Psychological Resistance to Facing Election Fraud

From: TruthIsAll.net

The Silence of the Scams:

<snip>

Ignorance of Extent of Dirty Tricks

If people knew about the amount and extent of dirty tricks, 3,500,000, or 1,750,000 may not seem so insurmountable. Some of the tricks documented include throwing out of Democrat voter registration forms, broken machines, misplaced machines, machine errors, reduced numbers of machines in Black and predominantly Democratic areas, less than in 2002, causing long lines, unmailed absentee ballots, absentee ballots requesting 86 cents, insufficient postage, which were returned, certification of more votes than registered voters in some areas, reversal of percentages of registered Democrats and votes for Bush in many counties, modem connected voting machines and tabulators, different standards for provisional ballot recounts in different areas, many provisional ballots, also called “placebo ballots”, not counted at all, voting machines defaulting to a Bush or 'jumping' by recording a vote for Bush when Kerry's button was pushed, phony companies registering voters and then tearing up the registrations of Democrats but not Republicans, exit polls not corresponding with reported votes in counties with no paper trail, while exit polls matched reported votes in counties with paper trails, voting elections officials creating what look like phony election machine poll tapes and tossing original, signed tabulations in the garbage, people posing as technicians coming in and tampering with machines, Republicans posing as Democrats, a lock down, refusing to let observers in, with the excuse of terrorist alert to observe the counting of votes in a country in Ohio, misinformation about the date and location of voting in Black neighborhoods, threats of arrest for voters with traffic tickets or any record, unusual discrepancies between numbers of votes for Kerry and Democratic candidates on same ticket, and widespread refusal of media to report on any of these, and a media campaign trashing exit poll data with made up reasons. And these are just the ones we know about.

Discomfort with Numbers

The best evidence for fraud in the 2004 election is statistical, according to Josh Mitteldorf of TempleUniversity's Statistics Department. Many are uncomfortable with numerical and statistical science that quantifies judgments about likelihood. For example, statistician Dr. Steve Friedman of University of Pennsylvania, and graduate of MIT found that the discrepancy between exit polls and the actual vote count in each of three states, Florida, Ohio and Pennsylvania, is 1 in 1,000,000, but the likelihood of all three states being discrepant in the same direction is 1 in 250,000,000. What people heard in the news was a smear campaign invalidating the credibility of exit polls, even though they are considered highly accurate, are used in many countries as indicators of fraud, and that exit polls in counties with a paper trail matched the official vote count, and in counties where there was no paper trail and evidence of computer irregularities, the official count was different than the exit polls and always favored Bush. They even made up fake reasons for this discrepancy regarding response bias--which did not exist where there were paper trails.

Disbelief

Many people don't believe the allegations of fraud because they didn't read about it in the New York Times or hear it on CNN. (The only mainstream media to report it was Keith Olberman on Countdown, MSNBC.) We might wonder about the media censorship on this story and intentions to promote disbelief in the populous, in addition to ignorance.

Conformity and Herd Mentality

Because of the media blackout, ignorance, and emotional tone of reporting, Americans have a false perception of consensus about objective reality. The majority conforms to this misperception and most do not have the psychological make-up to challenge the status quo. The few that are courageously addressing this are not heard, or else they are severely shamed, ridiculed and viciously accused of causing problems. Thus, even the thought of questioning is suppressed.

more at link:
http://www.truthisall.net/Diane_Perlman/diane_perlman.html

????

A much larger spread predicted than actually happen - so large some say not reasonable?

blue states if not the deepest Kerry was polling in the low 70's.

His vote tally was in the low 60'a. Ahh we use the hightly riggable optical scam machines. There tabulation software can rig stuff so if there is a higher than 5% margin of victory, the paper is never even reviewed.

That's how * got his alleged mandate. By shaving votes off even the bluest of states. He couldn't bare to lose the popular vote again IMHO.

corporate news monopolies then CHANGED their exit polls, late on election day, to MATCH the results of Diebold/ES&S's NON-TRANSPARENT, "trade secret," proprietary vote tabulation formulae--that is, they doctored the exit polls, in absurd, impossible ways, to make it come out a Bush win. This is one of the things that TruthIsAll discovered. Exit polls--which are used worldwide to verify elections and check for fraud--were likely a much more accurate gage of the vote than "results" tabulated by Bushite corporations under a veil of secrecy. In other countries, exit poll data is kept separate, specifically to check for discrepancies. Not here. (See **, below.)

TruthIsAll found a wave of exit poll discrepancy (veering off from the non-transparent "official result") that starts on the east coast (highest discrepancy) and peters out on the west coast--indicating that the fraudsters wanted to secure an early victory. The US map of the "red shift"--degree of discrepancy in each state, between the exit polls (likely more accurate) and the Diebold/ES&S vote results (likely tilted to Bush)--shows New York in the gray part of the spectrum (a fairly accurate official vote count), but a big concentration of red-shifted states on the east coast overall, with more balance in the west, of red-shifted vs. accurate.

Very good discussion and maps, here:
http://en.wikipedia.org/wiki/2004_U.S._presidential_election_controversy,_exit_polls

This analysis at wikipedia moreover shows a distinct relationship between the red-shift to Bush and electronic voting. The states where paper ballots predominated yielded more accurate counts (closer to the exit polls) than states where the Bushite corporate election theft machines predominated.*

(I can't help it. Diebold/ES&S's very close ties to the Republican Party, Bush-Cheney and rightwing causes, and their "trade secret" vote tabulation code, often with not even a paper trail--all put into place in 2002-2004 by the two biggest crooks in Congress, Tom Delay and Bob Ney--abetted by Christopher Dodd and by the SILENCE of the Democratic Party leadership--makes me want to scream: WHAT. DID. YOU. EXPECT. ???!!!)

TruthIsAll and others also found the highest discrepancies in battleground states (such as Ohio, Pennsylvania). So what apparently happened is this: Working with the Diebold/ES&S secret formulae, the fraudsters switched or otherwise changed votes to favor Bush--a few percentages here, a few percentages there--at first heavily on the east coast, and then wherever it was easy to do, flipping a 3% Kerry win into a 2.5% Bush win, in the popular vote, and also used this capability to turn the closest battleground states to secure the Electoral vote.

One wonders why they did all that visible, illegal suppression of the Democratic votes in Ohio, and other places, given the electronic control that they had acquired. I think the answer is that, 1) Ohio was a battleground state, essential to the Electoral Vote; there had been resistance to Diebold machines and it was not a fully Diebold state, thus other kinds of Dem vote suppression were needed; 2) the machines had to be pre-programmed in general (not so easy to change on election day), Kerry won by a higher margin than expected, so supplemental Dem vote suppression was needed; and 3) they had to be careful not to tip their hand with overly-dramatic vote stealing; they grabbed small %'s here and there, even in some big blue states with fairly honest election systems, like NY and Calif (Calif no more--Bushite/Diebold shills in charge now), but there was a big amount of uncertainty and iffiness about what could be done without detection--overt vote suppression in Ohio on election day was a contingency plan, or insurance, and, as it turned out, they needed it.

**2004 was the first nationwide testing out of direct fascist control of our elections via electronic voting and its secret code. We caught them, because alert bloggers and analysts took screen shots of the early exit polls, showing a Kerry win that, upon analysis, they realized could not have been turned around without vote rigging. Next time it won't be so easy to catch the fraud. The corporate news monopoly "consortium" exit pollster, Edison-Mitofsky--which has admitted that Kerry won the exit polls--has promised to never let the public see the real exit polls again.

ASK THE DEMOCRATIC PARTY TO FUND *INDEPENDENT EXIT POLLS* THIS YEAR. They owe us this, for their lack of vigilance--and in some cases, corruption--on Bushite-controlled electronic voting and the country's loss of transparent elections.

take a few more years. If it was/is this spring - CONGRATS :toast: - and if not, I am sorry I brought it up!

Both TIA's registered voter use rather than likely voter for pre-election polls and his undecided voter allocation do indeed account for his claim that Kerry led Bush going into the election - and he does ignore the possibility that last minute events could change those pre-election poll results.

