become largely irrelevant.

 

Comments
Spike:

Mark.

I wonder whether your model should pay more attention to the most recent polls at this stage. I agree with your approach intuitively (I'm no expert) for most of the campaign. But at his point, the mathmatical prediction must collapse to an actual state of reality in the electorate on Tuesday. So a poll taken today ought to have more significance now, than what it would under typical circumstances (say a month ago).

____________________

jsh1120:

I believe I noted this in another post, but if so it bears repeating.

I'm hesitant to accuse a national pollster of cooking the results of their work, but the evidence from Zogby's work over the entire year at least raises the question.

As nearly as I can tell, the best predictor of Zogby's results is whatever can be touted as "news." The changes in his results over time; the ridiculous "internet polls" he has published, and his "analysis" all point to a pollster less interested in measuring public opinion than in making headlines.

Now, in the last few days before the election we see Zogby's results beginning to converge with those of other organizations. While that may, indeed, represent what his organization finds, I have to say that the hypothesis that he is simply attempting to avoid being pegged as wildly off the mark in the election postmortems cannot be rejected.

____________________

wjbill49:

@jsh1120
well said re Zogby's results (Zogby's ego)

I believe it is too late for McCain in PA .... there are also so many easy scenario's for Obama even without PA

____________________

enadknock:

about pennsylvania, it sounds like most of the undecided voters are going mccain. do you think this will be a national trend? is the "bradley effect" hiding here? in which case, any state poll which shows obama with less than 50% should be viewed with a pound of salt.

in another vein, can you refer me to reputable polls that have early voter results?

____________________

NM:

In New Mexico, an Albuquerque Journal poll released this morning shows Obama by 8. A month ago they had Obama by 5.

http://www.abqjournal.com/elex/0210257elex11-02-08.htm?splashtop1


____________________

RWCOLE:

Polls released today are the first numbers that will be held against the tracking polls come election day- at least those doing three day rolling averages. If anyone's been screwin with the numbers, it's likely to end today.

____________________

NC-Steve:

1- Fabulous site, I love it.

2- I agree with the suggestion that more recent polls be given more weight. I defer to your statistical prowess, but it seems to me that the end of the trend line isn't necessarily reflecting current reality.

3- Minor question - on the individual state polls, the color of the datapoint dots doesn't necessarily correlate to the party who scored higher in the poll. For instance, a red dot might correspond to a poll where Obama had a significant lead.

____________________

NM:

NC Steve "Minor question - on the individual state polls, the color of the datapoint dots doesn't necessarily correlate to the party who scored higher in the poll."

Steve, if you see a red data point for a poll, you will always find a corresponding blue data point for the same poll.

____________________

NashvilleLefty:

Pollster guys: thanks for all your work this election season.

____________________

Bryce:

Spike,

I too have the same skepticism, so I tend to look at the individual states with the high sensitivity filter turned on. This gives a more accurate picture of short term trends in the race.

Well, theoretically at least. I've yet to go and see what the latest Ohio Univ. OH polling anomaly has done to that race.

____________________

RWCOLE:

Seems like more than normal variation in the Pennsylvania polling- someone tryin to put their thumb on the scale?

____________________

MattF:

I'm glad to see your clear discussion of statistical variation.

One thing I sometimes get irritated about is when people misuse the word 'outlier'-- one poll out of ten being lower than the others is --totally normal statistical variation--. If, on the other hand, all the polls gave -exactly- the same result, -that- would be an outlier that required an explanation.

Thank you for letting me vent on this subject.

____________________

dw:

about pennsylvania, it sounds like most of the undecided voters are going mccain. do you think this will be a national trend?

I'm thinking no, if this really is what it is. McCain is absolutely pounding PA senseless with commercials and continuous news cycle domination. Look at the overall traction McCain's getting in the PA vote -- +4.6 in a week -- compared to other states. McCain may be cajoling and scaring undecideds to switch in PA, but it's also the only place he's really spending his resources.

I really doubt McCain will get all the undecideds to flip GOP in all states -- there's been some great analysis on the undecided vote here and over on 538, most of which suggests undecideds will break pretty evenly. Even if McCain still does win all remaining undecideds, he'll still lose the election. PA in the current tally is 51.6%. OH is 49.7% -- effectively 50%. Obama wins PA + OH, election over. In order for McCain to win this election, he'll need to win all the undecideds AND erode 4-6% of Obama's highly motivated base across the board.

is the "bradley effect" hiding here? in which case, any state poll which shows obama with less than 50% should be viewed with a pound of salt.

