Exit Polls: Will They Be There in 2006?

I have heard that Edison-Mitofsky is not doing them, and I have also heard that no one is.

Is this info correct?

Thanks,

Rick

Exit polling is used internationally as a barometer to measure results with with the answers people give as to how they voted. There is little reason to believe people would lie on a massive scale as to how they voted.

after the 2004 election. Now keep in mind i heard this on MSM, CNN i think. But if the Die-bold fix is in, it makes sense that the powers to be do not want exit polling.


Peace!

No, your information is not correct.

http://www.mysterypollster.com/main/2006/05/fooled_again.html

However, if it's Edison-Mitofsky again, this will not be a good thing. The corporate media-supported institutions are not there to be accurate. That was proven in 2004.

I think the time has come for independent, regionally based market research firms to begin exit polling in key races. This would be relatively easy to do on a smaller scale. No doubt it would take a fair amount of $ to do, but it would be worth it.

Paging George Soros...

although I'd put a different gloss on it. Certainly the US Edison-Mitofsky exit polls are not designed to give an independent estimate of the vote-count. They have designed exit polls to try and do that (they did it in Azerbaijan) but you have to design the right poll for the right job, and the E-M US exit polls are not designed to do that job.

In fact, it is not a job that any exit poll (contrary to popular belief) is well-designed to do, although it may be all you've got in Azerbaijan, indeed, it may be all you've got in the US, although I'd suggest that resources would be much better spent on good direct monitoring of precincts and counts by diligent election reform activists. But the reason exit polls have problems that other polls, e.g. pre-election polls do not have is that they are conducted on a single day, with a vast number of respondents, and, therefore, with casually hired, minimally trained, and thinly spread interviewers - only one per precinct. A pre-election poll, in contrast, is conducted over several days, by professional interviewers, using randomised dialing, and with opportunity to follow up non-respondents. The sample size may be smaller, but sample size is no protection against biased sampling. Thee one thing an exit poll has going for it over a pre-election poll is that at least you can be pretty sure that your respondents have actually voted. You know, at least, that they got as far as the precinct.

A good exit poll (for example the BYU one) would have teams of well trained interviewers (trained to stick rigidly to their Nth voter protocol) ensuring good random sampling throughout the day. This would cost several times more than the already expensive (because so huge) E-M poll. The E-M poll, however, because it is designed to project the counted result, not provide an independent check on the count itself, is able to achieve "accurate" results (in the sense of projecting the counted results accurately) despite a fairly rough and ready precinct-level data collection system because it incorporates other data, including pre-election polls, and, eventually, incoming vote returns, into its projections. In other words, it compensates for bias in the poll by triangulating the responses with at least two other sources of information.

The problem, of course, is that it therefore can't be used as an independent check on one of those sources of information.

So, you are right - if you want an exit poll that is an independent check on the vote-count you need a different kind of exit poll to the one conducted by E-M. But it will cost a vast amount of money. And, I'm afraid, it still won't be immune from non-response bias, because no poll ever can be. If your vastly expensive exit poll gives you a different answer from the count, you still won't know which one was wrong. Although there will be ways of analysing the data that will provide strong indications one way or the other. But then that is also true of the E-M poll.

Totally disagress with you. Totally.

Now, who am I to believe?

I actually was an exit pollster in Indianapolis for Edison-Mitofsky. The training was exemplary for a one day job. There is absolutely no reason that I can see why their exit poll couldn't have been used as a predictive model AS LONG AS they didn't mix in actual vote totals. The non-response issue should randomize and no doubt in my mind that it did.

What was suspicious was that not only the Presidential election but most of the key Senate battles turned rapidly from blue to red late at night, contrary to decades of history where urban (D) centers reported late. Also quite suspicious was Mr. Mitofky's ridiculous and ever-changing excuses as to why his polling didn't match the results.

I know this kettle of fish has been stirred a lot since 04. Although I am a regular on DU for over two years now, I don't post a lot and don't particularly read this forum much since the final days of the "recounts".

So I don't wish to exhume the battle that you, Febble, and others who have a different view than you have fought over the many months.

I do appreciate your information, however.

Thanks,

Rick

And it's good to hear from an actual interviewer. But when you say that "the non-response issue should randomize" you appear to misunderstand the nature of non-response bias. Sometimes it does, sometimes it doesn't. It depends entirely on whether or not one group of voters were more likely to be non-responders than another.

