Homepage | Photoblog | Weatherblog | Photos | Old blog archives

About me

I'm Brendan Loy, a 26-year-old graduate of USC and Notre Dame now living and working in Knoxville, Tennessee. My wife Becky and I are brand-new parents of a beautiful baby girl, born on New Year's Eve.

I'm a big-time sports fan, a politics, media & law junkie, an astronomy buff, a weather nerd, an Apple aficionado, a Lord of the Rings and Harry Potter fanatic, and an all-around dork. My blog is best-known for its coverage of Hurricane Katrina, but I blog about anything and everything that interests me.

You can contact me at irishtrojan [at], or donate to my "tip jar" by clicking the link below:

June 2008

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          
Pajamas Media BlogRoll Member

« Happy Earth Day! | Main | UFOs attack Arizona, Florida »

Don't trust the unweighted exit polls!

[UPDATE, 8:12 PM: Welcome, new readers! The scenario I predicted in this post is now happening. Some leaked exit poll numbers show Obama winning, others Clinton leading by a slight margin. Regardless, it's all meaningless. Likewise, CNN's statement that the race is "competitive based on the exit polls" should be not taken as predictive of the final outcome. CNN initially said Ohio was "competitive," too.

Anyway, I'm liveblogging the results on my homepage.]

Following up on a point I made yesterday -- in a post that just got Instalanched -- as we political junkies giddily await the first hints of what's happening in Pennsylvania today (finally, another primary!! election results tonight!! exit polls!!! colorful maps!!! whee!!!!), it's very important for us, and even moreso the media, to remember that Obama almost always does well in the leaked, unweighted exit polls, and almost always does less well in the final results. For instance:

  • In New Hampshire, the "first wave" showed Obama leading by 4; he lost by 3.

  • On Super Tuesday, the "first wave" had Obama winning New Mexico by 6 (he lost by 1) and losing California by just 3 (actual margin: 8). The "second wave" wasn't any more accurate: it had Obama winning Georgia by 50 (actual margin: 35), Illinois by 40 (actual margin: 32), Alabama by 22 (actual margin: 14), Delaware by 14 (actual margin: 9), Connecticut by 7 (actual margin: 4), Arizona by 6 (lost by 8), New Jersey by 5 (lost by 10), Missouri by 4 (actual margin: 1), Massachusetts by 2 (lost by 15), and losing Tennessee by 10 (actual margin: 13), New York by 14 (actual margin: 17), Oklahoma by 30 (actual margin: 23) and Arkansas by 46 (actual margin: 44). So Oklahoma and Arkansas were the only states on Super Tuesday where Obama did better than the leaked, unweighted exit polls suggested.

  • On March 4, the "second wave" showed Obama winning Vermont by 34 (actual margin: 21), Texas by 2 (lost by 4), Ohio by 2 (lost by 10), and tied in Rhode Island (lost by 18).

Averaging all those numbers together -- and I recognize that this is very unscientific -- you get an average discrepancy of 7 to 8 points. That is to say, Obama generally does 7-8 points worse in the actual results than he did in the leaked, unweighted exit polls.

This is crucially important, because it has the potential to significantly affect the post-primary "spin." That's exactly what happened on Super Tuesday, when Clinton was able to initially spin a "victory" out of her lukewarm performance, largely because the media was expecting Obama to win some "big states" based on those early, favorable numbers. Likewise on March 4, Hillary was able to claim "success" for her Texas and Ohio "firewall," even though she really needed much larger margins to make meaningful delegate progress, in part because the leaked exit polls again conned the punditry into expecting better showings by Obama, possibly including a win in one or both states.

It'll be a travesty and a farce if that happens again. Hillary Clinton needs to win big -- like, double digits big -- and make significant delegate gains in order to claim any kind of a meaningful victory in Pennsylvania tonight. That basic fact will not change one iota if Drudge and The Corner and Talking Points Memo and the Huffington Post (and, er, the Irish Trojan) publish initial, meaningless numbers this afternoon that show a "OMG A DEAD HEAT IN PENNSYLVANIA!!1!" and then Hillary "pulls away" and wins by 6 or 8 points or whatever.

It's important to remember that these leaked exit polls do not actually represent any version of reality; they are not something that a candidate can "come back" from. In those instances where they differ from the actual numbers, they are, and always were, simply wrong. The exit-poll-fueled "seven-hour presidency of John Kerry," for instance, was always an illusion; Kerry was never "ahead," and Bush never mounted a "comeback." That's all pure perception, and has nothing whatsoever to do with reality. Same thing here. If Hillary, yet again, does substantially better than the exit polls suggest, nobody should be surprised, and still less should she get favorable, "expectations"-based spin as a result. Obama's early "lead," in that event, will have been (again) a complete chimera. So please, for heaven's sake, let's not get all excited if history repeats itself again.

As I said yesterday: Fool me once, shame on you. Fool me 21 times, shame on me.