No one has disproved the idea that the GOP is just about as corrupt in any given state. I believe that percentage of corrupt GOP in each states GOP party is about the same.

The PDF you link to http://inside.bard.edu/~lindeman/surprise01.pdf
shows that it unlikely that say Florida was more corrupt than Utah - and indeed TIA does assert that it was in the Eastern US that the stealing of the election occurred.

I still think the election was stolen based on my observation of the GOP in a dozen states around and during elections.

You folks did find a "fraud" correlation to the size of the "Black" polulation in any given area - albeit a small correlation, but fitting TIA's thought. So our Southern State GOP are crooks relative to black votes even finds support from your research!

Also,you found, in my opinion, an excellent reason to end absentee voting - and indeed that idea has been suggested all my life by those that have observed GOP election theft.

I found it interesting that there was little attenpt to analyse the vote versus exit poll data on it's own - or did I read the PDF incorrectly? It seemed to consist of comparisons of pre-election poll versus vote and exit poll versus vote data - with little on what the exit poll versus vote can tell us on its own.

so I my defense is postponed indefinitely:(

One of the problems in using the exit poll data to try and pinpoint where fraud occurred is that although there were a very large number of actual respondents in the poll, there are rather few precincts per state ( 30-40, typically). And the relevant unit of analysis for vote-stealing fraud is the precinct, nested within county, or state. So even when we look at precincts within states we don't have much statistical power, and we certainly don't have any county-level power - typically fewer than one precinct per county is sampled, so many counties go unsampled.

However, looking at the precincts nationwide there is a fair bit of statistical power - 1250 precincts in which the proportion of rolled-in absentee ballots was low enough to make analysis possible (and I do agree with Kathy Dopp that breaking out results by absentee, early and on-the-day is an important bit of transparency that would go a long way to ensuring honest county, as is the ability to check precinct counts against county tabulations). The trouble is that, as I keep on saying, examination of the relationship between the amount of redshift at precinct level and the amount of "redswing" since 2000 would seem to indicate that there is no discernable relationship. If fraud was responsible for redshift, it doesn't seem to have translated into any discernable benefit to Bush!

In fact, in many ways, Bush's 2000 vote share at precinct level is a much better baseline for fraud than the exit poll anyway. Bush's 2004 counted vote share is actually more highly correlated with his 2000 counted vote share than with his 2004 exit poll share (i.e. there is more shared variance - that doesn't mean there wasn't both redshift and redswing - there was both). If people really want to find out whether the election was stolen on DREs, then the obvious way to do it is to obtain 2000 voteshare data (or 2002 data) for the precincts in question, and test out whether swing is greater in precincts using DREs (or using equipment by a particular manufacturer; or any other factor hypothesised to be associated with fraud). That could be easily done at both county and state level, because there is colossal amount data (far more than in the exit poll data) and, unlike the exit poll data, it is public data . This, of course, is what Herron and Mebane did in Ohio.

http://macht.arts.cornell.edu/wrm1/Ohio2004/Ohio_precincts.pdf

Swing analysis won't of course show up inequities that were also present in the reference year, but it should certainly show up theft of any new magnitude, associated with new technology, for example.

The trouble, as far as I am concerned, is that while there are plenty of a priori reasons to be suspicious of the election, the only a priori reason to postulate massive vote-theft is the exit poll data - and my view, after fairly extensive consideration (!) is that the exit poll data, if anything, suggests that massive vote-switching fraud did not occur. But if anyone would like to look at the swing data from 2000 they might be able to prove me wrong.

So, as a psychologist (getting close to a PhD!) I would say in answer to BeFree's OP that the reasons perhaps that there is resistance to the idea that the election was stolen is simply that the evidence is not as compelling as I admit it looked to me in November 2002. I know that many here still find it compelling - but the more closely I have looked at the numbers, the more I have had to confront my own "cognitive dissonance" and come to the depressing conclusion that more voters voted for Bush than Kerry (although not as many more as would have liked to vote for Kerry). Although the good news, if you believe me (and I realise some don't) is that it means that there is still hope for next time!

But can I commend, as ever, Shanikka's great DKos rant:

http://www.dailykos.com/story/2004/12/15/9734/4841

Never mind who lost, look at who was disenfranchised!

and I have my own "cognitive dissonance" that prevents me from easily coming to the depressing conclusion of a no fraud Bush election - indeed I find the fraud everywhere idea much more likely - both in 2000 and 2004.

I think even your analysis showing the black correlation makes more likely the assertion that the Ohio fight is about African-American voting rights and not about John Kerry. And Shanikka's great DKos rant on this has only inspired 623 responses so far!

http://www.dailykos.com/story/2004/12/15/9734/4841

My take away from your post is that the lack of precincts per state (30-40) in the exit poll makes statistical conclusions by state or county impossible. My take away is not that we can not do a by machine analysis, or a by precinct controlled by GOP either directly or by a Sec of State running the election. And those require the data to be released by the exit polling firm.

The idea that the GOP IS UNIFORMLY CORRUPT IN EVERY STATE works for me! :-) And the analysis so far has not disproved that - although as done so far the analysis can not prove it either.

to a DKos diary is huge. It made a very great impression! But it's archived now, although I keep it alive by frequent links!

I am not quite sure what you mean by "My take away is not that we can not do a by machine analysis, or a by precinct controlled by GOP either directly or by a Sec of State running the election. And those require the data to be released by the exit polling firm."

You don't need exit poll data to analyse swing by machine, or by county control, or by any factor that you hypothesis may have been associated with fraud. Precinct level swing is a very powerful measure of the expected value of a candidate's vote-share in a precinct, and if it is greater where certain factors you suspect would contribute to fraud are present, then they will show up when you correlate those factors with swing.

As for the exit poll data - sorry to keep saying this, but the data ARE released. What is not released is precinct identifiers. This makes them of limited use for your purposes, but is essential if the confidentiality of the respondents is not to be violated. People assume that this is just an excuse but it isn't. In my own field we have to take extraordinary care to ensure that participants cannot be matched to their responses. But my point is that far better data with which to investigate fraud 2004 is actually public, namely the precinct vote counts for each precinct for 2004 and previous elections. Not being an American I don't know where you get it from, but it must be available, even if you need to invoke the FOIA to get at it.

And as an actuary, you will know what to do with it! Instead of "following the money" - follow the swing!

This uses TIA's data from this post:

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=330449&mesg_id=330449

I have ranked the states from left to right by magnitude of exit poll discrepancy, and what is plotted is Kerry's percentage of the two party vote.

The blue triangles represent his exit poll percentage,and the red squares are his counted percentage. Generally the blue triangles sit above the red squares, showing that his counted percentage was lower than his exit poll percentage. Also, generally, the green diamonds, representing TIA's pre-election estimates are generally higher than the counted percentage (TIA, as we all know, predicted a Kerry win).

If you look at the left of the plot, where the exit poll discrepancies are greatest (and presumably fraud is postulated to be more likely) the curious thing is that Kerry often does as well or even better in the actual count than predicted by TIA's model. This is true of the four states with the greatest redshift in the exit poll (VT, NH, NY, DE). As we move to the middle of the plot, TIA's projections more often coincide with the exit poll than with the count, but by this stage we are getting to be within the MoE of both sets of polls. Nonetheless, both exit poll and TIA's predictions tend to be slightly higher than the count. However, at the right-hand end of the plot, where the shift in the exit poll is actually blue (Kerry does better in the count than in the exit poll) TIA's predictions are for Kerry to do even better still. It would appear, therefore, that if the exit polls for these states are indicating either no fraud or even pro-Kerry fraud, TIA's model is over-optimistic for Kerry. On the other hand, if TIA's model suggests fraud in these states, then the exit polls would seem have an even more markedly pro-Bush bias.

A more parsimonious explanation would seem to me to be that while the exit poll tended to over-estimate Kerry's voteshare in blue states, TIA's model tended to over-estimate Kerry's vote-share in red states. So I plotted this here:




And it seems to be the case.

