Comments on: free the exit poll data Blog, news, books Thu, 12 Oct 2017 08:56:00 +0000 hourly 1 By: abhi Thu, 09 Aug 2007 20:29:11 +0000 please vote

By: Ron Ulan Fri, 03 Dec 2004 20:11:04 +0000 I strongly suspect that the exit-poll actual differential is merely
a reflection of voters who are happy to run away from all pollsters, and that such people would more likely be skewed toward the
right then the left in this country, especially amongst the new voters who turned out for 2004. I was raised liberal. I worked for McGovern in 72 when I was in college. I grew conservative over the years. Once in my life, an exit pollster apporached me as I left the polling place In Whitestone, Queens County, where I reside.
I saw him and ran from him as if he was a disease. I regretted this later, but my instant reaction was that this Exit Pollster was a walking disease that needed to be avoided as if he were a contagion. If voters for Bush in 2004, more so then voters for
Kerry reacted as I did, it would be impossible for exit polls NOT
to be skewed from the actuals. And as for Florida, land of
retirees from the North, I would not be surprised at all if many
lifelong Democrats voted for Bush at the top, and then Democratic the rest of the way….and then ran from Exit Pollsters. Bush haters cannot, will not, even try to imagione this
as possible. It is so possible, I would be surprised if it were not the case.

By: Max Lybbert Thu, 18 Nov 2004 10:45:21 +0000 For the record, a partial recount would be fine if there were some credible reason (such as fraud or illegal activity) for one county to be recounted above all others.

Gore’s claim that Palm Beach (with the butterfly ballot) needed a recount was credible, but thrown out in court. His claim that punch-card ballots were error-prone, however, was a pretty transparent excuse. If he were truly concerned about punch-card ballots, why didn’t he ask for a recount of all counties that used punch-card ballots? Really, he asked for the recount, and found an excuse later.

By: Max Lybbert Wed, 17 Nov 2004 20:39:00 +0000 Sorry, Jon, Bush doesn’t have a “proven record” of voter fraud. I’m amazed how people still believe Bush stole the 2000 election. Here’s a quick rundown:

  • the initial count of ballots said Bush won Florida, so did the automated recount;
  • Gore asked for a manual recount in only the four most liberal counties of Florida (out of 60+ counties), intentionally ignoring the votes of nearly all Floridians, he then started talking about counting every vote (some votes are more equal than others?);
  • Gore failed to convince a court that the butterfly ballot was illegal — disarming the only legal argument for a recount in a single county (compared to a statewide recount);
  • Gore failed to explain why he hadn’t asked for a statewide recount, when the election was essentially statewide, he also never explained how a manual recount could be more acurate than an automated recount (think thousands of different people in different counties with different standards);
  • the US Supreme Court ruled 7 to 2 that Gore’s recount requests in only four out of 60+ counties were unconstitutional, the Court sent the case back to the Floridian Supreme Court;
  • a different aspect of the case reached the US Supreme Court, this time five justices decided that the recount requests were unconstitutional;
  • the 5-4 majority refused to give the Floridian Supreme Court another bite at the apple;
  • five members of the court explained that the Constitution prohibits anything less than a statewide recount for a statewide election;
  • three of those same members of the court also claimed that the Floridian Supreme Court had “interpreted” the law to mean something that wasn’t ever in the law;
  • the remaining justices of the Supreme Court wrote opinions about how the Supreme Court shouldn’t get involved in politics, etc. but never explained how a “partial recount” in a statewide election could be Constitutional;
  • people started claiming that Bush stole the election, even though Gore was the one who requested an error-prone manual recount that intentionally would not count votes in conservative counties.
By: Jon Spitz Wed, 17 Nov 2004 13:30:20 +0000 Mr. Lessig,