P.S. Politico adds an additional cautionary note, referring to actual returns rather than exit polls:

Don’t be fooled by early results. The cities and suburbs usually report their returns first, which gives the candidate favored in those areas a quick – and sometimes fleeting – lead. The conservative-leaning small towns through the center of the state usually filter in much later in the evening.

This tendency has wreaked havoc in past elections: A Democrat goes to bed thinking he or she is the winner, but wakes up the loser. The last time it happened was 2004, when the Associated Press called the state attorney general race for Democrat James Eisenhower and retracted it later in the night as the numbers closed. Some newspapers went to print with the wrong results.

So Obama could show a lead in the early results, but it might be short-lived. If Clinton is ahead at the start, she may never lose it.


TrackBack URL for this entry:

Listed below are links to weblogs that reference Don't trust the unweighted exit polls!:


Ok, Obama does well in unweighted exit polls. My question is: why?

I'm tempted to cite the Bradley Effect, but I've been scolded for that before. I honestly don't know. But, even accepting the premise that unweighted exit polls are inherently unreliable, the overwhelming pattern of unreliability in one particular direction certainly does cry out for some sort of explanation.

does politico really know what he is talking about? missouri was the exact opposite of what he is claiming. clinton got ahead early but the cities got their returns in late and obama won late night. the same deal happened a few times with mccain/huckabee with huckabee getting early leads and mccain coming back once the urban votes came in. am i crazy?

also, my prediction for today
clinton 54%
obama 45%

"...certainly does cry out for some sort of explanation."

Well isn't it Obvious? FRAUUUD! Massive rigging of the "real" "voting" "results" by Rove/Diebold/Halliburton/Penn/Ickes, to
steal it for Hillary. Flor-i-DUH! O-hi-YO! Zim-ba-BWEY!!!


(Good post, Brendan.)

This probably depends on the state. Some cities seem to take forever to get their results in, while the rural areas report right away. Pennsylvania has traditionally been the opposite way - Pittsburgh and Philly have their votes counted very QUICKLY, and the rural vote gets in really late in the evening. Thus, Obama is likely to jump to the early lead in my former home state.

It will become known as "The Obama Effect". More people will claim to vote for him than actually do. I predict this will be around 10-12% in a national election and will be studied for years.

What a lucky guy you are to have a billion hits from Josh.

What a lucky guy you are to have a billion hits from Josh.

This is an instance where the Politico is right: the cities and suburbs in PA (as opposed to MO, where the situation is different) are usually first in.

You are Irish, 26years and with a degree from USC and Norte Dame! Ha!

Given that the exits appear skewed in one direction, one has to wonder about racial bias. After all, someone who votes against Obama based on race isn't likely to admit it in an exit poll (which are typically conducted face-to-face). On the other hand, it seems odd that someone who voted against Obama--for any reason--would turn around and say they voted for him, which is what would need to happen to create the kind of discrepancy highlighted here.

Given that the exits appear skewed in one direction, one has to wonder about racial bias...

I believe Mystery Pollster at has talked about exit polls skewing towards younger voters -- given Obama's strength in this demographic (and HRC's strength with seniors), it's not surprising that he reads better in exit polls than in the final results.


It's likely the Bradley Effect.
It's more pronounced in FACE TO FACE polling.

I call it the Voter Disparity Theory (VDT).


Congratulations on a fantastic post.

My own (entirely unsubstantiated) guess is that we're witnessing an enthusiasm gap. Obama's supporters, who skew younger, more urban, and more affluent, are more likely to line up at the polls at 7am waiting for the chance to cast their ballots (with the notable exception of black voters, who tend to turn out quite late in the day). Bear in mind that first-wave numbers are taken through 9am; second-wave numbers include results through 3pm; and third wave numbers are added up after the polls close. Since the early morning rush still dominates the total turnout until after 5pm, we're looking at a different demographic subset. They have models to account for this, of course, but I think your data argue fairly compellingly that those models are out of whack in this unusual cycle.

One reason that unweighted exit polls are misleading: self-selection by those who participate. No doubt that Obama-cans are seeking out the exit pollsters to share their joy over voting for their candidate.

The exit poll refusal rates at U.S. polling places are very high compared to other countries. In the 2004 Presidential election, Republican voters refused to participate in exit polling in greater numbers than Democrats; most likely another manifestation of the Shy Tory effect, as some GOP voters may have regarded the exit polls as an extension of the “liberal media.” In fact, the predictive shakiness of the 2004 and 2006 national exit polls has been linked to problems with voter sampling.

I have written more on the problems of opinion polling for campaign 2008 here:

Jefferson Flanders

Intesting post. We'll be watching tonight.

I have to ask, though, how do you handle the big game?

Is it FIGHT ON! or Fighting Irish?

Must be tough!

So why is it racist when a few Whites vote based on race, but when 90% of Blacks vote based on race its just "ethnic identification."