Cheers

Lizzie

"What was suspicious was that not only the Presidential election but most of the key Senate battles turned rapidly from blue to red late at night, contrary to decades of history where urban (D) centers reported late."

The exit poll discrepancies were at least pretty comparable across most of the partisanship spectrum -- if anything, smallest in the most Democratic precincts. So I think it isn't surprising that the results would tend to flip rapidly (within the first few hours) as vote counts arrived, no matter what explains the discrepancies. The models are less influenced by the timing of returns than raw totals would be.

(As for Mitofsky's "excuses," I don't think they have changed very much, actually. I read a transcript of what he said on NewsHour, I think Friday after the election, and it seemed pretty familiar. Presumably he knows more now than he did then.)

In my view these were done electronically through the back door and pre-programmed opportunities for fraud in the machines and centralized tabulators, in addition to the types of fraud that occurred in the Warren County Ohio piece of the scandal.

What I think happened is that Bushco waited to see how many votes they needed to dump into the system and then they did that, through various pre-arranged means. It's no accident that Bush ended up with 3 million more votes than Kerry. No way they wanted Kerry winning the pop vote and losing the electoral vote after what happened in 2000.

In the earlier context, I thought you were referring to changes in the exit poll projections. Now I don't know at all. (This is not a gripe, just a report!)

Time for me to go to bed.

I am referring to the rather sudden shifts in the vote totals for several Senate races that showed D's ahead until late at night. Historically, the last precinct blocks that tally come in from urban centers which tend to be heavily Democratic. But for some reason, most of these races (Bunning in KY, Coburn in OK, Murkowski in AK, a few others that I can't specifically recall) mysteriously went the other way late at night, much as Bush v Kerry did.

The hypothesis is that Bushco carefully monitored each race and had pre-arranged dumps of R votes set to go, with only the quantity needed to be determined based on the exit polling and the actual vote totals as they accumulated. Then, about 11pm EST, they flipped the switch.

I can't prove this, and I'm no expert. I just got on this subject because of my concern about 2006 exit polling.

Like I said earlier, I didn't mean to stir the hornet's nest, it's been stirred enough. I know that Febble, you have a hardened position on this and I don't expect to convince you of my position.

I am always open to learning more facts but I doubt that I will see anything that convinces me that

a) The exit polls by E-M weren't compromised in some way
b) Bush really won fair and square
c) The Senate was really won by the Republicans fair and square

Hence, I am petrified about 2006, like many DU'ers. It's time for me to start thinking about what I can do.

First thing, I guess, is I'm going to re-enlist with E-M for 2006. I got pretty upset when one of E-M's defenses was that their contract employees weren't trained. Bushit. They did a great job and I posted this back in November and December 04, both here and on my blog.

I absolutely knew that I had to ONLY talk to every nth person. That was basic basic basic.

There is no way that I will ever be convinced that E-M's unpolluted data could not be used as a check against the actual polling. Truth is All and AutoRank have done an ample job of proof in this circumstance. Including the "non-responder" issue.

Like you, On the Other Hand, i'm heading to bed myself.

Thanks for listening!

I would like to testify on behalf of Febble that she doesn't generally have a "hardened position" on much of anything. She works harder at keeping herself honest than anyone else I know. I have heard that there is an old saying in military intelligence, which Febble appears to live by: roughly, "Assume that half of what you think you know is wrong, but you don't know which half." (Of course if one takes that too literally, then one is just rudderless.) I'm sure that, like all of us, she is stubbornly wrong about something despite it all, but she is an excellent person to disagree with IMHO.

I will never, ever try to convince anyone that "Bush really won fair and square." It can get strange around here really fast; if I say that I don't think that Bush won by 7 million votes, someone is liable to say that I am arguing for DREs. Sigh. --Let me say just a few things.

I have no idea whether what you say about KY, OK, AK, ? holds up; I don't know of a source that would let me reconstruct the sequence of returns. I guess streaming video of cnn.com might help. It seems to me that if the Republicans were in a position to do that, they could have done it with or without the exit polls; by 11pm EST they would have had ample vote count data. No? (Well, not for Alaska, but for the others....)