By definition the Bradley Effect can't hide in the undecideds, since the Bradley Effect is about lying to pollsters about supporting Candidate A while voting for Candidate B.
And anyway, the Bradley Effect has been pretty well debunked at this point.

If anything, there may be indications of a "reverse Bradley effect," in that African-American support may be understated in polls.

McCain's last two hopes, other than the dead girl/live Osama scenarios, are that the Christian conservatives that saved Dubya in the final 72 hours will finally show up, or that the polls are missing a segment of the population that's very pro-McCain. On the former, it's looking more and more unlikely, and the McCain GOTV machine isn't anything like Bush's in 2004, so even if they did show up, would the McCain folks be able to get them to the polls? On the latter, the big problem is that the most conventional polling method -- calling people on land lines -- is already biases towards McCain, since his supporters are more than likely to have land lines. There's more evidence that Obama's supporters aren't getting picked up by classical polling methods (though how much they're getting understated is open to discussion).

If you're rooting for McCain, you're hoping for a miracle at this point. And it's not just a Hail Mary, it's a Hail Mary, recovering the onside kick, then hucking another Hail Mary. I don't think it's a question of whether Obama's going to win, it's now a question of how much he'll win by.

____________________

chuck412:

Unfortunately, I, like (based on what I've seen), many others, seem to be hard-wired to permit emotional investment to trump rational analysis. Nothing personal, but I'll be thrilled when it's all over and your great site isn't necessarily the one I hit first thing at sunrise.

____________________

chuck412:

Unfortunately, I, like (based on what I've seen), many others, seem to be hard-wired to permit emotional investment to trump rational analysis. Nothing personal, but I'll be thrilled when it's all over and your great site isn't necessarily the one I hit first thing at sunrise.

____________________

machomaas:

Good analysis by the previous pollster. McCain's only path to the WH must include some kind of nationwide shift in the polls of more than 5 percent, throwing all the toss-ups in the McCain column and getting him somewhere in reach of NV, CO, VA, OH and PA (although all these states are currently polling >5% there is the MOE). It makes me wonder if the O campaign really should be throwing money into red states instead of trying to keep/expand his leads in say OH and VA and insulate these crucial states from a nationwide shift.

____________________

Mark,

Can it be that Zogby is doing his tracking polls in different geographical regions in any given day? There seems to be days with surges for McCain, which would happen if you were concentrating on the South and Mountain West and other days with upticks for Obama. Overall, this would keeps his numbers not to far from the totals from other polls, but produce some 'newsworthy' results.

I do not know if you are aware of the Zogby poll done in 1998, a few days before the Schumer DaMato race. His results were neck and neck (front page New York Post headline), while everyone else was seeing a landslide. Schumer won by 10 points. Zogby later stated that he had made some unfortunate
'adjustments' about who was going to vote.

____________________

It seems that certain factors, such as a higher than normal voter turnout of African-Americans and young people, will be working in Obama's favor as well as the large number of newly registered Democrats. To what extent are these factors being considered in the polling? If these factors are being underweighted then Obama may win by a greater margain than expected. What do you think?

____________________

bowlrmark:

Cluallen, so far in early voting (at least if FL), young voters are the worst performing demographic. That would be worrisome to a degree in that state. Hopefully, the "reverse Bradley" effect will occur (whites not wanting to admit voting for a black). Anyone have any data that shows a reverse effect?

____________________

enadknock:

DW, thank you for your insightful response. i'm a worry wart who will never forget the 2000 election.

by the way, realclearpolitics moved ohio into its toss up column because of today's mason-dixon poll. and yet mason-dixon has never polled obama as ahead in that state for the past two months. i guess the recency of this latest poll has more gravitas.

____________________

tom brady:

I'm not worried about the Bradley effect - I'm worried about the social conformity effect. We know that about 12% of Americans tell pollsters they have voted, when they haven't, and a similar number say they are registered when they aren't. The reason is that poll respondents want to conform to socially acceptable norms of behavior. In this election cycle many Americans view the election of Obama as a change to finally put the legacy of racism in this country to bed once and for all. They can see the long lines, favorable meeting coverage, record registration, etc., as visible indicators of support for Obama. My worry is that they feel pressured to tell pollsters they are supporting Obama, when in fact for reasons of political ideology, worry about Obama's lack of experience, etc., they do NOT support Obama, and wouldn't support him even if he was white. That's my worry about polling bias - not that a significant subset of Americans are racist, but that they are conformists.

