Here are state-by-state results with (exit polls) and counted vote (Obama-McCain), and the difference on the Obama side between exits and official count:
Florida (52-49) 50-48 -2
Iowa (58-42) 53-44 -5
Missouri (52-48) 49-49 -3
North Carolina (52-48) 49-49 -3
New Hampshire (57-43) 56-44 -1
Nevada (55-45) 55-42 -0-
Pennsylvania (57-42) 54-44 -3
Ohio (54-45) 51-47 -3
Wisconsin (58-42) 56-42 -2
Indiana (52-48) 49-48 -3
New Mexico (56-43) 56-41 -0-
Minnesota (56-39) 54-43 -2
Michigan (60-39) 57-40 -3
Georgia (47-51) 46-52 -1
West Virginia (45-55) 42-55 -3
What to conclude? Too soon to know anything, but this is consistent with 2004 – the exit polls are consistently saying that the Democratic candidate is getting more votes than the final tallies say. In 2004, they tried to blame it on the “Reluctant Republican Respondent”, but that was disproven. There was no such effect.
In a perfect world, one would expect exit polls to fall within margin of error on either side of the ballot, with Republicans as often coming up short as Democrats. But it doesn’t work that way. Something is amiss.
Piece of Mindful 
Hey Mark, when was the last time an absentee voter was exit polled?
LikeLike
I have some stainless steel sheet metal screws if you need help fastening down that tinfoil hat of yours.
Gee, let’s understand the “quant” of statistics. A typical poll says something like there is 95% confidence interval that the margin of error is within 3%. Get it, Mark? In other words there’s a 5% chance that the margin could be greater than 3%. That’s not nothing and most of these polls were pretty close to the margin.
Or is this just the “fear” that you espouse rather than international bogymen?
The guy won. Enjoy your day and forget about the man behind the curtain. Sheesh, Moonbat!
LikeLike
Swede – it’s all factored in.
Budge – I know about MOE and 95% and all of that stuff. Also, exit polls are cluster samples, and so MOE is difficult to calculate. The remarkable feature here is that all of the polls are skewed the same way – So tell me what the odds are, genius, of the skew, even if within margin, all going one way – eh? Eh? I’m all ears.
Anyway, these polls are much closer to the vote count than 2004, when many many states fell outside the MOE. I’d have to go look it up. Apparently installation of paper ballot systems in places like Florida did some good, though there is still potential for malfeasance in the scanning equipment, as demonstrated time and again. Random auditing of results and follow-up on errors would fix a lot of this. As of right now, our voting system is not secure.
Lastly, Budge, you leave a bad feeling behind you – I’d appreciate your not coming around. Thanks. F/O.
LikeLike
Apparently, Dave’s a little pissed off. Perhaps he needs a little tin-foil chapeau for himself …
The exit polling isn’t exact; Dave is correct about that. But G’Damn, he sounds really pissed.
LikeLike
Do they figure in the honesty factor Mark? What makes a exit pollee tell the truth and nothing but the truth? Where’s the lie detectors and sodium pentathal. Are the pollers subject to background testing or questioned if their biased? Every exit poller I’ve ignored fits into the young and foolish dem demographic.
You of all people should know people bend the truth. You do taxes.
LikeLike
The interesting thing about delusional people is that a their favorite delusion becomes more complex over time. This is a natural response because reality never compares favorably with a delusion, but the delusional person will never give up his delusion, also. Thus the delusion must be continually modified or updated to explain for new reality-based evidence as it appears. This leads to more and more delusional complexity as time progresses, until, ultimately, the delusional person spends all his time altering and rearranging the “logic” of his delusion.
It is usually at this point, when the delusional person begins to spend all his time working on his delusion, his aberrant behavior is noticed by his family, coworkers, and friends, and professional intervention normally results.
Woo-woo!
Go for it, Mark “Landslide” Trotsky!
Gong. Gong. Gong.
LikeLike
Mark, I gotta to hand it to ya, these obsessions still give you purpose for living, not like these losers.
LikeLike
Swede – what you don’t get is that everything averages out. That’s why they interview large numbers of people. Large numbers of people aren’t going to lie. You’re questioning a well-accepted science that has been around a lot longer than you – cluster sampling is used around the world for many purposes, from determining the number of earthquake victims in Pakistan to US-caused deaths in Iraq to elections. Along comes Swede who suddenly figures out that people sometimes lie, and all of the statisticians who do this stuff slap their foreheads and say “Damn – we never thought of that!”
Rook – it’s easy to be you, to never question the truth as it is handed you. I’m willing to bet you voted for Bush twice and McCain once, and I didn’t even have to interview your ex-wife to figure that out.
LikeLike
Sorry Swede – Budge riled the beast and you caught it. The point is this: Exit polls are a long established scientific practice – the people who do it are very good at their trade. Exit polls a better reflection of elections than elections of exit polls.
It is true that the 2008 exit polls, from what I’ve seen, are looking much better than they did in 2004 (where a remarkable number of states fell outside MOE, and always in favor of Bush), but it is troublesome that the 2008 deviances are always on the Obama side. That should not happen – the results should fall in a scattered pattern reflecting a somewhat balanced skew between Bush and McCain. That tells me something – what I am not sure.
LikeLike
Mark, if you delete it’s your call. First, I think if you look at the history of both pre-election polls as well as exit polls they generally have tended to have a Dem bias. I have looked at this for years and it’s well documented.
Secondly, the infrastructure for exit polling changed completely after the 2000 election where the media used a common service for exit polling. This is no longer the case and there are now about three different exit polling organizations taking place.
Third, I’m not at all pissed-off – as Wulfgar would have me be. I didn’t have a dog in the race and don’t feel particularly compelled to think one candidate is better than the other. That, not withstanding my feeling about Congress and the lack of divided government.
Next, I’ll leave your comments section from now on but I find it really sanctimonious and hypocritical for all of the things you’ve said about people over the years. In the words of your son: “Your a dick!”
LikeLike
When we have these disagreements your response is, prove it. Well here’s your change to become credible, supply the link that shows the adjustment for early voters, and the honesty and biasis of the polees, pollers, and company which tabulates and broadcasts said results.
And while your at this please explain why, verified in your original post sample, it’s always eastern and midwestern states which show the greatest variance between polled and actual, when those states are reported first on election night.
LikeLike
There are some biases in exit polling that skew the results, the largest being that older people don’t participate, and most polling is done in the large urban areas. I’m guessing that would be large enough to give the 2% or so difference.
>>>>You’re questioning a well-accepted science that has been around a lot longer than you – cluster sampling is used around the world for many purposes, from determining the number of earthquake victims in Pakistan to US-caused deaths in Iraq to elections.
Things may be “well accepted” and “used around the world”, but they still have their problems. You have to choose the proper sample, and the results are only as good as the field work. The lab techs think they can account for all of this, but you know they can’t.
LikeLike
Dave – you’re welcome here and I value your comments. You just pissed me off with your shitty personal attack. But I’m better now.
All of you have it in your minds that elections are regulators of exit polls, and not the opposite. In fact, exit polls are meant to measure the accuracy of the vote count. If they are off, and persistently off in one direction, that points to a problem. Instead, people point to the exit polls as the problem. It’w wrong-headed.
Pre-election polls are not as accurate, but if they persistenly over-estimate Democratic vote, it might point to the success of Republican vote-suppression efforts. That’s been going on for decades. It’s a truism in politics that high turnout favors Democrats. That’s why the huge suppression effort this year, using the ACORN “scandal” as cover fire.
Fred – you’re right about proper samples and all of that – that’s basic. It’s what they do, and they constantly fine tune their work. To say that polls must be wrong because sampling is difficult is to hide your head in the sand.
LikeLike
>>>>To say that polls must be wrong because sampling is difficult is to hide your head in the sand.
You have to look at individual polls. Some polls are done very well and are highly accurate. Some are a joke. Contentious man-on-the-street type polls that have to capture info in a short period of time have problems. Young people are more open and likely to talk to a pollster (who is most often young and more likely to engage a young person) than an older person. To get the sample size, and for $$ reasons, the polls are almost always done in urban areas. The field people are given careful orders, but it is a messy affair. And then you have to make sense of the raw data while taking into account the limitations. The nice clean number at the end belies the messiness of it all. Don’t believe for a minute that they’ve got it figured out to eliminate a 2% error. Election administrators have the edge over pollsters.
>>>>Pre-election polls are not as accurate, but if they persistenly over-estimate Democratic vote, it might point to the success of Republican vote-suppression efforts.
You are too eager to believe conspiracy theories. The only successful Republican vote suppression effort was nominating John McCain. Fewer people voted in 2008 than 2004, and it wasn’t Democrats staying home.
LikeLike
Don’t worry Fred, we can trust’em. From wiki.
A possibly unwritten secondary mission of the Voter News Service was to provide election results as quickly as possible on election night—a point which came to haunt the VNS in the 2000 Presidential election.
The VNS received intense criticism for its ‘flip-flop’ calling of the state of Florida in that election[citation needed]. During the course of the evening, it first called the closely contested state of Florida for Al Gore, then George W. Bush, and then as ‘too close to call’. Critics argued that the state should never have been called until the state’s fate was clear. The Voter News Service also received specific criticism for calling the state of Florida for Al Gore before the polls closed in the Florida panhandle, which was located in the Central time zone and heavily Republican.
In 2002, the VNS intended to make calls in the November U.S. Congressional and Senate elections. It attempted to use a computer designed for VNS by an outside contractor to do this. A system failure occurred in this computer on election night, making quick delivery of data impossible. In fact, collecting and delivering the data took ten months[1].
In January 2003, the Voter News Service was disbanded largely because of failures in 2000 and 2002. Murray Edelman, VNS editorial director, criticized the decision as making the VNS a scapegoat. [2]
In the 2004 presidential election, a new organization called National Election Pool was set up by the same organizations, utilizing consultants Edison/Mitofsky for exit polling and Associated Press for official returns. However, the NEP had controversies of its own for 2004 when it released exit polling data early that was significantly different than the final results.
LikeLike
Fred – you’re stubbornly refusing to consider the possibility that people manipulate the vote, either by suppression or miscounting. Your arguments are the equivalent of a monkey with his hands over everything but his mouth. But explain to me why the Republican party, for years under court order not to do voter caging, still does voter caging, challenges, voting list purges and other dirty tricks? They know, as you probably do too, that the more people who turn out, the more likely they are to lose. Democrats, on the other hand, know that bigger turnouts favor them, and engage in energetic voter registration drives, which were this year demonized (even criminalized) by the Republicans. Them’s the hard nuts.
Swede – I think it rather humorous that people criticized pollsters for getting Florida wrong in 2002 when they got it right – Gore won Florida. And there has been increasing pressure since 2000 to keep exit polls secret – 2004 being the reason why. They shine a light on the process, a little disinfecting sunlight. But our leaders want us to believe our elections are squeaky clean when they are far from that. Exit polls are a measure of election honesty, and along with spot audits and random recounts of paper ballots, are a tool to ensure honesty. The fact that they are disparaged is a sign that the elections are troubled.
And both of you stop and think – the software used to count votes is not under control of election supervisors and cannot be reviewed. It is controlled by private corporations. How do you spell A-B-S-U-R-D? For so long as that situation exists, our elections are not secure.
LikeLike
There is almost too much here to debunk.
>>>>the software used to count votes is not under control of election supervisors and cannot be reviewed.
It is reviewed by election supervisors.
>>>>Gore won Florida
He did not win Florida. Bush got more votes.
>>>>And there has been increasing pressure since 2000 to keep exit polls secret
Pressure to keep them secret during the election. Not afterward.
>>>>But our leaders want us to believe our elections are squeaky clean when they are far from that.
They are pretty clean, and such “dirt” is not going to be found by the current state of exit polling. It is easier to game exit polls than actual elections. You worship the wrong gods.
>>>>But explain to me why the Republican party, for years under court order not to do voter caging, still does voter caging, challenges, voting list purges and other dirty tricks?
Ahh, these aren’t dirty tricks. Evidently Democrats can’t stand any scrutiny. I wonder why. Don’t tell me Democrats don’t pull some shady deals. I don’t like busing people to the polls. Would you call that a dirty practice?
>>>>Democrats, on the other hand, know that bigger turnouts favor them
Not necessarily. Democrats target certain demographic groups. Call me when they conduct registration drives and get out the vote drives equally everywhere.
LikeLike
Software is not reviewed by election officials. It is proprietary. You obviously haven’t been following this controversy. That’s at the heart of it – companies say that software is a trade secret. No one gets to see it. No one. Get up to speed.
Gore won Florida. That was the conclusion of the major press investigation, but it was quashed because of 9/11.
There is controversy surrounding release of raw exit poll data, unlike in other countries like German, where government does exit polling to make sure that votes are correctly counted. We should too.
Paul Weyrich: “Now many of our Christians have what I call the goo-goo syndrome — good government. They want everybody to vote. I don’t want everybody to vote. Elections are not won by a majority of people, they never have been from the beginning of our country and they are not now. As a matter of fact, our leverage in the elections quite candidly goes up as the voting populace goes down.”
That’s from an early leader of the conservative resurgence. It’s just a fact of life – Democrats try to register voters, Republicans try to stop them from voting. That’s why long lines in Democratic districts – they short them voting machines.
LikeLike
>>>>Software is not reviewed by election officials.
It’s true that officials probably don’t pore over lines of code, but they can test the machines and their results are often checked against paper ballots. Some manufacturers make their source code available. Diebold is the famous exception, mainly because they use a variant of Windows. Local officials are the ones who keep elections fair. They make the call as to which machine to buy, if to buy machines at all. Some precincts have gone back to paper ballots, at great cost. Think back to the days of hand counting: there was a time for miscounts. Technology has made our elections fairer, crabs like you notwithstanding. link
>>>>Gore won Florida.
Gee, the initial vote count had Bush ahead by over 2000 votes. The final certified tally, after a recount, had Bush winning by 537 votes. What is your problem? Probably this (from Wiki): “Following the election, recounts conducted by various United States news media organizations indicated that Bush would have won if certain recounting methods had been used (including the one favored by Gore at the time of the Supreme Court decision) but that Gore might have won under other scenarios.[19]” You must be from that group that wants us to sit around recounting ballots until the Democrats win. I’m reminded of when a media outlet sends the same financials to several different tax preparers, and they all come up with a different tax liability.
>>>>There is controversy surrounding release of raw exit poll data
A group “of the usual suspects” of media groups (you know, the unbiased folks at AP, NBC, ABC) commission the polls. I heard you can buy the data. Normal folks prefer to put their efforts into good election results, not worrying about sketchy exit polls. The raw data won’t add much to the public debate.
>>>>Paul Weyrich…our leverage in the elections quite candidly goes up as the voting populace goes down.”
Maybe, but maybe not. Weyrich isn’t necessarily correct. I think it is best if informed and engaged people get themselves to the polls, not be led there by activists gaming the system by herding in their favorite demographic.
>>>>That’s why long lines in Democratic districts – they short them voting machines.
“They”? Who’s this “they”? Why don’t “they” Democrats get some more machines?
LikeLike
Trust and don’t bother verifying. Is this your new ethos? The word gullible is not in the dictionary. Exit polls are a means of verification. They go haywire, and you tell me that the exit polls are the problem. That’s nonsense.
There’s so much going on here that you’ve not followed – a Wiki check satisfies you. Electronic voting systems are not secure, per se, because every district needs a software technician to review the code, which they cannot do because it is proprietary anyway. Now you’re telling me that many machines in use use open source software. That’s not true. You’re wrong.
The problem with electronic counting is that there is no tie-in between the voter and the vote, and no way of verifying that each voter meant to vote the way it is recorded. That’s the nature of the beast. Computers cannot solve that problem. The only way around that problem is a paper record – not just a paper record, but paper documents being the official record of the election.
See the movie Hacking Democracy sometime for a demonstration of how easy it is to compromise optical scanners. Even more so for touch screens, which are “faith based” anyway. Pre-election dry runs did not solve the problem – it doesn’t take a genius to get around that test. With election administrators watching, the machine flipped votes and gave the wrong winner. The thing was, it was so easy. The machines are eminently hackable. That’s been the problem from the beginning – the machines are crappy – Stephen Spoonamore, software engineer who is involved in a lawsuit against the State of Ohio concerning the 2004 election, says that the machines are so hackable that it is almost as if they are designed to be hackable.
Then you have proprietary code – all of them have it. It’s the standard. Montana uses e/s machines with proprietary code, like counting votes is some Star Wars fantastic achievement, and not BASIC 101 code. It’s ludicrous.
Here’s how you do an election: Paper ballots, electronic scanning for compilation, reliance on exit polls to spot problems, and random auditing of results – that is, random districts are hand counted to see if tallies match, if they don’t more tests are done – eventually, the whole election is recounted if there are enough problems. Right now exit polls are telling us there is a problem. It’s as plain as the nose on my face. People like you, head in sand, assume that the exit polls are wrong.
Gore won Florida in the following scenario: Statewide recount, which was underway and stopped by the gang of nine. Bush won only if recount was severely limited to a few districts. This is all old news.
LikeLike
>>>>Gore won Florida in the following scenario: Statewide recount, which was underway and stopped by the gang of nine.
Maybe, maybe not. Wisdom is for the Court not to intervene. If they involve themselves and the count changes by some sketchy recount protocol the other side doesn’t accept, you’ve got big problems. The four left-wing Judges who voted for a recount were irresponsible.
>>>>Here’s how you do an election
I like paper ballots, scanning, and auditing. Exit polling can be useful, but it really is too expensive to do on a widespread basis. I haven’t been involved in exit polling, but I’ve been behind the scenes with other polls, and it’s hard to get good field data. You need to sample the same population several times. A one time shot like elections has big time field data problems. I’m not unsympathetic to your view on exit polls, but I’m surprised you put so much stock in them.
>>>>Now you’re telling me that many machines in use use open source software.
I didn’t say they were open source, I said some companies will release their code to election officials.
>>>>The problem with electronic counting is that there is no tie-in between the voter and the vote…The machines are eminently hackable.
I have more faith in the system than you do. Just because something is hackable doesn’t mean it’s always going to be hacked. At some point you have to have trust. If there are a bunch of crooks running our elections, they will find a way to spoil the results eventually. Most people, enough people, are fair.
LikeLike
I came across a discussion of exit polls by a number cruncher: here
LikeLike
My above link was to the comments section. You have to punch the “show entire article” link.
LikeLike
Trust, but verify.
I am an accountant – the dreary profession. Here is what I know – if a system has flaws, people will exploit those flaws. Most people are as honest as they have to be, but enough aren’t so that you always have to be vigilant. If you give them a way, they will use it.
LikeLike
That will be the umpteenth article I’ve read that said that election results invalidated the exit polls. I am only suggesting the opposite.
LikeLike
>>>>Trust, but verify…you always have to be vigilant.
I agree. We also have to be vigilant about exit polls. They really aren’t that good. Go work on one sometime.
LikeLike