But if that is how it went down, then it would entail sudden consternation at NEP Central as, quite abruptly, all the estimates (or at least the ones you mentioned) started heading south. I think there would have been a fair number of people, some of them Democrats, who saw that happening, and I don't expect that they would all spontaneously suppress the knowledge. (For that matter, lots of people who were monitoring those races would have seen the same thing.) That is why I am initially skeptical of that scenario as I (now) understand(?) it.

There is good anecdotal evidence that not all the exit poll interviewers were as clear as you were: http://www.mysterypollster.com/main/2005/01/professor_m.html . However, I agree with you that interviewer training wasn't the crucial problem, and I don't remember E/M or NEP saying otherwise. It seems to me that their number one point has been getting closer to the polling places, which seems like an excellent idea. Seems to me that if you are 100 feet from the polling place, next to a MoveOn volunteer, and you are at a busy polling place, you will have trouble conducting a random sample no matter how well you are trained.

Actually, I partly agree that the exit poll data can be used as a check, but I totally disagree with TIA and Autorank about the results. I won't "stalk" you with my reasons! A lot of the things I worry most about (like "disappeared" registration forms) would hardly, if at all, show up in exit polls. There is a lot of work to do. Hang in there!

is at least half right about me (never mind which half) and in support of myself I would draw your attention to an ongoing conversation with eomer in the current psychology thread:

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=430061&mesg_id=430212

Obviously I'm more hardened than I was, but I'd certainly hope it is because of the amount I've learned, not because of the degree to which I've dug my heels into my starting position (which, as it happens, was that the exit poll discprepancy might well have indexed fraud).

OK, now to a couple of other points:

First of all, to read the E-M report as "blaming" the interviewers, is, I believe to fundamentally misunderstand it (just as to regard the theory that Bush voters were less willing to participate as "invented" is also to misunderstand it).

What was presented was a causal theory - that there was a slight but pervasive tendency for Bush voters to be less willing to participate in exit polls than Kerry voters. There is reasonably good precedent for this, in that for at least the last five presidential elections, the within-precinct discrepancy has been in the redshifted direction, most notably in 1992. Now, the reason could be historical fraud (although is is hard to explain why the effect was so small in 2000, but I accept it is possible). However, it cannot be measured directly, as, by definition, we cannot know the vote of those who should have been in the poll but weren't.

However, we can postulate that where precinct level factors existed that would be likely to make strict random sampling difficult (e.g factors that might allow more opportunities for unwilling voters to evade the pollsters, or factors that might might keeping track of Nth voter difficult (crowded precincts; long interviewing interval) then, if the theory is true, these factors should correlate with redshift. And, it turns out, many of them do.

This is interesting as it indicates that adverse conditions did not merely increase unsigned error (which it would do if there was no underlying tendency for Bush voters to be less willing) but increased signed error (redshift), lending support to the differential participation rate hypothesis.

In other words, the signed error was not the fault of the interviewers, but arose from an underlying differential propensity to participate which was reflected as redshift where conditions likely to increase errors were present.

This is not, and cannot be, proven, but when a theory makes a priori predictions that can be tested statistically, and those predictions are supported in the findings, then the theory looks to be in good shape.

But to turn to the other theory, that fraud, not differential propensity to participate, caused the exit poll discrepancy: this, also, cannot be measured directly, but, like the differential participation rate theory, gives rise to testable hypotheses. I will suggest two:

One is that discrepancies should be greater where a suspect technology is used (e.g. digital voting technology). However, it turns out that whilst in rural, suburban, and small urban places, there is overall less discrepancy (it tended to be greater in urban precincts), there is also no difference in discrepancy between precincts using different voting technology, including paper ballots. However, when we look at urban precincts (serving areas with a population of more than 50,000) we find that there is a difference depending on technology used. However, contrary to expectations, the discrepancy is greater where older technology was used (levers, punchcards) than where DREs and optical scanners were used. Now this is at least interesting, and suggests the possibility that something associated with those technologies was responsible for at least some of the discrepancy. But it doesn't support, very well, the hypothesis of massive digital vote theft as the cause of the overall discrepancy.

The other testable hypothesis is the one referred to on the thread I linked to, which is that unless fraud was universal and pretty well uniform, you would expect discrepancies to correlate with Bush's performance relative to some baseline, e.g. his 2000 vote share or pre election expections. But, it turns out, it does neither.