____________________

green baby green:

Looking good!...It's amazing how calm these posts are now the Boom has been given the boot!

____________________

george:

@tom brady

What is the source of the data about respondents misreporting prior voting and registration? Is it, in fact, from "self-reported" info in polls? On the face of it, it would seem to be extremely difficult to have objective data on the question.

____________________

Tom Brady, you are correct that the "Bradley Effect" is one based on social conformity and it's not all about race. Yes, that's one factor, but I think it's over-stated. There's no doubt that people are going to report being LV for Obama because they know they are "supposed to" socially.

What we have to count on is that this effect is countered by the LV pool being skewed too old. As much as 45 of this electorate will be Gen-X and Millenial, and I rarely see them polled at that rate.

____________________

douglasdao:

hsh1120:

I agree with you and though I posted this earlier in another thread it bears repeating:

Zogby isn't deliberately crooked in his numbers, but he easily succumbs to the temptation to chase headlines and his press releases are misleading. This has always been the case with him. Zogby will bring out a "look at me" press release no matter what his actual numbers say. As Thatcher pointed out - there is no way what Zogby is reporting to be his numbers is actually his numbers. I suspect what he is doing is playing fast and loose with his splits because he is the only pollster I am aware of who takes two daily samples group - AM and PM - rather than a larger single day sample. A single one of those splits is a very small sample with a very high margin of error. What I think he does is when it suits him to get a headline, matches up his splits to get a result he wants. So, his normal Wed. PM + Thu AM not showing a McCain closure? He can pick Thu AM + Thu PM, or whatever. Numbers don't lie, but you sure as hell can lie with numbers.

I think Zogby polling is mostly about Zogby. Outside of totally obsessed political wonks, who can name the head of Gallup, Hotline, Battleground, R2K, or any other polling company? Zogby puts himself just way too much out there for a reputable pollster.

____________________

Aristotle:

machomaas:

It's an interesting point you bring up. In fact it really tells you alot about the candidate himself.

First, it's like watching an artist manipulate the controls of his media (not meant as punditry but rather as "canvass") He is completely attacking in the areas that keep his opponent off balance. If his internals showed him a race too close to call in PA or VA I'm sure he would compensate.

Second, He is doing exactly what he said he would do. Run a canvass campaign. He wanted to go to the red states and he has not only done that but done that excessively well.

Third, look at what it means for him as the leader of the country. He didn't leave a large demographic out of his campaign. He talked to everyone. This makes his chances of effectively managing the country all the more likely. Just because he is going to have a majority doesn't mean that it will be easy for him to manage that mess. They're going to want to force a bunch of personal junk down his throat. He has to be able to go to the people and ask for support in saying no.

I truly think he is willing to give up his presidency for his plan and ideals. Look at how long range and effective his planning has been. This guy is like a savant.

I don't mean to sound like I've drank the cool-aide here people. But this guy is going to change the way we look at leadership. When Obama's done as president in 8. (And I say this with all do respect) The leaders like Sarah Palin and L. Dole are going to be completely off the menu. The American People will want People like Obama and maybe people like Jindal. Politics will be secondary to Capability me thinks.

My 2 cents!

____________________

RWCOLE:

Investor's Business Daily now shows Obama ahead by only two points....

What do you call a continuous outlier? A liar?

____________________

douglasdao:

RWCOLE:

I think we'd call that an out and out liar.

____________________

carl29:

Investor's Business Daily is for rigt-wingers consuming, so obviously that they are rallying the troops at this point :-)

____________________

SusanD:

Mark,

Thank you for running such a great site. Your steady, reasoned approach and presentation has given me tremendous peace of mind during this voting season. Your steady communication of the facts has been a source of comfort, and your work spared me a great deal of anxiety over the last six months - I mean like improving my physical health level of anxiety.

Best of luck to you and your family. Keep up the good work.

Respectfully,

Susan Dunlap

____________________

Clearly, everyone tracking voters who have already voted at the national level is finding that roughly 20-25% of their samples have already cast ballots.

Those who report the preferences of those having already voted, at both state and national levels, generally show Obama ahead, but the amount of Obama's lead in the national polls among early voters varies considerably. Zogby and CBS/NYT have Obama up by 17-19 points, whereas Hotline has him up by 5 points among early voters.

I have compiled the results of all early voting exit polls--both national and state--on my blog at: http://xcurmudgeon.blogspot.com/2008/10/early-voting-exit-polls-and-demographic.html

____________________

Basil:

Thanks Mark. You're a model of rationality, which is something of an outlier in cyberspace.