I do find it hard to devise a fraud narrative that would account for these kinds of pattern. I keep coming back to the point: if polls (pre, post) and swing are all "redshifted" (which they seem to be) and the redshift in each case has a common cause (fraud) then why are the distributions of shift in each case uncorrelated with each other? Why is swing not greater where exit poll redshift is greater? Why is redshift from the exit poll not greater where redshift in the pre-election poll is greater?

The simplest explanation seems to me to be that there was bias in the polls, which is perfectly precedented.

Which is not, of course, to say that there was not also some fraud in the count (or voter suppression, which might show up in the pre-election poll discrepancy but not in the exit poll discrepancy), but it is to say that the redshift in the polls does not give us anything like an estimate of its magnitude, and may be to say that its magnitude is unlikely to be in the order of millions of votes.


edited to fix typo

DC is not shown on the scatterplot as it has an outlying Kerry proportion.

What you can do is say that fraud was ubiquitous and uniform and that polling error was sporadic and highly variable. That way the variation seen in the plots is caused by polling error while the shifting of the entire plot off center is caused by fraud. In other words, the pre-fraud plot looked (the premise says) just like the final one but was centered at zero. Fraud lowered the entire plot below zero without changing the ups and downs of it, sort of like if you shifted the axis without changing any individual data points.

Of course, my narrative doesn't go so far as to explain how fraud like that could have been implemented but maybe we can defer that for a moment and see if we agree at the theoretical level.

I've just been setting up a simulation of the swing-shift plot, where I can manipulate the variance in fraud (also variance in swing). The value of of the fraud-induced correlation in the swing-shift plot depends on the variance as well as the magnitude of the fraud. I can adjust fraud variance by making it more or less ubiquitous and/or by increasing the distribution of the magnitude.

My output tells me how many votes I've stolen.

So sure: if we have absolute control over all counts used to give the vote totals to E-M on election night (60% at the precinct, 40% from county tabulations), we can do it.

edited to add inexplicably missing words

because the real-world implementation is where it gets tricky.

I think there is one more theoretical point to consider first, which is the following. We agree you could do it if you have total control and can make the fraud uniform and ubiquitous. We also agree you can't do it (without creating correlation) if your fraud mechanism is localized and variable. My current point is that there is some wiggle room in between the two extremes. How much can you depart from ubiquitous and uniform without creating detectable correlation? How much variability in polling error do you need to cause it to overwhelm and disappear whatever variability you have in fraud? If you've got two knobs in your simulator, how far up can you turn the fraud variation knob and how far down can you turn the polling error variation knob before fraud variation becomes detectable correlation?

The point of all that is to allow the mastermind of my implementation (see photo upthread) to not have omnipotence (although reports indicate he may have).

So now to implementation. I guess the best bet would involve a fraud algorithm built in to central tabulation software. Without the wiggle room I mentioned then you would need to have that algorithm virtually everywhere. Maybe the wiggle room lets you have the algorithm in some number of places that is less than everywhere, actual number TBD. The algorithm would pad by a percent, either pre-determined or let's say adjusted realtime so we can say that is what the mastermind is working on in the photo. Does padding by a percent get us where we need to be or do we need some other algorithm (I know you had some thoughts on this before).

And there are a few other things we can look at like how we define "swing" and therefore "uniform". I've got to get on with some other stuff right now, but what I am doing is assuming that what might be uniform about fraud might also be uniform about genuine swing e.g. uniform defection rates; uniform rate of Kerry vote flipping. Both of these would induce greater percentage point swing to Bush in high Gore precincts (more Gore voters to defect; more Kerry votes to flip) than in high Bush precincts, just as genuine swing to Kerry or pro-Kerry fraud would produce a greater percentage point swing to Kerry in high Bush 2000 precincts than in high Kerry 2000 precincts.

And there are other knobs to turn, like the amount of unsigned non-sampling error in the poll, and simply the amount of unsigned genuine swing. We do have rather too many free parameters, so I'm trying to ball-park some constraints.

See you later!

Lizzie

...TIA's analysis of the Miscountski report that shows the method used to establish the final exit-poll numbers were totally bogus?

a conversation in a blizzard.

The numbers I am talking about are not the "final exit poll numbers" which were indeed weighted to the vote returns, as they announced they would do in advance, as they do every election, and as has contributed to their reputation for "uncanny accuracy".

What I am talking about are the raw, precinct exit poll returns, before any kind of processing whatever has been applied. These numbers show that the actual raw returns, when compared with the precinct count made mostly at the precinct (60%) but in some cases from the county tabulations (40%) were substantially discrepant from the count, with discrepancies in both directions, but with the mean discrepancy substantially "redshifted" ie. on average, the percentage of poll responses for Kerry was higher than the percentage of Kerry votes in the count.

Two hypotheses could account for this: bias in the sampling of voters or bias in the count of votes, and I am trying to distinguish between these two. We cannot do this by looking at the "final exit poll", and I am not even looking at the "final exit poll" numbers. I am looking at the correlation between the precinct level discrepancy between the two sets of numbers (raw poll; precinct count) and the degree to which Bush's vote share changed between 2000 and 2004 in that precinct.

OK?

Release the precincts that were polled for establishing the exit-poll data. Then we can all work together to find the answers that have escaped you for 15 months.

OK?

As you know, for reasons I have given, that is not either possible or ethical.

If you want to help, then help to develop testable hypotheses, as eomer is doing, as I have been doing, as many on this forum have been doing including OTOH, but to do that you will have to find out a little more than you seem to know about the nature of the data.

Yes, that's right, and we will never know the true nature if E/M don't release the data.

The data is a secret, I got that. And as long as it is a secret, we will never know. That's the point.

There are trusted people that the data could be released too, eh?

I take it that you are able to access the data? Well, just have our people sign a confidentiallity agreement that keeps personal info a secret, and let us get on with a full and open, bi-partisan examination of the data.

Or is that a problem?

like the difference between projections, final projections, precinct level data, etc.

BeFree, you seem determined to bug me. I am doing my best to find out what happened. If you want to snipe at everything I say, then frankly it won't do either of us any good.

I was contracted to do some work for Mitofsky on the data because he wanted to know what happened, and, as it happened, so did I. I happen to be a good data analyst who posted some critical things about the E-M evaluation, which was why I got hired. I also happen to post on DU. It's probably as close as you are going to get to the data unless you too manage to hire yourself to Mitofsky, in which case you will also be bound by conditions of confidentiality. It is simply the way it is with confidential data.

What I can do is explain what I tested and what I found. I'm trying to do that.

If you don't like it, feel free to ignore my posts. I think some people find them informative, but of course I may be wrong. Nevertheless, they are informative, and you can choose whether or not to find them so.

cheers

Lizzie

Boy, you've got a huge range on the vertical axis there, makes it hard to tell a small exit poll shift from a large one.

One problem I have with this post is that you presume that there is not a diverse, 50 state (or less) strategy that uses techniques that differ (based on different voting machine systems, different access points, etc) and therefore affect the variables you cite differently.

is directly linked to from the post, if you want the numbers. The point of the graphic was to show the relative relationship between the three stabs at who people voted for, and I ranked by discrepancy so you know the larger exit poll discrepancy is always on the left. I could have plotted them to the same baseline, but then you'd have lost the between-state variance in vote-share, which is fairly informative. Whatever.

But to take your second point - I do not presume anything. Eomer has the interesting idea that if the strategy is uniform across states, it won't show up in the swing-shift correlation. So I have tried to model that assumption, although I think it is unlikely, as does eomer, which is why he wants to see how much wiggle room there is for variance.

The answer seems to be not a lot. The more variance there is in the mechanisms and/or magnitude of fraud, the less likely is it to fail to show up as a positive correlation between swing and shift.

Except that I'm not sure this is your point - in what context are you assuming that I am presuming anything? If you are simply referring to the plot in this post, all I am saying is that the degree to which the exit poll was discrepant from the count (which may reflect fraud) seems to be uncorrelated (actually slightly negatively correlated) with the degree to which TIA's pre-election polls were discrepant from the count (which might also indicate fraud). If fraud is responsible for both sets of discrepancies, it is odd, to say the least, that they go in opposite directions, whatever the method of fraud used. One might expect that in states with little fraud, both pre-election and exit polls would be closer to the count than in states where there was a lot, where one might expect both pre-election and exit polls to indicate a higher Kerry vote-share than the count. But this is not the case.