Your analysis of the discrepencies between Exit Polls and the actual vote count fails to address the most relevent points of the controversy. Although I agree with you that the polling data should be made public to discount your theory that it was incompetence that caused the disecrepencies, other facts that we already do know make this possability highly unlikely. First, Edison/Mitofsky has a long track record of polling results that are dead on; they know what they are doing. Second, in non-swing states, their polling was within the expected margin of error .5%. Third, it is not in their interest to falsify the data since their business depends on their polling results proving accurate. Fourth, they have no record of partinship to suggest that they would falsify the data. Fifth, Bush not only has a strong motive to commit election fraud (to stay in power), he also has a proven history of commiting voter fraud (Florida in 2000). Sixth, with computerized vote tallying (not the voting machines but the vote counting machines themselves) it is now very easily possible to commit voter fraud on a massive scale. For your analysis to be convincing, you must address these issues.

Jon Spitz
Laytonville, CA

By: Max Lybbert Tue, 16 Nov 2004 14:50:33 +0000 Well, my linking abilities seem to have been degraded. That should be:

“In fact, I now agree with Lessig that the raw exit poll data should be released so that intelligent statisticians can take a look at it. …”

Teach me to skip the Preview stage. …

By: Max Lybbert Tue, 16 Nov 2004 14:46:58 +0000 Alex, I’m finding your comments and corrections very insightful.

In fact, I now agree with Lessig that the raw exit poll data should be released so that intelligent”>intelligent”>”>intelligent statisticians can take a look at it. I don’t expect anything to be found, but I do understand the concept of auditing even without proof of dishonesty or fraud. I signed a petition for an independent investigation of Abu Ghraib for the same reason. And while I’m not holding my breath, I’m interested in BlackBox Voting’s FOIA requests and the Green Ohio recount request.

Can the electronic voting machines be compromised? I believe so. Is there a better way? Yes. Should we take it? Yes. Were the machines compromised on a wide scale on Nov 2? I doubt it.

By: Bernie Tue, 16 Nov 2004 02:18:39 +0000 Does anyone know anything about the effort to get access to the exit poll data by county?? That is what is missing in much of the discussion.

By: Alex in Los Angeles Mon, 15 Nov 2004 18:55:03 +0000 About the Caltech/MIT study:

They use exit polls that have been revised to MATCH the election results. They seem not to realize that per standard exit poll methodology,, their exit poll source, is now reporting exit polls that have been reweighed to match election results.

It is quite surprising, and I imagine will be somewhat embarrasing once they realize their error. Obviously, if they want to compare exit polls to election results, they would need to analyze exit polls that are not revised to match the election results.

Explanation of exit poll methodology:

Exit poll study that recognizes exit poll methodologies:

By: Paul Hughes Mon, 15 Nov 2004 18:15:42 +0000 Larry,

I’ve come to respect you tremendously over the last year, but your telling us that you that these claims are bunk is equally bunk. Why? Because neither you, nor any of us have any idea what the actual vote count was, or if it was manipulated or not. Unless you are psychic, and you have never claimed to be, then you can’t claim these suspicions are unfounded. Nobody knows, and that is the problem. If you do the math, these numbers don’t add up. There IS definitely a SIGNIFICANT statistical descrepancy here. To deny this is silly. So any wise and judicial person would want to investigate this descrepancy and get to the bottom of it at all costs. This is our democracy that is at stake. Playing it down as you do, only furthers the problem.

Right now we should all be fighting to have open-source elections and accountability (and re-countability) at every level of the electorate. Anything less and it’s a stab our democracy.

By: Beth Mon, 15 Nov 2004 05:48:15 +0000 Why not simply count the ballots in those questionable counties? That would show conclusively whether the inconsistencies were the fault was of the polls, the machines, or some of both.

By: H�r�ur Helgi Helgason Sun, 14 Nov 2004 12:20:41 +0000 Excuse me, but has nobody answered the latter part of Mr. Carnell’s colorful commentary, e.g. “All you have to do is look at how these inverted counties � Baker, Franklin, etc. � voted in 2000. Baker is >69% Dem by registration but went for Bush by 69% in 2000 and 72% in 2004. You will find similar results for Franklin and the other counties that have been mentioned as being excessively inverted.”?

