Exit Polls: Can someone debunk or explain this?

[Bill Bored]

I have read that states with paper trails matched the exit polls and those without them did not. But the only actual example I've seen is OR.

TIA did a graph once that showed little or no correlation between machine type and Exit Poll Red Shifts by state. Meanwhile, I've tried to find a list of states with and without paper trails and have not been successful, much less voter-verified paper ballots and actual random auditing procedures.

So what's up with this? I appreciate the Oregon example, but that's just one state. And OH had LOTS of paper, but ballot orders and other scams could have been used to mess with it.

NY has nearly all mechanical lever machines. We recheck every one of them within 2 weeks of the election. This is like doing a 100% audit or in effect, a full manual recount. Yet we had a large Red Shift in our exit poll too. I'm still trying to find out about how the NY totals are tabulated, but to be sure, it's not done on a networked GEMS server, or if it is, the totals MUST be manually entered. Might as well just use Excel.

When you look at machine problems on election day, NY had it's share, but based on our large population, this was SMALL compared to other states with no people and lots of DREs!

So how can we explain all this?

[merwin]

1. It's not the voting machines that were the problem, it was the tabulators.

[Bill Bored]

2. See comments about NY nt

[Botany]

3. True .....

In Illinois, Maine, and (?) Wisconsin that had paper trails the
the exit polls matched the vote. In all states that swung to bush
that did not have paper trails the difference between the exit polls
and the raw vote was bigger than the margin of error.

Try Winkipedia .......?

Blog help for the histograms.

:bounce:

[TruthIsAll]

4. Here's the graph

[Bill Bored]

19. Thanks TIA

From this, it looks like machine type didn't matter much.

1. FL and TX are mirror images of each other with the same machine types.
2. NY, again had no e-voting.
3. Lots of states had the vote come pretty close to the exit polls, even with e-voting, although this could be due to small populations. Do you have a percentage weighting? I remember seeing one and it looked similar. The title of that thread was "It's Graphic..."

Also, I assume these are the unadjusted exit polls, right?

Were there any deviations at the state level compared to the Adjusted (11/3) version?

Also the BBV/DRE category does not distinguish whether or not there's a paper trail, i.e., VVPB, or not.

It's fair enough to say it was the tabulators that were screwed up. With Diebold, GEMS is used for both DRE and OpScan tabulation. But this still doesn't backup the assertion that states with paper matched the exit polls any better than those without paper, which was my original question, right?

Maybe this needs to be checked at the county level, but we don't have exit poll data for that do we?

[emcguffie]

5. look at maine. n/t

[southwood]

6. Wish I could help!

But check out the Edison/Mitofsky evaluation report

http://www.exit-poll.net/election-night/EvaluationJan192005.pdf

page 40. The only way of voting that comes sort of close to the exit polls is good old paper ballots. (They suggest that this is because paper ballots are only used in rural areas.)

[TruthIsAll]

7. ROTFLMAO: They really said that? Have they no shame whatsoever?

The only way of voting that comes sort of close to the exit polls is good old paper ballots. (They suggest that this is because paper ballots are only used in rural areas.)

[andym]

8. Do you have more info on what you mean that evey voting machine is checked

What do you mean that every voting machine in NY was checked? Are the numbers actually re-recorded and compared to the original vote? If so, then this really the first state that underwent a full recount!!!

Is there any info or documentation available?

[Igel]

12. They canvas the machines.

They have to make sure that the people (like me) that fill out the forms don't screw up. We amateur poll workers (just 4 of us) read off the numbers, and then close the machine; then sometime later (at least in Monroe County) they double check our work.

There's a reason the election night returns aren't official.

(edited to remove references to seals; the seals just keep more votes from being cast; the backs of the machines are simply locked.)

[Bill Bored]

21. So Igil, why do you think the NY Exit Poll didn't match the vote?

And was this the official vote that didn't match, or the initial canvas, which we know is always somewhat flawed?

And how are the votes finally tabulated to get the statewide totals?

Thanks!

[Igel]

24. I have absolutely no idea.

It's possible to rig the machines, but that sort of tampering would involve dem involvement and be easily discovered by examination. After all, *they're not making spare parts*.

Once you shave the gears on one of those machines, you have to scavenge parts from broken machines, and that's not a trivial task these days.

And since the ballot order is random, the next time the machines were used it could affect the Working Families Party as much as the Dem party.

I'd like to see a county-by-county breakdown of the non-Hubble "red shift". Just too lazy to look.

[Bill Bored]

16. This is the law in NY. nt

[FogerRox]

9. Paper trail--no such thing in the USA-IIRC

[skids]

10. Ah, but what was FOUND during the machine "recheck"?

Don't assume you'd know if there was a problem.

I heard the seals were broken on the back of the NY machines when they went to look. There were news articles in a local NY paper about some strange stuff going on, and there were lots of machines in Bronx and Queens "stuck on the Republican side" according to complaints filed at EIRS.

[Igel]

13. I don't understand your post.

Unless the machines were different from the ones used in Monroe County, the only thing the seals do is keep more votes from being cast. (And they're a pain to get on sometimes: we had one that wouldn't seal properly, and had to just leave it that way.)

What does "stuck on the republican side" mean? That the little wheels were stuck?

[skids]

14. Look, it's your state.

So I really don't feel the need to do your homework for you.