"Ethnic identification" - interesting term I've not heard one person of color utter yet. So you contend 1 in 5 or 20% whites voting in Pennsylvania is just a "few"?? AA are not voting for Obama based so on race, and it's beyond insulting for you and MSM to continue to make such a case.

I'm with Irish Trojan on this one. Never trust ANY exit poll ANYWHERE. We all remember the networks calling Florida for Gore by 6% before the polls closed in the panhandle, and reality was Bush by a smidgen. Exit polls also predicted Kerry winning Virginia in 2004, reality was Bush by 8%.

The simple reality is that exit pollsters can't go everywhere at once, and it's hard to know whether precincts selected for exit polls really represent an entire state. Even at a given precinct, voters with little else to do will be more inclined to talk to exit pollsters than a working person in a hurry to vote, then rush off to work, or a mother in a hurry to take her kids back from a baby-sitter. Exit polls also completely miss the trends of early or absentee voters, who were 10% of the electorate in Florida in 2000.

Exit polls can be useful after the official results are in for analyzing how certain groups of voters (by sex, race, income level, etc.) voted, but cannot be used to predict election results. Forget early curiosity--we'll see who won after all the votes are counted, including people who never met an exit pollster.

Actually....the only thing that really matters in the end is the number of delegates. Expecially when your candidate is 140 plus behind ... exit polls and political spin are like sports statistics....they only seem to matter to the loser......

Larry is wrong on this. The exit polls were right in 2004- it was the machines, the Ken Blackwell fake felon lists, the dirty tricks that changed the official results in Ohio. If the media (owned by several large corporations) was really "liberal" and biased towards Kerry, we wouldn't have had an election season where nothing good was ever said about Kerry and nothing bad ever said about Bush.

People should read more of Greg Palast's work on the GOP election fraud, which the corporate media doesn't cover. Yes we know many reporters are registered Dem, bla bla bla, or whatever argument GOPers will retort with. Conservatives actually own the media and the editors above the reporters are statistically conservative- reporters know who feeds them.

I think it's pretty simple; liberal voters are more enthusiastic about politics and government, so they're more likely to give pollsters (who often look and sound "like them") their choices for office.

Also, liberal candidates are usually more socially chic, so their voters are less reluctant to have a conversation about supporting them in public.

In remember with a snicker the iteration of this theory back in 2004, when Democrats spent nine months trotting out every celebrity, every entertainment channel, every talk show gaggle, every everyone to endorse Kerry and belittle Bush voters...and then were enraged - enraged - when Kerry's exit poll numbers bore no resemblance to his actual performance.

Tom forgot the black helicopters and the UFO involvement in the Ohio results.

PS - It's called the Bradley Effect.

Obama has the most fervent supporters that will make time to attend rowdy caucus, donate lots of money and/or drink cyanide laced beverages for him.

Too bad the general election isn’t a caucus, eh?

Competitive to CNN means they know who is going to win but they do not want to rush to judgment. If the race was truely close they would say it was "too close to call."

Hillary will win tonight. Barack will win both on May 6th and that will be the end.

Great analysis, Irish guy. I think the Pope's visit will spike Catholic participation and that will seal a Hillary win of around 10 points.

Exit pollster employees are young and click with young voters not elderly curmudgeons so that's where some bias comes in.

Could the exit polling discrepancy simply be a result of the fact that older voters (Hillary's base) are not so comfortable with telling anyone, let alone strangers, who they voted for (not to mention how much money they make, how often they go to church, how old they are, et cetera) and therefore successfully avoid exit poll workers? On the other hand, young voters (Obama's base) are more likely to show up with an Obama poster, scream OOOBAAAAMAAAAA as they pull the lever and seek out exit poll workers as they leave the building.

I exit polled California for the consortium. I would say that Hillary voters, as blue collar workers, are more likely to be rushed and decline the survey. You have the old fashioned voters that refuse to disclose who they voted for, believing that their secret ballot is sacred (or that the networks screwed it up in 2000), the "Operation Chaos" crowd which doesn't want to admit what they did, and moms who have to juggle their kids and need to take care of them. That is my guess for the skew.

"AA are not voting for Obama based so on race, and it's beyond insulting for you and MSM to continue to make such a case."

That's why exit polls are showing that 92% of blacks voted for Obama. Please. Explain those results in any other way.

What a stupid blog. How did this get on thi site?

Obama was never ever going to win Penn. White females have HRC's back bigtime in Penn.

Exit polls never had Obama ahead. This is total Bullcrap

What a stupid blog. How did this get on thi site?

Obama was never ever going to win Penn. White females have HRC's back bigtime in Penn.

Exit polls never had Obama ahead. This is total Bullcrap

Chris Matthews, who is from Penn., says that traditionally Philadelphia, the Obama stronghold, is a very SLOW count and will be late coming in. This has me wondering if his gain on HRC will strengthen even more later in the evening.

it is no secret that obama is a stronger candidate among educated middle class population segments. As the artcle denotes -- exit polls pull from cities and suburbs first -- typical enclaves for this particular pop. Segment. That said it makes perfect sense that obama would have a perceived lead. The same logic follows for Kerry as well or any other candidate that has not developed a populist platform...good article.