Because my views though possibly fairly viscous at this stage are not set quite hard, I am still messing about with the possibility that widespread, more or less uniform fraud might account for a substantial proportion of redshift without showing up as a positive correlation between redshift and performance, but it is not at all easy to see how anything plausible could do this. I shall keep trying though. (Another possibility is fraud carefully targetted at precincts where Bush was expected to do badly, but this scenario also presents fairly serious statistical - and practical - problems).

But I confess that my heart isn't as much in it as it should be because the alternative theory is actually very well supported. Nonetheless, there is room for the possibility of fraud in those thar plots. What I don't think there is room for is miscounted votes as an explanation for anything more than a small proportion of the undeniably large exit poll discrepancy.

But I will end by saying that none of the above has anything to do with the kinds of electoral injustice that we saw in Ohio, most of which would have been completely invisible to the exit polls.

"There is no way that I will ever be convinced that E-M's unpolluted data could not be used as a check against the actual polling. Truth is All and AutoRank have done an ample job of proof in this circumstance. Including the "non-responder" issue."

I agree with your first sentence. I think that with careful analysis it can tell us a lot. It's just that I think that following careful analysis (yes, some of it done by me) it mostly tells us that it is unlikely that massive digital vote-stealing occurred.

I don't agree with your second sentence, but that will by now be obvious.

But I do agree that the election was unjust (in ways that would not, by and large, have shown up as discrepancies in the exit polls.

Lizzie

seems to have culled most of his information about exit polls from Freeman and Baiman. I think Freeman and Baiman have produced profoundly flawed analyses. Having them cited by Kennedy doesn't make them any less flawed.

the least they can do for their grass roots volunteers and donors--and for the country--in reparation for their lack of vigilance in permitting Bushite corporations to gain control of all the vote counting with "TRADE SECRET," PROPRIETARY programming code and virtually no audit/recount controls. (Jeez...)

2. What I heard is that Edison-Mitoksky vowed to never let the public see the REAL exit polls again. We will only see the DOCTORED numbers--the exit polls falsified to match the NON-TRANSPARENT results of Diebold/ES&S's secret formulae.

This situation is truly mind-boggling. Stalinist Russia, here we come. HOWEVER, we are NOT Stalinist Russia--not yet anyway. Stalin took over a country with NO democratic traditions. No Declaration of Independence in their background. No Constitution. No Thomas Jefferson. No Tom Paine. No Underground Railroad. No unions. No "Big Tent" Democratic Party. No FDR. No Republican patriots like Earl Warren and Dwight D. Eisenhower. No Martin Luther King. No democratic infrastructure. No democratic principles. No Bill of Rights. Nothing. Ten centuries of highly repressive, autocratic rule.

So, while Stalin and "1984" are damn scary models--and the Bush junta seems to following them to some extent, as well as paralleling Hitler's rise in other ways--WE have a population of Americans who, even according to the war profiteering corporate news monopoly polls, are NOT BUYING this junta or its policies. And rebellion is in the wind. Frankly, I think that if Gore and Kerry run together in '08--or if Gore heads the ticket, with, say, Feingold or Clark as his second--the Democrats will blow Diebold/ES&S's election theft machines, and Edison-Mitofsky's exit polls (if they dare to do them), right out of the water.

The scandal that is brewing over the '06 Congressional elections is going to be the biggest in our history. Let's hope it sends these election theft machines right into 'Boston Harbor'! We need to help that along with CLOSE MONITORING of the '06 elections, in whatever ways we can, in highly non-transparent conditions. Here are some ways you can help...

------------------------------------------------------

SOME RESOURCES FOR AMERICAN REVOLUTION II:

Practical suggestions for the immediate future:

1. ABSENTEE BALLOT VOTING. Promote absentee ballot voting. It's not the ultimate solution, by any means, but it at least provides a tangible paper record for challenging suspicious election results, recounts and investigations. (Absentee ballots were a great help to investigators in 2004.) Absentee ballot voting is also a form of protest against the machines. 50% of Californians are now requesting Absentee Ballots. If enough people do it, the machines will be obsolete; then we can work on getting rid of the central tabulators.