I did see the exchange between Mr. Carnell and Mr. Don (Whiteside?), about the correlation of voter registration to voter behavior, but that exchange does not address the really interesting point in this thread: Did many Florida counties break away from past presidential election trends in severe ways, i.e. did counties that have traditionally gone for presidential candidates of one of the two largest parties, by a large margin, suddenly now go for the candidate of the other one of these large parties, also by a decisive margin?

By: Sepp Hasslberger Sat, 13 Nov 2004 20:39:54 +0000 According to a News Update from Citizens for Legitimate Government
( November 10, 2004 ) :

Diebold Source Code!!! –by ouranos ( “Dr. Avi Rubin is currently Professor of Computer Science at John Hopkins University. He ‘accidentally’ got his hands on a copy of the Diebold software program–Diebold’s source code–which runs their e-voting machines. Dr. Rubin’s students pored over 48,609 lines of code that make up this software. One line in particular stood out over all the rest: #defineDESKEY((des_KEY8F2654hd4″ All commercial programs have provisions to be encrypted so as to protect them from having their contents read or changed by anyone not having the key… The line that staggered the Hopkins team was that the method used to encrypt the Diebold machines was a method called Digital Encryption Standard (DES), a code that was broken in 1997 and is NO LONGER USED by anyone to secure programs. F2654hd4 was the key to the encryption. Moreover, because the KEY was IN the source code, all Diebold machines would respond to the same key. Unlock one, you have then ALL
unlocked. I can’t believe there is a person alive who wouldn’t understand the reason this was allowed to happen. This wasn’t a mistake by any stretch of the imagination.”

By: Max Lybbert Fri, 12 Nov 2004 19:28:12 +0000 My only hang-up is the tradition of casting a blank vote,” meaning you vote for the winner — whoever that turns out to be. It’s part of the reason there are nearly always landslides in Brazilian elections. That, and PSDB is the single (large) conservative party opposed by about a million leftist parties, including PT (the party of the current President). It’s very similar to Mexico and PRI vs the world + PAN.

Anyhow, Brazil goes for a simple majority vote for President, and every 50 years has a referendum determining if the country should remain a republic, go parliamentarian, or become a monarchy again. International politics can get interesting.

By: Ygor Valerio Fri, 12 Nov 2004 19:07:22 +0000 Well, they actually do. And, a plus: voting in Brazil remains one person = one vote, also in presidential elections. It certainly makes our job easier.

By: Max Lybbert Thu, 11 Nov 2004 15:45:47 +0000 For what it’s worth, I’ve seen the Brazilian voting machines, and they look pretty cool. Here in NC, about half of the precincts use electronic machines that have gray-and-black LCDs and push buttons. In Brazil, they have touch-screen machines that include the name, party, and picture of each candidate.

The issue, of course, isn’t the “wow” factor — it’s whether the machine has a way to verify that the vote total is accurate.

By: Ygor Valerio Thu, 11 Nov 2004 14:39:54 +0000 On the electronic voting machines issue, the Brazilian experience has been a role model to some countries. In the recent municipal elections, which happened throughout the whole country, we had 0,01 percent of errors in machines. All of them were immediately replaced.

By: Max Lybbert Thu, 11 Nov 2004 10:49:11 +0000 Well, MIT and Caltech have been digging (PDF).

To twist Lessig’s words, “Anyone who is surprised that an exit poll didn�t work has been living on Mars for the last 100 years.” By “didn’t work,” I mean, “didn’t accurately predict a 3% spread when the margin of error is 4%.”