Frankly, I wonder sometimes if there aren't some bad apples in the media who deliberately don't ask a representative sample and/or don't go to a polling station that is representative of the general public. If the reporter is out there alone, who's to know?

The exit polls were always considered the gold standard until companies like Dielbold, Es&S, etc. took over the voting counting. As Stalin said "Those who cast the votes decide nothing. Those who count the votes decide everything." Why in Europe can exit polls be almost exactly in tune with final results, while since the takeover of the counting by right wing companies in the US they now have to be "weighted" (adjusted to accord with the fixed final "vote"). Read or Mark Crispin Miller or Greg Palast to see how our democracy is being malipulated by right wing companies (why are private companies even counting our votes?!). The "liberal media" nonsense is an added absurdity, given that the Clear Channels and Murdoch, Moonie (Washington) Times and almost all of talk radio are in the hands of right wing idealogues. Even Repubs should be concerned about the theft of our (now) deMOCKery and get rid of all the machines and insist upon HAND COUNTED PAPER BALLOTS, counted by precincts in public view. The restoration of democracy in the US is at stake.

My feeling is that exit polls are skewed according to personality/social norm bias. That is, the people most likely to respond to exit pollsters (rather than being annoyed or even offended by the intrusion) are most likely to be socially gregarious, younger, hip, socially confident, and socially conscious (as in peer pressure) voters. These personalities will be dispropotionately attracted to particular candidates in an election, notably ones that have an aura of also being socially gregarious, young, and hip. Likewise, since the pollsters themselves are likely to fit this description, the exit polls tend to reflect more social networking than actual voter practice. Older or more introverted or more reticent, dare I say more 'conservative', voters simply don't respond to pollsters as often, and are more likely to be annoyed that anyone would even ask about what they consider fundamentally a private (even sacred) matter.

Kenfolk: Err.. because none of that is true?

"The exit polls were always considered the gold standard until companies like Dielbold, Es&S, etc. took over the voting counting."

Exit polls have been consistantly off since forever. Go back to any election and you'll find that the exit polls are wrong more often than they are right.

"Why in Europe can exit polls be almost exactly in tune with final results..."

Because they aren't. Take almost any recent European election. Taking the most recent I'm aware of, exit polls had Berlusconi winning by 2 points, but in the final tally he ended up winning by 6.

This whole meme about how accurate the exit polls are got started after the 2000 election and has been talked up by liberals with no regard for the facts ever since. You've got your cause and effect reversed. Before there were conspiracy theories, everyone expected the exit polls to be off by some fairly significant amount - because every statistician knows that they aren't a true random sample and that the pollsters actually taking the data are a bunch of bored college students. The 'gold standard' myth is a creation required to sustain your conspiracy theory.

Incidently, I was opposed to using electronic ballots back when the Democrats clamored for them in the wake of the 2000 elections, and I'm still opposed to them. But incidently, I'm also opposed to hand counted ballots because it introduces so much human error (and increased risk of fraud). Machine tallied paper ballots using setups similar to those that score multiple choice tests are absolutely the securest and least error prone method because you have both a difficult to tamper with ballot and reproducability.

Has anyone looked at the actual results for a district compared with what the exit polling said they would be?

E.g. is there any chance that people are 'lying' to the pollsters and saying they've voted for Obama when they actually voted for Clinton?

For people who do not want to be accused of being 'bitter' they may just say, "I voted for O" when they actually picked Clinton.


"In Greek mythology, Cassandra was a daughter of King Priam and Queen Hecuba of Troy. Her beauty caused Apollo to grant her the gift of prophecy (or, more correctly, prescience). However, when she did not return his love, Apollo placed a curse on her so that no one would ever believe her predictions" -- Wikipedia on "Cassandra"

Gotta love Brendan Loy, 26 yr old graduate of USC and Notre Dame and self-professed big-time sports fan. In his blog at, he's done a great job worthy of the best baseball statistics nut of bringing together leaked exit poll information from the 2008 Democratic Primary showing that in New Hampshire, New Mexico, California, Georgia, Illinois, Alabama-- the list goes on and on-- "Obama generally does 7-8 point worse in the actual results than the leaked, unweighted exit polls suggested." He concludes that "these leaked exit polls do not actually represent any version of reality". Brendan says "The exit-poll-fueled 'seven-hour presidency of John Kerry', for instance, was always an illusion; Kerry was never 'ahead,'... that's all pure perception, and has nothing whatsoever to do with reality."

So, we should all relax because the nation-wide 7-8% pro-Clinton/anti-Obama shift in actual results compared to exit polls is nothing more than what happened in November 2004 in the Bush vs. Kerry run for the presidency.

Whew! I'm so relieved.