Why do you think that I presume uniformity?

Thanks Febble. I hadn't seen that before.

I'm amazed Mississippi has such a high minus. Seems impossible. The Mississippi vote margin was almost identical to projection.

do not include pre-election expectations. The discrepancy for the Composite estimate for Missippi was -7.4.

The best Geo is an interesting one to look at, as it incorporates weighting by characteristics of non-responder but not by either pre-election polls or, of course, vote-returns. The fact that the best Geo discrepancies are considerably greater than the composite estimate discrepancies is also interesting, as it tells us that even before the vote-counts were coming in, E-M would have been noting potential problems with the poll, i.e that there were conflicts with their pre-election expectations.

Thanks for the summary. I only glanced at Best Geo in E-M pdfs I saved from early 2005 and didn't understand the numerical relationship until today

Some people are stuck on the idea that the tabulations posted on CNN.com (or even the leaked tabs) are the 'real exit poll results.' But they all incorporate prior expectations and/or some vote counts. And the irony is that the exit poll results are even worse.

More on the other branch tomorrow.

Hope he's healing up.

My thoughts are with him, and hope he soon returns to health.

Seeing bush go down because of election fraud would be good medicine, eh?

and that on both counts are equally my greatest fervant hopes.

abominable in the minds of citizens that they are incapable of believing that it is true.

We've got a pretty long and officially recognized history of fraudulant elections, don't we? I mean, especially on the local level, but in the late 19th century on the national level too. Wasn't the 1876 election so manifestly fraudulant that the states just punted it to the House? Or what about the Chicago "machine" turning Illinois greatly more pro-Kennedy than polling suggested? Why wouldn't people believe it can still happen?

It occurs when there is a need to accommodate new ideas.

Cognitive dissonance is a psychological phenomenon which refers to the discomfort felt at a discrepancy between what you already know or believe, and new information or interpretation.

If someone is called upon to learn something which contradicts what they already think they know — particularly if they are committed to that prior knowledge — they are likely to resist the new learning.

(Some day psychologists and political science majors are going to collaborate, trying to figure out how a dictatorship came to America.)

cognitive dissonance which leads to outrage which leads to action.

An aside:
"Some day psychologists and political science majors are going to collaborate..."
Isn't this what got us into dictatorship in the first place???

by psychologists. How to manipulate people into buying something.
The politicians are sold like products.

Yep you are right. Those damn shrinks :(

I don't know of any democrats, or very many who think the election was legitimate.

Many things, the machines, the crazy antics of Kenneth Blackwell
to supress the vote,

the missing thousands of absentee ballots in Florida.

But with paperless machines, or without audits, with control of the courts by
the very administration fraudulently elected,
how do you legally change things?

I believe that if this 2006 election doesn't change the makeup of
Washington seriously, then you can expect some people to abandon
legal recourse and the demonstrating, the out right boycotting,
the work stoppages, etc will become rampant.

More so for 2008.

Pretty much everyone has heard about 2004 over and over for the past 2 years,
and most have read all of the books about what happened.

I darn sure want to prevent another election like 04 and 00.

realizations that are corollaries to the fraudulent election of 2004 are very heavy, and they are the complete failure of all of the political institutions that we depend upon, including,

a) state/local election officials (corrupted very fast, in the 2002-2004 period, by the $4 billion HAVA boondoggle that Tom Delay, Bob Ney and Christopher Dodd arranged for, and by lavish lobbying by the Bushite electronic voting corporations);

b) news organizations (war profiteering corporate news monopolies, all of them)--actual collusion in the stolen election, with their falsification of their exit polls--and the exit pollster itself (Edison-Mitofsky);

c) the Democratic Party leadership--the most shocking failure of all (combo of corruption, fear, collusion on electronics in gov't, support for Bush's war, and I don't what-all...insanity?);

d) the courts, the legal system--should never have permitted "trade secret" vote tabulation; how did this get legal?;

e) Republican Party fairness (I mean, some believed in it--that they were still loyal, patriotic citizens of the US, believers in fairness and good government; true at one time, I guess; Bushite fascist coup now);

f) Congress, as a "check and balance" in our overall political system--even if full of Bush "pod people," you'd think that "trade secret" vote tabulation and no paper trail wouldn't fly. But it did.

-------

No one is surprised at Bush rigged elections. People expect it of that crowd of swindlers. What they can't take in is all the other failures.

I have an intelligent and I thought well-informed friend, to whom I described the facts about our election system and its takeover by Bushite corporations using "trade secret," proprietary software and no paper trail, and here was her reply...

"But the Democrats wouldn't let that happen, would they?"

The issue is not "Would they?" The issue is "They DID." And we need to ask why. But it's very hard for people to absorb this and the grand failure of everything.

That's what we must face. It's been building up for a long time--especially with more and more corporate/&corporate news monopoly influence on elections, and on Dem leaders, since Reagan, crescendoing in this junta, this fraudulent fascist regime, which is stealing the country blind to lard the super-rich, has slaughtered tens of thousands of innocent people, and tortured many more, and is absolutely dismantling good government and the Constitution.

The only way to reverse it is to get transparent elections back--because, you know, it's not just Repubs/Bushites; a lot of Dems are beholden to Diebold/ES&S as well. You could probably pick them out by their support for Bush appointments, and maybe their votes for Bush's war. You wonder why 60% to 70% of the American people--those opposed to the Iraq war, torture, the $10 trillion deficit, larding the rich, etc.--are represented by only about 25% of the Congress? Think about it.

I had previously talked about it only with family. These peoples' eyes bugged out when I told them a few of the facts TIA points out. I also said that the USA is the only major democracy in the world with vapor-voting (electronic) machines, and that the largest democracy (India) uses paper and gets quick reliable results. I think they were a bit rattled.

the biggest hurdle remaining.

This is something that Mark Crispin Miller mentions and I think it's a key element in the suppression of the truth.

To admit or suggest that such a thing could happen around the country is to endanger your position on the newspaper, as a political science teacher at a university, as the head of polling company, as an elected (supposedly) rep of your state or region, etc., etc., etc. People just can't really take the chance these days. What would happen if they lost their jobs? It's a thin rope a lot of people are holding to keep their heads above water. Some of it is just fear as well, fear of being thought of as a "conspiracy theorist" or scoffed at by the loud-mouthed minority trained and deluded by Pox News etc.

- Under what conditions do millions of allegedly "free" people knowingly acquiesce to being deceived, dominated and deprived of their own political will?

- How is it that even those who were politically engaged for the first time resign themselves to an unjust fate, refusing even to consider what happened to our country?

- Why do progressive citizens actively dismiss and even malign a small group of courageous, devoted people working day and night on their behalf to uncover, calculate, analyze, and evaluate the extensive, varied forms of criminal sabotage that undermined their democracy?

- How are Americans becoming complacent with escalating fraudulent activity? In other words, how do so many people live with the knowledge that they have been tricked before, were just tricked again--and then submit to life under the power of those who tricked them?

questions worth pondering for sure . . . and answers worth pursuing . . .

is our "psychological resistance" acquiescence to fascism? . . .

However, I decline to place myself in the "we" column and do believe that the evidence is more than sufficient to show rampant fraud and without active grassroots work (and sadly, possibly even with it) the next election will be a "surprise" win to the Thugs again.

many people have a hard time accepting that the 'national parent figure' (government, those who are in power) can be that abusive toward its 'children' (the people, us).

It is like the beaten wife who stays with the wife beater - pleading, begging, trying to reason with the abuser, in hopes that things will change for the better.

"If people knew about the amount and extent of dirty tricks, 3,500,000, or 1,750,000 may not seem so insurmountable."
It didn't help Gore.

not the key reason, but it doesn't help

Many people, even highly educated people, don't understand the concept of statistical sampling - how you can sample 200,000,000 people by calling only about 1100. "no one has ever polled me, so it must be wrong".

But then again, if they did understand statistics, the gambling and lottery industries would fall apart, so.....