By: garcia oliver Wed, 10 Nov 2004 18:10:15 +0000 Lawrence Lessig says, �Always, and in every election, voting machines fail. That fact should force us to a sensible architecture for voting machines � one which we don�t have just now for electronic voting machines. But it isn�t, itself, evidence this election was �stolen.� �

Of course not. There is no �smoking gun�. At least not yet. But there is some circumstantial evidence, e.g.:

(1) �While the heavily scrutinized touch-screen voting machines seemed to produce results in which the registered Democrat/Republican ratios largely matched the Kerry/Bush vote, in Florida�s counties using results from optically scanned paper ballots – fed into a central tabulator PC and thus vulnerable to hacking � the results seem to contain substantial anomalies.

�In Baker County, for example, with 12,887 registered voters, 69.3% of them Democrats and 24.3% of them Republicans, the vote was only 2,180 for Kerry and 7,738 for Bush, the opposite of what is seen everywhere else in the country where registered Democrats largely voted for Kerry.

�In Dixie County, with 9,676 registered voters, 77.5% of them Democrats and a mere 15% registered as Republicans, only 1,959 people voted for Kerry, but 4,433 voted for Bush.

�The pattern repeats over and over again – but only in the counties where optical scanners were used. Franklin County, 77.3% registered Democrats, went 58.5% for Bush. Holmes County, 72.7% registered Democrats, went 77.25% for Bush.� []

While there are possible innocent explanations, they are addressed in the commondreams article.

(2) Kerry Margins: Paper-ballot and Non-paper ballot states show big differences. If accurate and not cherry-picked, these data are damning. However, I don�t know the selection criteria used, so don�t accept this study uncritically. []

(3) Many (most?) States that use e-voting have refused to include paper audit trail/backup despite nearly unanimous recommendation to the contrary by computer experts concerned about hacking or manipulation of results. This is a recipe for fraud.

�[T]hree separate reports issued by computer security experts have determined that electronic voting is a risky business. In fact, the most recent report issued in Maryland showed that the leading touch screen voting system offered by Ohio-based Diebold Corp. was vulnerable to tampering.

�We were genuinely surprised at the basic level of the exploits that allowed tampering,� said Dr. Wertheimer, a former employee for the National Security Agency now working as a director at Raba Technologies LLC, a Maryland-based security consulting firm.

Raba teams apparently found it easy to hack into the Maryland Touch Screen voting system built by Diebold. Raba�s report on hacking the Diebold voting machines noted that the system was insecure and required some basic changes before it could be reliable enough for a general election.

�The State of Maryland election system, as configured at the time of this report, contains considerable security risks that can cause moderate to severe disruption in an election,� noted the report.

�Further steps could be taken to ensure a safe general election in November. � Ultimately, we feel there will be a need for paper receipts, at least in a limited fashion,� stated the report.� []

(4) And, speaking of Diebold, �Walden �Wally� O�Dell, [is] the chairman of the board and chief executive of Diebold. For years, O�Dell has given generously to Republican candidates. Last September, he held a packed $1,000-per-head GOP fundraiser at his 10,800-square-foot mansion. He has been feted as a guest at President Bush�s Texas ranch, joining a cadre of �Pioneers and Rangers� who have pledged to raise more than $100,000 for the Bush reelection campaign. Most memorably, O�Dell last fall penned a letter pledging his commitment �to helping Ohio deliver its electoral votes to the President.� � []

Is this who you want in control of unaudited electronic voting machines?

Is this conclusive? No. But there is sufficient evidence to be highly suspicious of the official results. And reason to keep digging.

By: James (incredulous in Canada) Wed, 10 Nov 2004 14:33:09 +0000 “…nor do I think Diebold stole the election for Bush”

I see this sentiment everywhere these questionable results are discussed. I don’t understand it. When will Americans wake up and realize that their govt. is in the hands of thugs who will stop at nothing to get what they want?

I realize Prof. Lessig is a lawyer, and due process is his bag, and I respect that. But they (the administration) do not. Not a bit.

I appreciate that the professor is at least doing his bit to draw attention to this debacle, but the US, and the rest of the world, are running out of time for polite demands for inquiries, and dark intimations of that old bugaboo “incompetence.” Always with the incompetence. I am so sick of that word and it’s synonyms being so consistently used to explain away actions of obvious ill-intent.