2. MONITOR THE ELECTIONS. Join with others to closely monitor the coming elections and gather and document evidence. See www.UScountvotes.org, and other resources, below. UScountvotes.org needs donations!

3. DEMAND INDEPENDENT EXIT POLLS. Demand that the Democratic Party fund INDEPENDENT EXIT POLLS. Exit polls are used worldwide to verify elections and check for fraud. The war profiteering corporate news monopoly exit polls cannot be trusted (they are doctored to match the results from the voting machines' secret programming code; rather than being used to verify elections, they are used to confirm NON-TRANSPARENT "official results"). The Democratic Party owes us, big time, for their lack of vigilance--and in some cases corruption--on electronic voting. This is one critically needed thing that they can do to help.

4. THINK LONG TERM. Saving our democracy promises to be a long hard struggle. We obviously can't get rid of these machines before the '06 elections, so focus on doing our best with the Diebold/ES&S handicap (a 5% to 10% "thumb on the scales" for Bushites and warmongers), and getting rid of these machines afterward, for '08 and beyond.

5. TELL PEOPLE THE TRUTH. They NEED to know it. Engage them in the fight. Bumper sticker: "Help Us Beat the Machines--VOTE!" There is nothing more demoralizing and disempowering than constantly losing and not knowing WHY. There is evidence that the machines CAN be beaten by massive turnout. Get people involved! Help them to SEE what's happening! THEY will solve the problem, ultimately--if they can only IDENTIFY what it is!

6. PRESSURE LOCAL/STATE ELECTION OFFICIALS. Right now, the best place to fight this fight is at the state/local level, where ordinary people still have some influence. Bush's Congress is NOT going to give us back our right to vote--they are the ones who took it away (with the collusion of some corrupt Dems). Don't look to the Feds--look to your local county registrar, your state election boards, your secretaries of state. Demand TRANSPARENT elections. Also educate and mobilize your local Democratic Party groups.

History has chosen us to save this great democracy and pass it on to the future. The first priority in this historic fight is restoring our right to vote--the mechanism by which we exercise our sovereignty as a people. Without it, we have no power. We MUST change this.

Never give up on our right to vote! NEVER!

----------------------

ADDITIONAL RESOURCES FOR AMERICAN REVOLUTION II:

Hopeful signs - latest news:

California voters sue the state over Diebold:
www.VoterAction.org is suing the state of California and 18 Calif county registrars on behalf of 25 California voter/plaintiffs, on the illegal Diebold "certification" by Schwarzenegger appointee Bruce McPherson.
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x2180496
Seven of these counties just promised the judge they would use PAPER BALLOTS, and were dismissed from this lawsuit (4/27/06).
http://kcbs.com/pages/29285.php
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x2249205

Maryland rejects Diebold:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x418263

Florida - anti-trust accusations against Diebold, ES&S and Sequoia, re: heroic Florida election official Ion Sancho:
(FLA AG subpoenas the companies)
http://www.computerworld.com/governmenttopics/government/legalissues/story/0,10801,110192,00.html
http://www.tbo.com/news/politics/MGBKSY8W8LE.html
(info & discussion)
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x2183630

Utah county clerk fights back!
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x419226

(Tide turning?) New York Times: "New Fears of Security Risks in Electronic Voting Systems" (5/12/06)
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x2278829

-----

INFORMATION AND ACTIVIST RESOURCES for American Revolution II:

www.votersunite.org (MythBreakers - easy primer on electronic voting--one of the myths is that HAVA requires electronic voting; it does not.)
www.UScountvotes.org (statistical monitoring of '06 and '08 elections--they need donations)


(Activist sites with links to state activist groups or info)
www.votetrustusa.org (news of this great movement from around the country)
www.votersunited.org (good general info, and state links)
www.verifiedvoting.org (great activist site)
www.solarbus.org/election/index.shtml (fab compendium of all election info)

www.freepress.org (devoted to election reform)
www.bradblog.com (also great, and devoted to election reform)
www.TruthIsAll.net (analysis of the 2004 election)* :patriot: :applause: :patriot:
www.votepa.us (well-organized local group of citizen activists in Pennsylvania, where important legal issues are at stake, including state's rights over election systems)
Provisions of the PA lawsuit:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x423739