IMO is that of those who do understand the concept of statistical sampling, few understand the concept of random sampling, and the ways in which polls always violate the random sampling assumption.

Polls are not like casino games. Getting a true random sample from 200,000,000 people is impossible, although using random dialing helps. Getting a true random sample of 120,000,000 voters in person is even more fraught with problems, and both forms of sampling are vulnerable "non-sampling error" including response bias and selection bias.

People, unfortunately for pollsters, do not behave like roulette wheels, and there is no reason to suppose that the way they behave in regard to pollsters is unrelated to the way they vote.

can this not be factored into the design of the weighting via an overweight?

I see the feedback loop, but I find it hard to see why there is no solution. I really hope the adjustment calculates out as a infinite series whose sum approaches some limit.

it is not necessarily constant from election to election, although it tends to go in the same direction each year (and in 1992 was nearly as great as in 2004). What seems clear is that while there is an underlying trend from year to year for Democratic voters to be better represented in the exit poll than Republican voters (I say this knowing the risk of being flamed, but contrary to frequent claims, there is actually good evidence for this), where polling protocol is good, the effect is minimized. In other words, given good conditions, training, close adherence to random sampling protocol etc, Republicans have less opportunity to evade the pollsters, and/or Democrats have less opportunity to increase their chances of selection. The trouble is that the factors likely to provide opportunities for the underlying phenomenon to affect the sampling are only known after the election.

In fact, of course, you have put your finger on what actually happens, which is that the pollsters expect bias, and use the vote-count itself as the "overweight". As the vote-count comes in, they spot the what they assume is bias, and weight accordingly. The trouble here is that the procedure assumes the vote-count is correct! But this is at the root of a lot of misunderstanding of the exit poll. It is not designed as a check on the count - it is designed to project the counted result. So the pollsters, while they try to get the interviewing right, are not too worried about bias (unlike pre-election telephone interviewers who are trained interviewers and collect a much smaller sample over a longer period, the election day interviewers are casual hires with a short telephone training) because the intention is always to weight any observed bias by the vote-count itself. What they are really worried about is projecting a result that results in a "call" for a state that turns out to be "incorrect" (i.e. not in accord with the final counted result).

If you wanted to design an exit poll as a check the count (which would be prohibitively expensive, which is why they don't do it) you would have to have teams of trained interviewers rigidly sticking to Nth voter protocol throughout election day, and even then your results would be vulnerable to non-response bias (more refusals from voters for one candidate than the other). Ironically, ploys to improve response rate (e.g. giving away free folders) don't necessarily eliminate bias and may even increase it.


ETA: results are also weighted for the observed age/race/sex of non-responders, but this won't help with selection bias, only with non-response bias. Pre-election polls are also used in the weightings - if exit poll tallies are out of whack with pre-election polls, that is another indication that bias may be present in the exit poll sample, and they can be weighted accordingly.

the answer to the question "Was the election stolen".

We are left with our gut feeling that while non-response/shy voter does exist, the size of that non-response needed to get from exit poll to official results is unreasonable, plus the knowledge that the hispanic result is contradicted by other exit polls plus the knowledge that even in the "official exit poll analysis" there are indications that blacks are getting screwed.

It will be interesting to see if election theft becomes an issue in 06.

designed to keep the 2008 election honest? Construct the model and accompanying polling protocols, procedures, and training manuals in open forum and vetted by your statistician peers. Concurrent with your effort, we all campaign and lobby the League of Women Voters and other non-partisan organizations (and perhaps multiple-partisan organizations for inclusiveness) to amass the necessary volunteers, fund the training and infrastructure, and provide the logistics well in advance of the election. Call it Poll America 2008.

I know you want fair and accurate elections for the U.S.. This is a way for you to apply your expertise and talents to achieve that goal.

for the past 18 months?

Just because my checks didn't find the fraud, doesn't mean they weren't good checks. Maybe they were very good checks and there wasn't much fraud, huh?

But I simply don't recommend polls as a way to check on the count. There are far better ways. You can never rule out bias in a poll, however well it is run, and to run one of the quality of the BYU poll (which is an excellent model) nationwide would cost several times more than the E-M poll which is funded by several wealthy TV networks.

If you want to regard exit polls as a check on the count, then you need to set up testable hypotheses (as I've been doing) to disambiguate bias from fraud. But it's all correlational, and not conclusive, plus it has low statistical power at precinct level.

Much better to go for random audits; close bipartisan supervision of the custody of the ballots; checks on county tabulations broken out by election day voting, early voting and absentee voting, with canvassing of precinct counts; correlation between of "swing" from previous elections and any change factors that make you suspicious.

I'm sure others in this forum have more ideas than I do - my ideas on voting are very simplistic, living as I do in a backward nation that uses paper and pencil and people.

And then there is voter suppression to tackle, which, if anything cost Kerry the presidency, IMO it was that.

exit poll against which to check the official tally within any reasonable and effective MoE. This will come as a surprise to the Ukrainians who lacked your insight (though perhaps the attempted assassination of the initially losing contender put the people on the right track in spite of exit polls) and is too bad for us given we have little hope of influencing election processes (or having access to data) sufficiently to pursue your alternatives.

In conclusion, all I can say is: Bring back #2 pencils to our voting and human eyes to our counting!

about 2B pencils, but re Ukraine: yes, as you will be aware, there was plenty of other evidence of shenanigans in Ukraine, like pouring acid into the ballot boxes.

I know it is generally believed that the exit polls were the tip-off, but as far as I can gather, that is not actually true (though it is true that they weren't accurate, but then, if an election is rigged, exit polls will only ever be "accurate" by accident - it doesn't mean that exit polls for an unrigged election will necessarily be "accurate".

I am not sure just how snarky you are being here. It may well be that Ukrainians lacked my insight, but I also lacked some of my current insight when I was suspicious of the US 2004 election, although my experience in the UK certainly didn't lead me to believe that exit polls were particularly accurate. I watched in dismay in 1992 as Major, contrary to both pre election and exit poll predictions, emerged with a working majority (and then held to power for five more miserable years).

Yes, I believe it is impossible to calculate a true MoE for a poll. You can give an MoE on the assumption of a random sample - what you can't do is to assume your sample is random, because it never is. Exit polls are an extremely expensive way to get an unreliable estimate of who got what share of the vote, for the simple reason that they are vulnerable to bias. Pollsters are obsessed with forms of bias precisely because it is such an intractable problem. If you are relying even on a good exit poll to tell you your election is rigged, something is desperately wrong with your democracy.

Let us agree on 2B pencils.

One had Yushchenko +9, the other Yushchenko +1. That's a bigger difference than between the U.S. exit poll and the official count.

As you said, there was lots of other evidence. The OSCE report on Ukraine didn't even mention the exit polls.

You can thank all of these Democrats pushing paperless voting, and pushing Diebold.

“Dear Colleague,” Rep. Robert Ney, Rep. Steny Hoyer (D), Sen. Mitch McConnell, Sen. Chris Dodd (D), March 3, 2004.
They warned that amending HAVA to require VVPB would cause
"numerous adverse unintended consequences. Most importantly, the proposals requiring a voter-verified paper record would force voters with disabilities to go back to using ballots that provide neither privacy nor independence, thereby subverting a hallmark of the HAVA legislation.”
www.house.gov/cha/dearcolleaguemarch3-04.htm

Georgia's Secretary of State, Cathy Cox, (Democrat) could easily be considered the official Poster Girl for Diebold Election Systems, Inc. ...
www.countthevote.org/cathy_diebold.htm


Maryland's Secretary of State, Linda Lamone (Democrat) is a staunch defender of Diebold:
Maryland is of national interest because Lamone is the President of the National Association of
State Election Directors (NASED) and the most vociferous advocate for paperless voting in the United States.
http://www.americanchronicle.com/articles/viewArticle.asp?articleID=5932

Joe Andrew, former head of Democratic National Committee became a PR person for Diebold
in August of 2005.
http://www.zoominfo.com/Search/PersonDetail.aspx?PersonID=714249

Joe Andrew is a lawyer, former election official and former National Chair of the
Democratic National Committee. He advises Diebold Election Systems on voting industry standards.
http://tinyurl.com/zy6ys (cache of Oakland Tribune )

Joseph J. Andrew, who was head of the Democratic National Committee from 1999 through 2001,
is now lobbying for Diebold, where he is demonstrating that he is even more mendacious than
the veracity-challenged Mark Radke, the PSI (damage control) firm's David Bear,
the protective Marvin Singleton, and the script-reading Joe Richardson.
http://en.wikipedia.org/wiki/Joe_Andrew

08/20/2005
Diebold hires top Dem for PR blitz
Former party chairman make the case for voting to California
With a phone call and a retainer, Diebold CEO Walden O'Dell has launched former Democratic National Committee
chairman Joe Andrew on a 50-state ambassadorship for electronic voting."
http://www.wheresthepaper.org/InsideBayArea08_20DieboldHiresTopDemForPRBlitz.htm

Diebold needs California - which is why they've hired Andrews to get it for them. "
California, the nation's largest market for voting machines and the place where
Diebold's fortunes as the largest supplier of e-voting machines in the nation could be made or broken".