Election 2004 was a consolidation of the 2000/01 coup d’etat. This seems obvious to me. When will it be obvious to you?

By: Anon Wed, 10 Nov 2004 11:13:52 +0000 I’m sure all the lefties have a subscription to Salon, so go read this.

“Salon has examined some of the most popular Kerry-actually-won theories currently making the rounds online, and none of them hold up under rigorous scrutiny.”

By: Max Lybbert Wed, 10 Nov 2004 11:13:13 +0000 Joe wrote (to Lessig):

You�re telling us to be responsible with our allegations. Based on the experience of 2000, why should we be? They stole an American election, Professor. They stole it.

So, in the presence of true legal scholars, I would like somebody to understand why a partial recount is Constitutionally valid.

That is, if I were running for school board, and lost by 50 votes, would it be valid for me to ask for a recount of all white votes, without counting votes cast by Blacks? What about asking for recounts of predominantly-white neighborhoods? What about asking for recounts of predominantly-Republican neighborhoods?

What if I were named Gore, running for President, and asked for recounts of predominantly-liberal counties, while intentionally ignoring predominantly-conservative counties? What if I were to get on TV to nobly declare “every vote should be counted” while I worked very hard to have only certain votes counted?

On top of that, why should I put any more stock in manual recounts that have incompatible standards, that one county didn’t actually finish, and where the same county twice changed the standard of a legal vote during the recount?

A good essay about this can be enlightening.

For the record, the same place hosts several essays about Bush v. Gore that I haven’t read (type “bush gore” in the Quick Search box). I’m getting them now to see if anybody actually answers these questions.

By: Max Lybbert Wed, 10 Nov 2004 10:47:44 +0000 Alex, thank you for pointing out my mistake. I took another look, and the original post doesn’t ask “why were the leaked exit poll numbers wrong?” like I thought it did. Instead it asked (as you pointed out) “why were the complete exit polls wrong?” Since I haven’t yet seen the exit polls, I don’t know.

I know that nationwide, Bush had a 2% lead, and that would be within the margin of error of just about any poll I’ve ever seen. But, since each state must be looked at separately, this isn’t relevant.

By: J.B. Nicholson-Owens Wed, 10 Nov 2004 00:01:35 +0000 Bev Harris is going about it the right way–she’s calling for important data by launching a series of Freedom of Information Act requests. I’m willing to see what fruit comes from this.

Why did Democrats in those counties overwhelmingly defect to the President while remaining �liberal� in their other votes?

I can’t address the second part of that question, but in 2000, many thousands of Florida Democrats voted for Bush. In fact, there were so many of them they dwarfed the number of Nader voters. I don’t know why Democrats in Florida voted that way then either.

By: Alex in Los Angeles Tue, 09 Nov 2004 22:21:16 +0000 “Sorry, but incomplete polls with small sample sizes are often wrong. When complete, the polls were supposed to have a margin of error of 4%, but they weren�t even half done when the information was �leaked.� I�m not a statistician, but I understand that would have given the polls a margin of error so large that they would have been utterly useless and inaccurate. So, yes, they can be wrong, and as wrong as they were.”


You are attacking a straw man. Prof. Lessig is calling for the complete Final Exit Polls, not the incomplete early exit polls. MysteryPollster explains that the complete Final Exit Polls are released after the close of polls, and are normalized by turnout, gender, etc.

Final Exit Polls differed from the vote count significantly, but less so than the incomplete early exit polls. MysteryPollster also repeats the call for the owners of the exit poll data to conduct an analysis to debunk, or not, errors/fraud in the vote count:

See also Mickey Kaus’ call:

And Keith Olbermann:

Note, most of us expect to find Exit Poll Bias, not fraud. But Exit Poll bias would be a big deal and unprecedented.
See, again, Mystery Blogger:

Hope that helps clear up some questions.