The Voter Confidence Resolution
http://tinyurl.com/rlnr2 (“We Do Not Consent”)
http://guvwurld.blogspot.com (GuvWurld blog main page)
http://tinyurl.com/amryg (Voter Confidence Resolution

www.debrabowen.com (Calif Senator running for Sec of State to reform election system)
www.johnbonifaz.com (running for Massachusetts Sec of State on strong election reform and antiwar platform)

*Some tributes to TruthIsAll, who is very ill:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x417007
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x417231
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=364x675477

Congressional bills:

Russ Holt's HR 550 requires a real paper ballot, bans secret software in "voting machines", and has more than 170 co-sponsors, but the audit required is too weak, it promotes electronic voting and centralized power, and the secret software might be permitted to continue in the central tabulators (the bill is not clear). At lot of discussion at DU of the loopholes/pitfalls in HR 550 (many DUers support the bill):
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x422926
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x421136
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=422967&mesg_id=422967
To sign the HR 550 petition: http://www.rushholt.com/petition.html

(Note: Senate Bill-SB 330 and House-HR704 simply require a "voter verified paper audit trail" (VVPAT), which may be best for the moment.)


Also of interest:

Michael Collins (Autorank)'s searing election reform article for New Zealand's Scoop.com
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=364x971363

Bob Koehler (-- four recent election reform initiatives in Ohio, predicted to win by 60/40 votes, flipped over, on election day, into 60/40 LOSSES!--the biggest flipover we've seen yet; the election theft machines and their masters are now dictating election policy! Title: "Poll Shock" 11/24/05)
http://commonwonders.com/archives/col321.htm

Bob Koehler's latest: "Trust us: Take this box and stuff it" (3/16/06)
http://commonwonders.com/archives/col337.htm
More Koehler:
www.tmsfeatures.com/tmsfeatures/subcategory.jsp?file=20051124ctnbk-a.txt&catid=1824&code=ctnbk

Amaryllis (Diebold, ES&S, Sequoia lavish lobbying of election officials - Beverly Hilton, Aug. '05)
www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x380340

HOWARD DEAN remarks on electronic voting machines 04/06
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=364x994507

------------------------------------------------

Throw Diebold, ES&S and ALL election theft machines into 'Boston Harbor' NOW!

:think: :patriot: :woohoo: :patriot: :think:

-----------

"That to secure these rights, Governments are instituted among Men, deriving their just powers from the consent of the governed, — That whenever any Form of Government becomes destructive of these ends, it is the Right of the People to alter or to abolish it." --Thomas Jefferson, The Declaration of Independence

You wrote: "What I heard is that Edison-Mitoksky vowed to never let the public see the REAL exit polls again. We will only see the DOCTORED numbers--the exit polls falsified to match the NON-TRANSPARENT results of Diebold/ES&S's secret formulae."

What E-M will take care NOT to release is the midday results - i.e. the results from incomplete samples. I have heard nothing to indicate that close-of-poll projections will not be released, as usual.

I was working GOTV and in mid afternoon I got a short break and checked out DU, the early numbers and all the euphoria. I contacted my friend and we bet Kerry at Tradesports.com. He was already the slight favorite then maybe an hour after our bet he was 1/2.

I felt like a world class dunce when I came home and looked at the early numbers more closely. Several of them were bizarre and all tilted toward Kerry, like up 17 in New Hampshire and up 10 in Pennsylvania. Then I checked Florida and Ohio, each only a point or two for Kerry. Numbing beyond description. I was sure they would err similarly to the obviously flawed numbers, and Kerry was finished. I turned off the computer and TV even whille DUers were still cedlebrating.

We'll still have leaks of the midday numbers. No way someone like Drudge is going to sit there and pretend he doesn't have inside sources.

They have to steal more votes than ever this time, and they don't
want those pesky exit polls getting in the way anymore.

don't build a case on a non-fact.

I'd hazard a guess that as the country moves towards Vote-By-Mail, like Washington and Oregon, that exit polls will go silently into the night.

Anyone know how it would be possible to conduct exit polls with VBM?

For so-called exit polling. I've seen it detailed. They take extra steps in that state, random digit dialing (RDD) along with selecting names from a registration list. Precautions to make sure they are sampling only one person from a given household. I know I have a more specific description in my files but right now I can't find it.