Harris Miller, running for US Senate as Virginia Democrat,
As head of the ITAA, Miller specifically lobbied Congress against verified voting, on behalf of the
interests of Diebold and other manufacturers of paperless e-voting machines -- members of ITAA.

The ITAA PAC, of which Miller was President, Treasurer, and Top Donor, donated 89% of its PAC Income to Republicans between 1997 and 2000.
http://www.zoominfo.com/Search/PersonDetail.aspx?PersonID=268428

The vendor community doesn't like it. "We oppose the idea of a voter-verified paper trail," says Harris Miller, president of the trade group Information Technology Association of America. Introducing paper into the mix, he says, defeats the improved efficiency and reliability e-voting promises. "There was never a golden age when paper ballots were accurately counted," Miller says. Adding paper to e-voting will only make the process of administering elections more costly and time-consuming without improving accuracy, opponents assert. http://www.cio.com.au/index.php/id;558873322;fp;4;fpid;21

More about Harris and Diebold, as well as other things not so positive here:
http://www.sctnominationwatch.tpmcafe.com/story/2006/1/6/23532/17147

Steve Metcalf, former democratic NC state senator turned Diebold lobbyist.
Diebold Election Systems, Inc. (Resigned 3/27/2006)
http://www.secretary.state.nc.us/lobbyists/Lobbyist.aspx?PId=5080

George Nixon Gilbert, democrat and Director of Guilford County NC Elections:
12/24/2005 High Point Enterprise. Company Root of Controversy- by David Nivens, staff writer.
Gilbert calls the Diebold critics “fanatical.”
“These machines are used in states from Georgia to California,” Gilbert said.
“Diebolds are just as reliable as other machines. They all use the same software and they all have software errors.”

A SCENARIO (HERD MENTALITY)

If you believe the 04 election was stolen you must be a "conspiracy nut." Do you think there was a conspiracy?

Let me answer your question with a question. Was the Jim Crow period a "conspiracy"?

That's not what I mean.

No I think that's exactly what you mean. Why don't people speak up about this electronic voting curse? It's not because they're all in formal cahoots. It's because of a whole range of psychological factors.

But that begs the question. You have to believe there's a conspiracy because of your obsession with this issue.

Yes, I believe there's a conspiracy, just as there were certain people who consciously tried to manipulate public opinion so as to maintain the inferior status of blacks during Jim Crow. That number was a very small number but thru coercion, literal or indirect bribery, cajolery at all levels, etc. this small number of people were able to maintain the status quo.

etc. You can carry on the conversation. This supports the herd idea in the original post and I think Crispin Miller suggested this idea. The analogy, like all analogies, isn't exact but it's close. Money is the big difference I think. The vendors are throwing around huge amts of money and I don't think that was such an integral part of Jim Crow, but the psychological aspect is almost exactly the same. Why is Motifsky voicing his theory about the Reluctant Repub Responder? Because he has been convinced by somebody or other that the election couldn't have happened all over the country as it did. It would have involved too many people and somebody would have spilled the beans.

But there are no beans to spill now. They've all been spilled. We know the machines are eminently hackable, they're used all over the country, exactly the same machines. If they haven't been hacked, if they weren't hacked in 04, then there's no such thing as "human nature." Given the opportunity and the means someone will always steal an election. What we have now in the US is the perfect storm for stolen elections.

You give quite the wrong answer to your very loaded (misspelling noted) question.

The "theory" is "voiced" because it is supported by rather good evidence, whilst the fraud theory is not only non supported by the evidence, but, if anything, is contra-indicated by the evidence.

I know this is hard to get your head round, but continuing to assert that Mitofsky dreamt up non-response bias because someone sat on his head and told him there couldn't be fraud in an American election is bullshit. Mitofsky is a pollster of vast experience and statistical expertise, and not only did he have a good hunch as to the likely cause of the discrepancy (non-response and selection bias are extremely well-established phenomena), he went to a lot of trouble to check it out. And, having done that, and published his findings, he then went to a whole lot more trouble to redo it because a poster on a leftwing blog (DKos) thought he'd done it wrong. Damn it, he hired the DKos poster to re do it. And she found the same thing.

However, not content with that, she checked out some other stuff while she was there, in the hope of finding out that maybe fraud played a major role as well as bias in the sampling. And when the data failed to support that hypothesis, she kept on going, and posting on DU, and soliciting ideas, in the hope that there might be something she'd overlooked.

And yet not only did she continue to find that the fraud hypothesis was not supported, she found that it became very difficult to formulate any pattern or mechanism for widespread fraud that was consistent with the data. Nonetheless she spent the last weekend modelling various scenarios in the hope of turning something up. Nada.

But watch this space, you never know. She never gives up.

Except that it doesn't matter because whether the election 2004 was or was not stolen by digital fraud makes no difference to the case that it could happen and that it must never be allowed to happen.

And don't forget voter suppression,which most certainly did happen.

Cheers

Lizzie

I am just absolutely rolling my eyes at the notion (expressed upthread) that political scientists are suppressing the evidence of a stolen election because they are afraid of losing their jobs. Mark Crispin Miller takes time out from his national book tour to accuse an entire profession of careerism? maybe I have given him too much credit. (The truly weird thing about this is that making a persuasive case for massive fraud would be more likely to make than to break someone's career in political science.) And then, here, the assertion that Mitofsky or others just assume that fraud is inconceivable -- truly wacky.

I would only add that if people want to avoid relying on Edison/Mitofsky for data -- even the evaluation report -- they can still partially replicate a lot of your non-findings. For instance, Charles Stewart's "addendum" paper used a bunch of different publicly available exit poll red shift measures to establish that at the state level, there was no apparent correlation between DRE use (or use of any other technology) and exit poll red shift.

http://www.vote.caltech.edu/media/documents/Addendum_Voting_Machines_Bush_Vote.pdf

One could use these measures (or do a painstaking compilation from exitpollz.org, or whatever) to verify that Bush didn't generally do better compared to pre-election expectations in red shift states compared with blue shift states. And so on.

If 2 million new voters voted for Kerry, and 2 million Kerry voters were illegally purged,
how would that show up in the numbers, to anyone who was analyzing them?

Can you say 'elephant in the room'.

We don't know how many regs were illegally purged or not entered in Ohio. I am skeptical of the highest estimates because the DNC's voter experience survey didn't indicate a high turnaway rate -- but that survey didn't have much statistical power. It does provide some reason to think that purges didn't alter the result. (I wanted to leave that sentence out, but my conscience wouldn't let me.) We should be interested in documenting and preventing registration purges regardless of whether they appear to be decisive.

Illegal purges shouldn't much affect the exit polls (except via uncounted provisional ballots), which may make them more credible given some of what we've learned about the exit polls.

Conceivably new voters could mask the influence of purges, although that is problematic because it doesn't appear that new voters in most places voted sharply, sharply for Kerry. In heavily Democratic precincts, lots of votes could disappear without affecting percentages very much.

By the looks of this survey, the sample size from Cuyahoga county could be as small as 33 respondents (57% is only divisible by 3).
They don't appear to give much information on specific methodology.
As far as I can tell, this survey could be severely watered down with respondents outside the problem areas.