But I'll throw you a couple quickies:

http://vote2004.eriposte.com/bluestates/newyork.htm

https://voteprotect.org/index.php?display=EIRMapCounty&tab=ALL&state=New+York&cat=02&start_time=&start_date=&end_time=&end_date=&search=&county=Bronx

[Bill Bored]

20. I know about voteprotect and there were dupes in those reports as I said.

Also, the 5 counties of NYC each have Millions of voters. A few bad machines for that many voters is much less than the per capita failures in the rest of the country.

I thank you for the other link though.

[skids]

22. Keep in mind EIRS is just the tip of an iceberg.

...of all the reports that didn't get called in to that database. Many of the reports got filed with local BOEs. No offense intended, but why we would trust the BOEs not to circular-file their own trouble reports I don't know. That's not a matter of politics, that's just a matter of good business sense -- you use a third party to monitor the effectiveness of any worker, not self-reporting.

EIRS is merely an indicator as to the nature of problems.

There were a few other online databases as well. It was a confused mess.

Anyway, if it is any consolation, NY has one of the lowest ratios of interviewees to votes cast (0.02%). Say what you will about the absolute survey size determining the MOE outright, when you're dealing with a survey that is cost-cutting rather extremely by using demographic shortcuts, the impact on the true MOE (which noone will ever know, there are too many variables both known and unknown, even to Mitofski) is likely to be there.

The true answer for a lot of states probably includes multiple variables. How many people walked out of the polls thinking they voted when their vote wouldn't actually count due to provisional rules? How many people had their vote switched? These are different phenomena, and will have different correlations. Given the variety of suppression/fraud techniques and the variety of ways vote counts can be just plain screwed up through mismanagement, trends are likely to be hard to find.

[Bill Bored]

17. No not lots of machines.

Some of those reports were dupes, stating the same location as Public School numbers and street addresses, making it look worse than it was. There are usually backup machines available and if not, paper ballots were issued. It's true the machines are old though and they do break down.

I havne't seen the documentation on the broken seals and would be interested in that.

[L. Coyote]

11. Percentages of USA Population Served by Voting Machine Types

Here are links to spreadsheets with data:

Percentages of USA Population Served by Voting Machine Types

Exit Poll and Actual Voting Discrepancy Spreadsheet

2004 Presidential Election Results

DU was a source for some of the data on this site:

Pobeka's Complete tabulation:voting methods/vendors by population in state

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=201&topic_id=5446

[Wilms]

15. During the RNC, Peter King made a remark about "counting the votes"

Ring a bell??

[Bill Bored]

18. Saw the video clip. It's so out of context though.

which election was he referring to, etc? His statement was made in 2003 I thought, on the White House grounds, right? Why doesn't someone just ask him? Is Alex Pelosi's video out yet?

[seaclyr]

23. Bill, I wouldn't make too much of some of the state-by-state comparisons

for example FL and TX. They needed votes in FL, they didn't need them in TX, in fact they had so many in TX that an extra few percent would have really looked odd. In the same way, they didn't need votes from some of the small states. Soooo... two states may have similar voting systems and different needs as far as "adjusting" the numbers is concerned.

[berniew1]

25. Its possible to rig any system if no safeguards, and it was done this time

But the discrepiency in exit polls and official results has been documented in most cases, due to a combination of widespread touchscreen fraud and dirty tricks/malfeasance to manipulate registrations, absentees, and provisional ballots- especially in minority areas throughout the country.
http://www.flcv.com/ussumall.html

[berniew1]

26. Fraud & manipulation shown to explain the exit poll/vote gap

http://www.flcv.com/ussumall.html

[mgr]

27. Depends upon how much you have emotionally invested in the red shift.

I have been of the opinion that the red shift is an artifact of exit polling bias and design, and the evidence now supports that position. That does not mean that voting irregularities sufficient to alter the presidential election results did not occur.

The original argument was that the West was not necessary for the republicans to tamper with the vote, and was all paper ballot to boot, so the exit polls were a fit. However, the only state in the election with a full hand recount was Washington (save two counties). A comment (just before Land Shark and team reported their findings) was made that the governor's race results were beneath that projected by the exit polls, and would imply a similar result may apply to the presidential race.

The one thing that no one has verified is the success of republican registration efforts. What could be hidden in the process is the registration of independents as republicans, this would effectively mask past voting patterns, as current party identification would change, but not whether they participated in past elections, and would in part address the zombie problem.

One thing democrats have to contend with is that our party has been a coalition party, and independents are more likely in the past to have splintered from the democratic party rather than republican party (the only exception I can think of is the Reform Party). Therefore it is unlikely that democratic voter registration efforts would pick up these voters, and have to look for new voters.

Now, with New York I would attempt to obtain the new registration patterns, and compare the number of registered republicans in 2004 to 2000 by county; I would do the same with democrats; then I would look at the pattern to see if the outcome is simply a function of increase in party identification, and map the residuals. Where the residuals lie are where the irregularity that you are interested in will be, and likely the result of local republican administration (I will bet the pattern will be strongest in the north).

Mike

[Bill Bored]

28. Ah, here it is from the master himself (Freeman)!

"...Mitofsky and Edison summarily dismiss the possibility that the official count was wrong. They reject the election fraud hypothesis because, they say, 'precincts with touch screen and optical voting have essentially the same error rates as those using punch-card systems.'"

"Indeed, they do. But this fact merely suggests that all three of these systems may have been corrupted."

<http://www.inthesetimes.com/site/main/article/1970>

I hope we can put this one to bed, at least for now.
A paper trail alone does not a fair election make!

[Carolab]

29. That's because it's not the voting, it's the COUNTING. n/t

[Wilms]

30. kick n/t