____________________

ricochet:

A letter to the Red States respectfully submitted by the Blue States.
Dear Red States:

We've decided we're leaving. We intend to form our own country, and we're taking the other Blue States with us. In case you aren't aware, that includes California, Hawaii, Oregon , Washington, Minnesota, Wisconsin, Michigan, Illinois and all the Northeast.
We believe this split will be beneficial to the nation, and especially to the people of the new country of New Pacifica.

To sum up briefly: you get Texas, Oklahoma and all the slave states. We get stem cell research and the best beaches. We get the Statue of Liberty. You get Dollywood. We get Intel and Microsoft. You get WorldCom. We get Harvard. You get Ole' Miss. We get 85 percent of America's venture capital and entrepreneurs. You get Alabama. We get two-thirds of the tax revenue, you get to make the red states pay their fair share.

Since our aggregate divorce rate is 22 percent lower than the Christian Coalition's, we get a bunch of happy families. You get a bunch of single moms. Please be aware that Nuevo Pacifica will be pro-choice and anti-war, and we're going to want all our citizens back from Iraq at once. If you need people to fight, ask your evangelicals. They have kids they're apparently willing to send to their deaths for no purpose, and they don't care if you don't show pictures of their children's caskets coming home. We do wish you
success in Iraq and hope that the WMDs turn up, but we're not willing to spend our resources in Bush's Quagmire.

With the Blue States in hand, we will have firm control of 80 percent of the country's fresh water, more than 90 percent of the pineapple and lettuce, 92 percent of the nation's fresh fruit, 95 percent of America's quality wines (you can serve French wines at state dinners) 90 percent of all cheese, 90 percent of the high tech industry, most of the U.S. low-sulfur coal, all living redwoods, sequoias and condors, all the Ivy and Seven Sister schools, plus Stanford, Cal Tech and MIT. With the Red States, on the other hand, you will have to cope with 88 percent of all obese Americans (and their projected health care costs), 92 percent of all U.S. mosquitoes, nearly 100 percent of the tornadoes, 90 percent of the hurricanes, 99 percent of all Southern Baptists, virtually 100 percent of all televangelists, Rush Limbaugh, Bob Jones University, Clemson, and the University of Georgia. We get Hollywood and Yosemite, thank you.

Additionally, 38 percent of those in the Red states believe Jonah was actually swallowed by a whale, 62 percent believe life is sacred unless we're discussing the death penalty or gun laws, 44 percent say that evolution is only a theory, 53 percent that Saddam was involved in 9/11 and 61 percent of you crazy bastards believe you are people with higher morals then we lefties.

Finally, we're taking the good pot, too. You can have that dirt weed they grow in Mexico

Peace out,

The Blue States

____________________

hjf:

Mark, congrats on the almost 100,000,000 hits recorded by sitemeter.

____________________

Banquo:

There would appear to be several unknowables at work that might favor Obama. One big one is how many newly registered voters will actually show up. The Obama campaign seems to be successfully expressing an urgency and inspiring an enthusiasm which should up what pollsters figure. (I think I saw a figure from past elections of only 20% of new registers actually voting.) Anecdotal evidence suggests early voters, contrary to past years, are voting strongly Democratic. That raises another question of how many later voters decide not to vote, especially if there are long lines or bad weather. Here again, Obama supporters seem to be more committed than many of McCain's. There is a good deal of conflictedness, especially on the Republican side (like him/hate her, love him/not so keen on him, don't like this ticket but can't bring myself to vote for the other one) and considerable disgust over the whole process which could translate into no-show voters. As for the so-called Bradley Effect, as others have commented, there may be factors working in Obama's favor. There may be as many voters reluctant admit they wouldn't vote for a woman or an old man as for a black candidate. In some parts of the country, there are probably a good deal of voters reluctant to admit they WILL vote for the black candidate. I'm trying not to hope too high, but it could still be a landslide in the Electoral College.

____________________

Vicente Duque:

To turnout or not to turnout, that is the question.