There must be other surveys on this.
If not...have any idea what one would cost?

I agree with you that a sample of about 1200 Ohio voters isn't much basis for reliable inference about a phenomenon that may have been intensely concentrated. (I doubt that the Cuya sample was as small as 33 respondents. Most likely it was roughly proportional to population size, which would put it somewhere over 100 respondents, maybe 150.)

I'd be cautious about thinking of the survey as "watered down with respondents outside the problem areas" -- the problem there is making assumptions about what the problem areas were. I mostly would have liked to see a substantially larger sample. I think I recently saw, on another board, someone say confidently that at least 300,000 people were turned away in Ohio. Not only am I not confident of that, but I think it's unlikely -- and this survey is one reason to think so. That said, there are plenty of numbers smaller than 300,000 that could be consistent with this survey and (in combination with other 'irregularities') could bring the Ohio outcome into question. (And if the registration rolls were messed over, I score that a Big Problem whether it puts the outcome in question or not.) I think some grassroots research continues on the registration issue.

It seems very late in the 'day' to field a study on 2004 election day experiences. That said... a professional telephone survey might run $15 or more per completed interview, but it might be feasible for one of the Cleveland schools to do a study in Cuya for less. In the electoral-vote.com database I see that Ohio University (in Hocking County) and U of Cincinnati fielded horse race polls in 2004; I don't know whether either did any voter experience polling.

(By the way, on a nearby branch of this thread, did you mean to reply to Awsi? He's the one who called you on the 2 million, not me -- I assumed that you weren't referring to Ohio.)

"the problem there is making assumptions about what the problem areas were."
I wasn't making an assumption of where the problem areas were, just pointing out the sample
could have been concentrated outside problem areas, I don't know how the sample was selected.
Since I don't recall anyone postulating that ALL of Ohio was fraudulent, I thought it was safe to
conclude that problem areas exist.

What was the source of information the poster used on the other forum for the 300k figure?

Without details, I have No faith in this survey at all.
Include that with the observation that the DNC didn't seem to want to fight for a recount,
and this company seems to be heavily used by them, gives me reason to look elsewhere.

"No two million new voters in Ohio, that's for damn sure. (No, wait... I am making assumptions again!)"
Sure sounds like this was for me.

Considering that the Lucas County investigation found 20,000 (from memory, cause Ohio changed the link)
valid new registrations sitting on a desk, unregistered,
why is it so hard to believe that there could be 2 million new voters in Ohio?
Considering Lucas is 4% of the population of Ohio one could extrapolate 500k new voters just from
the illegally neglected registrations of Lucas county (never mind any accepted ones).

I'm sure you know about the NYT article which stated the new registrations in Ohio, were running 10 to 1 for Kerry.

I didn't mean that you, specifically, were making assumptions either way; it was a general point about evaluating the sample. I do think we could find people postulating that practically the entire Ohio recount was fraudulent, from which it would be a small step to conclude that fraud occurred throughout the state (although not necessarily the sort of fraud that would show up in even the best Voter Experience Survey). I am not busting your chops here; I encounter a very wide range of arguments.

"What was the source of information the poster used on the other forum for the 300k figure?"

I couldn't tell, honestly. It seemed like it was probably derived from the exit polls. Not relevant -- I was just trying to give an example of a high-side estimate that went against this survey. Close to 200,000 in Cuyahoga County would be another such figure.

"Without details, I have No faith in this survey at all.
Include that with the observation that the DNC didn't seem to want to fight for a recount,
and this company seems to be heavily used by them, gives me reason to look elsewhere."

There are at least two ways to read this. One is as legitimate dissatisfaction that will in fact spur the writer to seek additional information. The other is as a convenient excuse to discount discrepant data, by imposing a much higher standard of proof than the writer would apply to data that supports his or her priors. It's the difference between omnidirectional skepticism and cherry-picking.

It's always legitimate to question a survey without a complete statement of methodology, but the indicated course would be to inquire about the methodology. Most likely it was Random Digit Dialing, possibly geographically stratified.

If -- if, mind you -- you are inclined to assume that the DNC somehow rigged this survey to understate the proportion of voters turned away at the polls, and if you wanted to convince a survey researcher that this is true, then you would be well advised to come up with stronger warrant than "the observation that the DNC didn't seem to want to fight for a recount." That said, if you go out and actually get some new data, then I will give you kudos no matter what you do or don't do with the DNC survey.

"'No two million new voters in Ohio, that's for damn sure. (No, wait... I am making assumptions again!)' Sure sounds like this was for me."

Well, the parenthetical does evince exasperation that you played a considerable role in evoking, but the part you seem to be missing is my subject header, "ah, yes, I assumed a topic-switch." In other words, my post #57 assumed that the two million voters hypothesized in the text of #55 weren't directly related to the Ohio scope of its subject header.

And yet, now you ask, "why is it so hard to believe that there could be 2 million new voters in Ohio?" Umm, OK. So does this mean that you are mad at me for assuming that the two million applied to Ohio and for assuming that the two million didn't apply to Ohio? Jeepers, what a week we are having! Really, I really am that confused. The numbers that follow may apply to some part of your scenario, if you have a scenario, but I am not at all sure.

Well, the Ohio exit poll indicated that 16% of Ohio voters were "first-time" voters. Comparing to the national results and allowing some wiggle for likely overreporting, probably somewhere between 20% and 25% of Ohio voters did not vote in 2000. So, if that is two million voters, then we would have a total turnout somewhere between 8 million and 10 million voters, which is hard to reconcile not only with the official returns (5.6 million votes) but with the population of Ohio (voting age population 8.7 million). That estimate would presumably apply to people who at least thought they had voted, not to people who stayed home because they had been purged.

"It's the difference between omnidirectional skepticism and cherry-picking."
The first reason was enough to not suggest cherry picking, just because a second, less objective reason is added doesn't discount the first.
'I think he's dead because he has no pulse...he's also not moving'

"It's always legitimate to question a survey without a complete statement of methodology, but the indicated course would be to inquire about the methodology."
Pot, kettle, black. You said you were displeased with the data...did you inquire about methodology?

"If -- if, mind you -- you are inclined to assume that the DNC somehow rigged..."
Let's not focus on the second reason, when the first was sufficient.

"And yet, now you ask, "why is it so hard to believe that there could be 2 million new voters in Ohio?" Umm, OK. So does this mean that you are mad at me for assuming that the two million applied to Ohio and for assuming that the two million didn't apply to Ohio? Jeepers, what a week we are having! Really, I really am that confused. The numbers that follow may apply to some part of your scenario, if you have a scenario, but I am not at all sure."

As I said, the 2 million number was pulled out of thin air, for an example.
That does not conflict with me asking you how you arrive at your statement which gives the impression that it is utterly impossible for that to be true.

I'm not sure why all my conversations end this way with you, I can only surmise it's by design.

Febble, I did my part and didn't get snarky.
I'm done.

on this board.

</joke>

<change subject>

You know what? I'm starting to get a bit excited at the idea of President Gore. He's been a big hit over here this week.

You might want to review a few of your comments in the last few days, and then ask yourself why you insist on placing the worst possible construction upon mine. Or not. The choice is yours. I gave you an out at every single point. What did you give me?

"You said you were displeased with the data...did you inquire about methodology?"

Nope, because my main dissatisfaction is with the sample size. I haven't really had reason to care whether they stratified the sample. However, when I get around to writing more systematically about Ohio, I will need more info -- the survey instrument, for starters.

"As I said, the 2 million number was pulled out of thin air, for an example. That does not conflict with me asking you how you arrive at your statement which gives the impression that it is utterly impossible for that to be true."

OK, fine, but it really shouldn't be that mysterious why I would be confused -- especially since you didn't ask the question when I first made the statement. Not as confusing as being accused of not telling the truth about tabulator fraud in some unidentified post two months ago, but confusing nonetheless.

Have I answered your questions? I have tried.

Ohio had less than six million votes cast, total.

I don't see how you get to the Kerry deficit, in terms of purged votes. The margin was 118,000. Kerry received 2,741,000 votes. So the 118,000 would be 4.3% of Kerry's official vote tally. I like to keep this stuff fairly simple and sensible. There is no way anyone is going to convince me Kerry was robbed or denied 4.3%, or anything close to that.

Just look at Florida 2000, which had almost the identical number of votes cast as Ohio 2004. Gore received 2,912,000 votes. Add 4.3% and that's a 125,000 vote margin for Gore. Not even the most wildly optimistic estimate gives Gore that margin. You can take all the 57,000 purged "felon" votes and assign them to Gore. Which is ludicrous, BTW. Tenfold ludicrous, or more. Eliminate every multiple page or butterfly ballot like Duval and Palm Beach. Give Gore the deserved net overvote advantage in those counties, 14 total. Then attribute the undervotes. Gore still comes up short of the 4.3%.

I don't think anyone is arguing Ohio 2004 was equivalent or worse than Florida 2000. Ohio had only 93,000 spoiled punch card ballots. An absurd 80/20 split and pretending none of them were intentional non-votes nets only 56,000, less than half Kerry's deficit. I know all about lack of voting machines in African American precincts, provisional irregularities, bogus phone calls, Blackwell nonsense like requiring standard weight paper for voter registrations. It's not 118,000 net.

Just imagine widespread reported irregularities in Pennsylvania and the GOP claiming they won that state. We would be on the floor in a seizure of laughter. But if you look at the numbers they aren't much different. Each candidate with slightly less than 3 million votes and Kerry winning by 145,000. That state wasn't even a major concern on election night since Kerry's margin was always secure. I like to turn things around from a perspective standpoint, to belt myself in the head and closer to reality.

Of course, Diebolders require half the 118,000, via the magic of switcharoo. That's why I intentionally responded to a post asking about purged.

No two million new voters in Ohio, that's for damn sure. (No, wait... I am making assumptions again!)

I never rarely say never, but to review what the DNC's Ohio Voting Experience Survey found: (1) About 2% of respondents said they didn't vote because of long lines; these voters "would have divided evenly" between Kerry and Bush. (2) One respondent out of 1201 (the report says "0.08%") wasn't given a ballot because of a registration challenge. (3) "...a third group of voters (equivalent to 0.83 percent of the voting population) did not go to the polls at all because they did not receive their absentee ballots, or had heard about long lines, registration challenges, and confusing polling sites." We don't know how they would have voted, although it is implausible to assume that they all would have been Kerry voters.

http://a9.g.akamai.net/7/9/8082/v001/www.democrats.org/pdfs/ohvrireport/section03.pdf

We also know that about 35,000 provisional ballots weren't counted. Some fraction of these may have been cast by voters who were wrongly purged. (Likewise, some fraction of the people in categories (1) and (3) may have been wrongly purged.)

So, how many Ohioans were illegally purged and attempted to vote? and how many net votes would they have given to Kerry? I don't know, but it is hard to drive six-digit figures through these data points.

--I got to spend a bit of time on "security moms," then was dragged underwater (metaphorically, of course). The aggregate exit poll data seemed to indicate that while women 'broke' more sharply to Bush in '04 than men did, unmarried women broke more sharply than married women did. While that (and the age tab as I remember it) argues against a literal emphasis on "security moms," it seems quite plausible that "security" was pivotal. I look forward to getting back to those data.

On the other thread I looked a few times then I understood you were probably swamped elsewhere.

Normally I isolate white women, regardless of marital status, in terms of switching to the GOP post-9/11. Perhaps I failed to do that in the specific thread. Most of the polls don't break it down to married or unmarried but it's definite unmarried women, while they voted in higher percentage than 2000, which killed Gore, did not support Kerry by expected margin.

Recently I heard something amazing on a network cable show on CNN or MSNBC, that half of Bush's gain in 2004 compared to 2000 came from women who lived in sight of tall buildings. It was not followed up on at all, nor was a "tall building" defined. I haven't seen the study but the source of the comment was reputable. I think it was Chuck Todd of the Hotline.

Half seems like a wild stretch, but as a Las Vegan I can tell you that was a persistent theme leading up to that election, local women I talked to worried the huge hotels were a logical target, especially after reports Atta and many other 9/11 hijackers met here months prior to 9/11.

I'd have much more energy toward Ohio 2004 if the deficit had been half or less. In the vital states, we need to identify potential problems now, instead of agonizing and speculating over lost numbers years later. Just like think tanks where the GOP got the jump decades ago, they're probably working on new low life vote suppression scams while we mentally Band Aid 2004. We need creeps who can at least partially anticipate the thought process of bigger creeps.

I wasn't referring to Ohio with the 2 million number,
It was an abstract to make a point.

Victoria Lovegren has done extensive work on this.

Over 165,000 possibly illegal purges in Cuyahoga alone.
http://ohiovigilance.org/Counties/Cuyahoga/Analysis/CuyCancelledNonVotingAnalysis.pdf

Did over 165,000 people in Cuya alone show up to vote and get turned away? That would be something like 20% of the intended electorate (some 670K presidential votes counted). Did even half of those people show up and get turned away? It doesn't seem likely. Kerry got about 70K more votes in Cuyahoga than Gore and Nader combined.

I'm not saying that the purge was copacetic -- I have little basis for judgment. But probably it was at least partly copacetic, because I really think it's unlikely that 20% of the intended electorate could be purged without leaving bigger traces than we see. Heck, according to the Census,
http://quickfacts.census.gov/qfd/states/39/39035.html
Cuya's population is about 1351K, of which 25% is under 18. That leaves just over a million voting-age population, some of whom presumably aren't eligible to register. So, the turnout in Cuya looks pretty darn good, substantially higher than the national average.

I think it would be helpful to randomly sample maybe 100 or 200 of the purgees and try to figure out how many of them were frustrated voters. It wouldn't take a large sample to get some idea of the order of magnitude (indeed, if we are trying to choose between 5% and 50%, a very small sample could be quite informative).

or lower turnouts than the rest of the state?

If my memory serves me well, I think it was an obvious difference in
favor of the Republicans.
Considering where the long lines were, that would be a hard pill to swallow.

it appears that voting machines were allocated at least partly on the basis of past turnout. As turnout was lower in 2000 in Democratic precincts than in Republican precincts, the result was that fewer machines per "active voter" were allocated to Democratic precincts.

Of course where turnout is low one year there is greater scope for increase in a subsequent year, so this is a pretty crass way of allocating a resource, and the result, in 2004, when turnout increased, was longer lines in predominantly Democratic precincts than in Republican precincts - which in turn probably suppressed turnout.

So it is difficult to calculate the extent to which lower turnout in Dem precincts in 2004 was due to the long lines and the extent to which it reflected historically lower turnout habits, which, in any case, might be due to past suppression of the vote. But I had a go at trying to calculate it in this paper:

http://uscountvotes.org/ucvAnalysis/OH/FranklinCountyReport_v2.pdf

And Walter Mebane has a rather more authoritive paper here:

http://macht.arts.cornell.edu/wrm1/franklin2.pdf

and an article here:

http://www.tompaine.com/articles/2005/07/21/timing_and_turnout_in_ohio.php

If you base it on registration, then urban areas will typically fare worse because they tend to have more "deadwood." Cuyahoga has, according to Leip, 1.006 million registered voters, which is almost as high as its voting age population (apparently about 1.013 million, based on Census figures). Leip reports that Cuyahoga had 67.0% turnout as a percentage of registered voters, below the state average 70.6%.

According to Mike McDonald, Ohio's VAP (voting age population) turnout was 64.72%. (I don't have the link handy, but you can find it on his George Mason website if you want.) Cuyahoga's VAP was apparently about 1.013 million, and its turnout was over 670,000, so its VAP turnout apparently was above the state average.

Bear in mind that illegal purges tend to drive registered-voter turnout up.

I see little doubt that Kerry lost votes due to long lines, but I cannot recommend using turnout differentials as an indicator.

the BBV fakes are obvious to me.

And no, this doesn't mean I don't think election integrity is not important. I think election integrity is crucial.

It's just that there are people that are ACTUAL election reformers, like John Conyers, and there are FAKES.