We can now settle back and argue about why 7-8% of voters in exit polls nationwide vote for Clinton, but lie to the exit-pollers. Pundits can openly speculate if it's because they don't want to be perceived as racist by a complete stranger and they aren't worried about being perceived as sexist. All kinds of rationale can be put forth, never to be disproved because, as with the 2004 Mitofsky-Edison exit poll data, the raw data will never be released for the public to analyze.

Eventually, they'll plug all the "leaks"-- people on the inside of the big media consortium that owns U.S. exit polls who release the "unadjusted" and "unweighted" poll numbers without permission. Then we won't have to worry about that crazy, lying "Exipolla" who can never seem to accurately predict who the votes will be counted for.

Never mind that exit polls are routinely used to detect rigged elections world-wide except for in the U.S. and we are treated to ironies like the Ukraine in 2004 when the U.S. mainstream media pushed for a runoff election on the basis of exit polls even while the exit polls in the U.S. showed Kerry should have gone to the White House.

Never mind that exit polls in Germany, where paper ballots are counted by hand, are generally never off by more than a fraction of one percent. (By the way, where paper ballots were hand-counted in the U.S. in 2004, the exit poll discrepancy was 1% or less, but ranged from 5% to 10% for other techologies including touch-screen, punch cards, optical scan and mechanical, see www.TruthInVoting.Org's home page)

Never mind that the pundit-spin stories like the "reluctant Bush responder" have been debunked by the little data that has managed to leak out.

Never mind that Edison-Mitofsky maintains raw exit poll data as a corporate secret, unavailable to researchers who what to study the true causes of exit poll discrepancies.

Never mind that the "weighting" and "adjustment" procedures aren't published, but essentially just modify the final, published exit poll figures to match the official results.

Never mind that the exit polls always seem to be wrong in the same direction.

Irony aside, however, we are truly fortunate for people who can be in possession of the facts and share those facts openly because they're as yet unable to see in that data how corrupt our elections sytems are. It's like, "Wow! 7-8% nationwide? Isn't that a *funny* thing?". If they could believe the facts in their hands and understand what kind of power must be behind a nationwide 7-8% shift in the official election results, they'd probably hesitate to speak up at all. Brendan, you're a courageous hero, even if you're only beginning to understand why. Please don't mistake this for condescension-- I'd say the same to my younger self-- I was much older than you are before I was able to face up to what the data mean. I didn't really see how bad our election system has gotten until 2005. God bless you, Mr. Loy. Keep asking questions.

I've said it before and I'll say it again, there should ALWAYS be hand countable paper ballots. I'd prefer an opening counting method that cost a little mroe money than relying on electronic machine tallying which is so incredibly open to single points of failure/hacking

Well Obama may agree with me, but the last comment was mine, name left over from the puppy thread

Well of course the name was left over. Once you go Barack, you never go back.

Earlier this spring when I voted in my state's primary, we had exit polling going on. I was very excited to get to share my opinion on the election. I noticed that not one single man that came out of the polls declined to share his opinion. What was striking was the number of older women who glared at the pollsters and walked on. "that's none of your business, young lady.".
If this happens in other states as well as mine, I say it certainly can explain errors in the exit polling.

Yes, weevil, we can speculate about anecdotal experiences all we want to, but we'll never get the data to test or confirm these ideas unless another party conducts large scale exit polling in the U.S.

But why would older women voters be more resistant to exit polls in the U.S. than in other countries (except when the U.S. women happen to use paper ballots that are hand-counted, of course)?

Brendan--You are making a critical error by referring to the early EPs as "unweighted." Weighting is applied by Edison Mitofsky to all EP data, including the early Call 1 and Call 2 data. It is stratified to their best-guess demographics, and in that sense is no different from the Call 3 data that theoretically appears just after poll closing. The difference is that from that point forward (and possibly including even the first media-posted EP results), the EPs are "adjusted" toward conformity with the "actual" votecounts coming in (and therefore become progressively more meaningless as a verification mechanism as successive adjustments bring about closer conformity). It is, contrary to your assertion, enormously significant that there is this persistent anti-Obama discrepancy between not only the early (but still weighted) EPs but the initial publicly posted EPs on the one hand, and the supposedly "actual" votecounts on the other. Anyone at this point who simply assumes that the votecounts are correct and measures EP "inaccuracy" thereby has drunk the koolaid served up by the MSM in service of the right wing takeover of America. We know for a fact that the machines are now virtually all black boxes, purveyed to us by corporations with avowed right-wing penchants, manipulated with the greatest of ease, and with no effective verification process. It is truly faith-based voting, with the EPs being perhaps the only "yeah, but" left standing. Please, in spite of your hard work and apparent good intentions, study up a bit more on EP-ology, before adding to the misdirection by apparently inadvertently posting inaccurate information about the EP data and its significance. I'd be happy to discuss with you at whatever length necessary.--Jonathan

Thank you Jonathan and a big "what-he-said". Thank you, especially, for your work saving the early 2004 Exit Poll data off the web on election night. Without your foresight, we would know even less about the unweighted exit poll numbers from that debacle. You are a true American Hero.

Honest: I'm not a tinfoil hat brigader but I notice that you say don't trust the unweighted polls because the "actual" results may vary later. Well, sure, but remember in 2004 when the polls showed Bush LOSING they began "weighting" them with "actual" numbers. What kind of methodology allows that? Absurd.

Remember too that in 3rd World countries, when the exit polls vary with the announced results, vote fraud is the first suspect.

And finally know that 85 percent of the Penn vote was done on electronic voting machines with no paper trail not way to audit or recount.

To believe that Bush won in 2004 you must also believe that…

Updated March 2, 2008 by TruthIsAll

1- The media exhaustively analyzed state and national pre-election /exit poll data and the documented evidence of vote suppression and miscounts. But raw exit poll precinct data has never been made public.

2- There are many explanations as to why the exit polls were wrong: Kerry voters were more approachable than Bush voters to be interviewed; interviewers sought out Kerry voters; returning Gore voters claimed to have voted for Bush in 2000 because they wanted to be associated with the “winner”; exit polls are not random samples; U.S. exit polls are not designed to monitor election fraud; early exit polling overstated the Kerry vote (women voted early); Republicans voted late; Gore voters defected to Bush at twice the rate that Bush voters defected to Kerry, etc. But none of these “explanations” are supported by the evidence. In fact, they have all been refuted.

3- The votes were fairly counted. But 2004 U.S. Census data indicates that 125.7 million votes were cast as opposed to the 122.3m recorded. And investigative reporter Greg Palast presented government data which documented 3 million uncounted votes.

4- Democrats failed to attract first-time voters. But according to the National Exit Poll, since 1992 the Democrats have won first-time voters by a 14% average margin. Ruy Teixeira wrote about it in The Emerging Democratic Majority.

5- Bush’s 48% Election Day approval rating was not a major factor. But since 1976 all presidential incumbents with less than 50% approval lost re-election (Ford, Carter, Bush I) while those above 50% won (Eisenhower, Johnson, Nixon, Reagan, Clinton). And there was a near-perfect 0.87 correlation between Bush’s monthly approval rating and the average national poll. The correlation was confirmed when Kerry won the 12:22am National Exit Poll by 51-48%. According to Frank Newport, Editor in Chief of the Gallup Poll, “The president's job approval rating is an important indicator of re-election probabilities. But like so much else in this election, this measure isn't giving us a great deal of direction right now. Bush's job approval has slipped to 48% among national adults and is thus below the symbolically important 50% point. If we take that 50% line seriously, then Bush is in a less-than-auspicious position. No president since Harry Truman has won re-election with a job approval rating below 50%”.

6- Bush gained 9% over his 2000 vote in heavily Democratic urban locations and lost 3% in highly Republican small towns and rural areas.

But that is an Urban Legend.

7- There is no evidence that the vote count was corrupted. So how did Bush win 51.5% of the first 115.8m recorded votes while Kerry won 54.6% of the final 5.3m, a 500,000 vote decline in the Bush “mandate”.

8- Final pre-election polls did not match the exit polls. But after undecided voters were allocated, pre-election state (Kerry 47.9-Bush 46.9%) and national (Kerry 47.2-Bush 46.9%) polls closely matched the national (50.8-48.2%) and state (51.8-47.2%) exit polls. The final Zogby battleground state polls showed Kerry leading in 9 critical states.

9- Bogus assumptions were used in the pre-election Election Model simulation model which forecast that Kerry would win 320-337 electoral votes. But the only input assumption was that Kerry would capture at least 60% of the undecided vote. The Election Model popular and electoral vote projections were confirmed by state and national exit poll analysis in the Interactive Election Simulation Model.

10- There is no evidence that undecided voters break for the challenger. But historical evidence indicates that undecided voters break for the challenger over 80% of the time - especially when the incumbent is unpopular - and Bush had a 48.5% average approval rating. World-class pollsters Harris and Zogby reported their late polling indicated Kerry would win 60-80% of the undecided vote. According to the Gallup poll: “In the final USA TODAY/CNN/GALLUP poll, President Bush held a 49-47 edge over Sen. John Kerry when the undecided voters were not allocated to a particular candidate. When Gallup, using a statistical model that assumes that 9 of 10 of those voters would support Kerry, allocated the voters, the poll ended as a dead heat with each candidate garnering 49%. The Gallup allocation formula is based on analyses of previous presidential races involving an incumbent”.

11- Bush was leading in the final pre-election polls. But Kerry led Bush by less than 1% in the state-weighted average vote and held a 3% lead after allocation of undecided voters. Kerry and Bush were tied at 47% based on the final national 18-poll average.

12- Non-response bias in the exit polls was the reason why 43 states red-shifted to Bush. But 3 million votes (mostly from heavily Democratic minority districts) were never counted - and Kerry won 70-75%. Response rates were lowest in Kerry urban strongholds.

13- It was just a coincidence that Oregon was the only battleground state which shifted to Kerry. But Oregon is also the only state which votes exclusively by paper mail ballots and the shift was a fraction of one percent. Any discrepancy would be minimal and favor either Bush or Kerry.

14- There is nothing suspicious in the fact that six of the eight states which deviated to Kerry from the exit polls were strong Bush states. But the exit poll discrepancies were all within the margin of error.

15- Final exit polls are always matched to the recorded vote. But this assumes a fraud-free election and no uncounted votes. And since 2000, elections have been anything but fraud-free. Therefore, the final exit polls (state and national) cannot be correct.

16- There are reasonable explanations why Kerry led the National Exit Poll by 51-48% at 4pm (8349 respondents), 7:30pm (11027) and 12:22am (13047) and Bush won the 2pm Final NEP (13660) by 51-48%. But they defy the Law of Large Numbers and the Central Limit Theorem.

17- The margins of error used to calculate probabilities of exit poll discrepancies were too low. But even assuming a 60% “cluster effect”, the probabilities were near zero. Assuming a zero cluster effect, WPE-adjusted state exit poll discrepancies exceeded the MoE in 29 states for Bush and just one for Kerry. Assuming a 30% cluster, the margin of error in was exceeded in 24 states for Bush. Composite (12:22am) exit poll discrepancies exceeded the margin of error in 16 states for Bush; none for Kerry.

18- There is nothing suspicious about the fact that all 21 Eastern Time Zone states red-shifted from the exit poll in favor of Bush. But 14 deviated beyond the margin of error with virtually ZERO probability.

19- Exit polls are not true random samples. But Edison-Mitofsky state in the National Exit Poll Notes and the 2005 Exit Poll Evaluation Report

that respondents were randomly-selected and the overall margin of error was 1%.

20- Bush voters were reluctant to respond to exit pollsters. But this was contradicted by the Final Exit Poll, which stated that Bush 2000 voters comprised 43% of the respondents, as compared to 37% for Gore voters. It was also contradicted by a linear regression analysis: non-response rates increased going from the strongest Bush states to the strongest Kerry states which suggests that non-responders were Kerry voters.

21- The Final Exit Poll was correct in matching to the Bush 51-48% win. But the 43/37 weights indicate that either Bush voters were over-polled or the Final was rigged - a classic Hobson Choice. In any case, the rBr hypothesis was debunked.

22- False recall (of Gore voters) is a valid explanation for the Final NEP “Voted in 2000” weights. But Bush 2000 voters could not have comprised 43% (52.6m) of the 122.3m votes recorded in 2004; Bush only had 50.5m votes in 2000. Approximately 2.5m Bush 2000 voters died prior to the 2004 election and at least 2m did not vote, so 46m was the maximum Bush 2000 voter turnout in 2004 (a 37.6% weighting). The “Voted in 2000” question was asked of 3200 of the total 13047 respondents. How come the majority of the 10,000 respondents( who were not asked how they voted in 2000) claimed to have voted for Kerry? Why would they lie?

23- Although the Final NEP Bush/Gore weightings were impossible, the DU Game thread showed that Bush could have achieved his vote assuming feasible weights. But in order to match the recorded vote, they had to compensate for the feasible weights by inflating Bush vote shares to implausible levels.

24- Bush won by 3 million votes: 14.6% of Gore voters defected to Bush in 2004; only 7.2% of Bush voters defected to Kerry; Kerry won just 52.9% of new voters. But according to the NEP, Kerry captured 10% of Bush 2000 voters while Bush won just 8% of Gore voters; Kerry won 57% of new voters and 71% of Nader voters.

25- False recall is a valid argument to explain the Final NEP 43/37 weights. But false recall is not a factor in pre-election polls. And the pre-election polls (after undecided voters were allocated) matched the exit polls.

26- An NES study suggests that a significant percentage of Gore voters told exit pollsters they voted for Bush in 2000. But only 3200 of 13047 NEP respondents were asked how they voted in 2000. How do they explain the other 10,000 who were not asked to recall how they voted?

27- Returning Gore voters misrepresented their 2000 vote because they wanted to be associated with the winner - Bush. But Gore was the official “winner” in 2000 - by 540,000 votes. And Bush had a 48% approval rating on Election Day. The majority of new voters were Democrats and Independents; they gave Bush a 25-30% approval rating.

28- Bush found 12 million new voters in 2004. But simple arithmetic shows he needed more than 16 million. He had 50.5m recorded votes in 2000, of whom approximately 2.5m died (assuming 1.2% annual voter mortality) and another 2.4m did not vote in 2004 (assuming 95% turnout). Therefore, approximately 45.6m Bush 2000 voters turned out to vote. Bush needed over 16.4m of 25m new voters (64%) to reach 62m. But according to the 12:22am Composite Exit Poll, he won 10.2m (41%) new voters.

29- Bush won because 4 million new fundamentalist voters were registered. But according to the NEP, Kerry won 57% of new voters (first-timers by 55-43% and others by 61-37%)

30- The sensitivity analysis of voter turnout which showed that Kerry won all plausible scenarios does not prove anything. But given the adjusted 12:22am Composite vote shares and assuming 100% Bush 2000 voter turnout, Kerry needed just 73% turnout of Gore voters to tie Bush. Gore voter turnout had to be 64% in order for Bush to match his recorded vote margin. Using pristine, non-adjusted exit poll shares, required Gore voter turnout was even lower.

31- Bush’s share of females increased by 4.2%. But his share of males declined by 0.2%. Females voted 54-45% for Kerry; Blacks voted over 90% for him; he won Independents by 52-44%.

32- Bush won Ohio. But there is substantial documented evidence of uncounted and switched votes, as well as massive voter disenfranchisement. Two election workers were convicted of rigging the recount. Fifty-six of 88 county voting records were destroyed. The final Zogby poll had Kerry leading by 50-47%. Kerry led the pristine unadjusted Ohio exit poll by 54.1-45.9% (10.9% average WPE). He even led the adjusted 12:22am Composite by 52.1-47.9%.

33- Bush won Florida by 52-47% (a 368,000 vote margin) even though the Democrats had a 41-37% registration advantage in Touch Screen (TS) counties and a 42-39% edge in Optical Scan (OS) counties. But Kerry won TS counties (3.86mm votes) by 51-47% and Bush won OS counties (3.43mm votes) by a whopping 57-42%. The Final Zogby Florida pre-election poll had Kerry leading by 50-47%. Kerry won the unadjusted exit poll by 50.9-48.3%. In 2000, Gore won 70% of 180,000 uncounted under/over votes. Had they been counted, he would have won by at least 60,000 votes. Dan Rather's expose on voting machines proved that poor-quality paper used in Florida punch card machines was a major cause of spoilage in heavily Democratic precincts.

34- The New York pre-election poll matched the recorded vote (59-40%), while the exit poll was wrong (64.1-34.4%). But even if the exit poll was wrong (it wasn’t), the discrepancy can easily be explained. The margin of error for the 600-sample pre-election poll was 4%, so there was a 95% probability that Kerry's vote was from 55-63%. Assuming a 30% cluster effect, the exit poll MoE was 3.2%, so there was a 95% probability that Kerry's vote was between 60.9-67.3%. A vote in the 60.9-63.0% range would fall within the MoE of both the pre-election and exit poll.

35- The NY 2004 recorded vote was fraud-free; lever voting machines are foolproof. But this implies that 100% of returning Nader 2000 voters defected to Bush. In 2000, the NY vote was Gore 60.5-Bush 35.4-Nader 4.1. According to the 2004 NEP Composite (12:22am), Kerry won Nader voters by 71-21%; 10% of Bush voters defected to Kerry; 8% of Gore voters defected to Bush. And Kerry did better in NY than he did elsewhere. Kerry won NY by 63-36% using weights based on the NY 2000 recorded vote, assuming his vote shares were two percent higher than in the NEP. Levers are highly vulnerable to fraud. Like touch screens, they don’t produce a verifiable paper record. NY votes exclusively on Levers; the WPE was 11.4%. CT, another heavy Lever state, had the second highest WPE (15.7%) of all the states. NY was number four.

36- Polling data was cherry-picked and assumptions set to favor Kerry. But three sets of exit poll data provided by exit pollsters Edison-Mitofsky which included response rates and within precinct error (WPE) all produced equivalent results. The Exit Poll Response Optimizer confirmed the USCV simulation of 1250 precincts categorized by partisanship which debunked the reluctant Bush responder (rBr) hypothesis.

The Exit Poll Response Optimizer models used WPE and response rates for:

a) 1250 precincts categorized by partisanship- Strong Bush, Bush, Even, Kerry, Strong Kerry:

Kerry won the two-paty vote by 52.15-47.85%.

b) Location-size categories- Big Cities, Small Cities, Suburban, Small Towns, Rural:

Kerry won the two-party vote by an identical 52.15-47.85%

c) States exit polls: Kerry won by 52.3-47.7%

d) National Exit Poll “Voted in 2000” category: Kerry won by 51.9-48.1%

So, Brendan, I'm wondering what you think of all this? Is it still your opinion that the unweighted exit polls should not be trusted? If so, what exactly are your reasons for supposing that the official results should be trusted? What is your understanding of the methodology and rationale for applying "weightings" to the exit polls?

What if democracy as we know it has been seriously compromised by new polling technologies and failures to enforce strict accountability re: chain of custody of ballots? How would we know? What would patriots do?


Yes, if it contradicts the biased corporate media narrative, it's a crazy conspiracy theory, right?

So impressed with how mainstream and normal your views, just like the media pundits. That's what matters, right?

The comments to this entry are closed.

Friends & family