I made a little table with the data from :

United States Election Project
George Mason University.
Dr. Michael McDonald

Last updated : Sunday, November 2, 2008

Totals of Early Voters ........ 2008 Early Vote / 2004 Total Vote

United States ......... 24,125,290 .......... 19.5%

New Mexico .............. 192,229 .......... 73.2%
Colorado ............... 1,477,836 .......... 68.8%
Nevada ................... 559,687 .......... 67.3%

Florida ................. 4,107,894 .......... 53.8%
North Carolina ......... 2,573,899 .......... 72.5%
Iowa ....................... 454,274 .......... 29.8%


Georgia ............... 1,994,990 .......... 60.1%

The Turnouts seem to be excellent, favoring Mr Obama, because of a better "Ground Operation". They can concentrate now on few "lazy people" or few "lazy counties" that they have on their computer screens.

http://tossUpStates.blogspot.com/

http://milenials.blogspot.com/


Vicente Duque

____________________

xcurmudgeon,

Thanks for the encouraging data on early voting results on your excellent website. Reading this really helped to increase my confidence that Obama will emerge victorious and perhaps even win in a land slide.

I am still looking for feedback on how the various polls are choosing their demographics. As you point out, the percentage of black voters is, at least in early voting, much higher than usual. This change, as well as what will hopefully be an increase in the youth vote, may not be sufficiently reflected in current polling.

Also these LV polls may not be properly factoring in the number of newly registered voters. What polls use the RV model as well as the LV one? I imagine the RV polls will show Obama up by a greater margin.

____________________

bmcfar01:

"I truly think he is willing to give up his presidency for his plan and ideals. Look at how long range and effective his planning has been. This guy is like a savant. "

1. No one who runs for President is willing to "give up" the office for his ideals. They just change their ideals. Obama is a politician--a great politian--but a politicians with all the instincts, disingenuity, compassion, and sagacity an any.

2. What closed the deal for Obama was the economy. Had there been no banking crisis, this would be a much closer race.

____________________

bmcfar01:

"I truly think he is willing to give up his presidency for his plan and ideals. Look at how long range and effective his planning has been. This guy is like a savant. "

1. No one who runs for President is willing to "give up" the office for his ideals. They just change their ideals. Obama is a politician--a great politian--but a politicians with all the instincts, disingenuity, compassion, and sagacity an any.

2. What closed the deal for Obama was the economy. Had there been no banking crisis, this would be a much closer race.

____________________

bmcfar01:

"I truly think he is willing to give up his presidency for his plan and ideals. Look at how long range and effective his planning has been. This guy is like a savant. "

No one who runs for President is willing to "give up" the office for his ideals. They change their ideals.

What closed the deal for Obama was the economy. Had there been no banking crisis, this would be a much closer race.

____________________

JoeyJams:

What is the significance in the difference of the solid Obama/Biden and the toss up numbers between RCP EC count & Pollster EC maps for 11-2-2008?

RealClearPolitics Electoral College

Obama/Biden 278 - 238 Solid 40 Leaning

McCain/Palin 132 - 118 Solid 14 Leaning

Toss Up 128

Pollster.com
Obama/Biden 311 - 264 Solid 47 Leaning

McCain/Palin 132 - 129 Solid 13 Leaning

Toss Up 85

____________________

NM:

@ Vicente Duque

I've seen other references to New Mexico's high early voter turnout, but 750,000 plus voted in 2004. So I'm not sure where that 73% figure comes from. I do not think the 192K number includes absentee voters, but still...

____________________

Aristotle:

@bmcfar01:

With respect I think your stance that a politician changes his views is both considered normal and expected. However, if that were the case How easy would it be for Barack simply to pour 100's of millions into P.A. and Ohio.

Thats all he needs. However, that is now what is happening. He is going places and spending money in places he has no need to be.

This would go against your point. My hope is that Cynicism my die a small death.

Peace

____________________

enadknock:

Vicente Duque:

check out the pew poll for detailed percentages of who the early birds voted for. pew reports that 32% of all voters have already voted and another 7% plan to before tuesday. its running 3-2 for obama. my old noggin calculates that mccain would have to win 56% of tuesday's votes to have a plurality (can't figure how 3rd party votes will go). let me know if i read the pew charts correctly.

____________________

Ham I Am:

The change in the Pennsylvania lines is such that they're actually sloping past vertical and going backward in time (more visible for both lines if you set sensitivity to "high", and reduce the date and % ranges to zoom in a little).

Is this representing a huge shift that we should look out for over the next day, or is it more a strange way that the data and the model are interacting that doesn't necessarily represent a huge shift?

____________________

Well, Mark, we're almost there. Many thanks for all of your work...tomorrow we'll see a last roundup of polls before the big day. I see that Gallup released its last poll: Obama 55%, McCain 44%. USA Today is similar. Now it's up to us.

____________________

jko:

Out of curiosity, how are the 27 million votes already cast accomodated (if at all) in the most recent polls? If there has been a disproportionate number of one party in the early voting, do the poll results then skew to the opposite party?

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR