Hand count versus machine count in NH.

[drm604]

Anyone here seen this yet.

http://www.bradblog.com/?p=5544#more-5544

Analysts at the Election Defense Alliance (EDA) have confirmed that based on the official results on the New Hampshire Secretary of state web site, there is a remarkable relationship between Obama and Clinton votes, when you look at votes tabulated by op-scan v. votes tabulated by hand:

Clinton Optical scan 91,717 52.95%
Obama Optical scan 81,495 47.05%

Clinton Hand-counted 20,889 47.05%
Obama Hand-counted 23,509 52.95%

The percentages appear to be swapped. That seems highly unusual, to say the least.

If those numbers are correct then it certainly does seem odd. The number of hand counted votes is about 20% of the whole which could be considered a more than sufficient statistical sample of the whole. Of course it isn't totally random, but it wouldn't be totally unexpected to see the percentages come close to matching those of the machine counted votes.

With that in mind, one could be forgiven for hypothesizing that the machine counts got swapped somehow. (You could, of course, also hypothesize that the hand counts were the ones that got swapped, but it seems to me that it would be more difficult for that to happen.) I'm not saying that that's the case. I'm skeptically minded by nature and there may very well be a logical explanation for this. It seems like it would be stupid for someone to do something so obvious. But it certainly is a striking coincidence.

Anyone know if the optically counted votes are all added up in one central tabulator?

Do we have any statistician's here who could address this?

[Mr_Jefferson_24]

1. Not unusual anymore -- not since 2000. nt.

[defendandprotect]

5. The steals don't begin in 2000; they begin in the mid-1960's . . . !!!

[Mr_Jefferson_24]

6. There may very well be some truth to that...

...but the reconfiguration of our electoral process in order to facilitate the easy and virtually undetectable theft of elections via machines seems to have been born around 2000.

[drm604]

8. Possibly earlier.

http://journals.democraticunderground.com/drm604/8

[Mr_Jefferson_24]

18. Good information -- thanks for posting it. nt.

[defendandprotect]

47. Thank you ... I wasn't aware of those articles . . . and I think it

is essential that Americans understand that this didn't begin in 2000 --
and that DU'ers understand this didn't begin in 2000 ---

It certainly changes one's overall concept of what has been going on when you undestand that!

[defendandprotect]

45. No . . . and it is frustrating because I've tried so hard to get this info out here ---

No . . . there is every evidence that this begins in the 1960's . . .
two journalists --- Jim & Ken Collier investigated the machines at that time and wrote a book called VOTESCAM which has been suppressed ---

Evidently, there are people here who think this has something to do with their nemesis
Bev Harris and so they seem to be frightened of hearing the story!!

Go figure --- !!! ???

But, again, the computers came in during the mid-1960's and the steals have been going on since then!!!

And when America --- or even DU members --- don't understand that, we are way behind in understanding this whole thing!!!

[Mr_Jefferson_24]

49. This book is online here...

http://www.constitution.org/vote/votescam__.htm

---snip---

NEW HAMPSHIRE CONFOUNDED MOST

POLLSTERS



Voters Were a Step Ahead of Tracking Measurements



By Lloyd Grove

Washington Post Staff Writer



For Vice President Bush and his supporters, Tuesday's 9-percentage-point

victory over Sen. Robert J. Dole (R-Kan.) in New Hampshire was a delightful

surprise; for Andrew Kohut, it was a horror story.



Kohut is president of the Gallup poll, whose final New Hampshire survey was

wrong by 17 points: it had put Dole ahead by 8; Bush won by 9. "I was

dismayed," Kohut acknowledged yesterday.



This New Hampshire primary was perhaps the most polled primary election in

American history, and in the end, the Republican voters in the state

confounded the predictions of nearly every published survey of voter opinion.



Gallup's glaring error and the miscalls of other polling organizations once again

raise questions about the accuracy of polls, their use by the media and the

impact they have on voters' choices and the public perception of elections. In

New Hampshire this year, news organizations' use of "tracking polls" to try to

follow the movement of public opinion night after night came to dominate news

accounts of the campaigning and the thinking of the campaigns themselves.



Tracking polls usually survey a relatively small number of voters every night:

150 to 400 in each party, in the case of The Post-ABC poll. The results are

averaged over several days. See POLLS, A11, Col. 1







Had the terms of Bush's "promise" to Sununu been met?



Whatever magic Sununu was able to conjure up during those final

hours preceding the overnight resurrection of the Bush campaign, it

worked.



There are those who believe that such a wild reversal of form would

have been subject to an immediate inquiry by the stewards if it had

happened in the Kentucky Derby. Any horseplayer would have

nodded sagely, put a finger up to his eye, pulled down the lower lid,

and signaled: "Fix."



Yet in New Hampshire, there was some wonderment expressed in the

press, and little more. There was no rechecking of the computerized

voting machines, no inquiry into the path of the vote from the voting

machines to the central tallying place, no public scrutiny of the

mechanisms of the mighty peculiar vote that saved George Bush's

career and leapfrogged the relatively obscure Sununu into the White

House. . . .

[defendandprotect]

50. Yes . . . have you browsed it?

It's been me here trying to get people to recognize that 2000 was simply a NOISY steal ---
but the steals have been going on since 1960's !!!

And, thanks for reposting that original NH surprise ---
good story to keep in mind ---

I forget how I came upon the information that these steals go back that far ---
I probably picked up some reference to their book in some other book ---
but I did get the book and then found the website ---

IMO, it is very important for DU'ers and all Americans to understand this just didn't
begin in 2000!!!

Meanwhile . . . there are some here who think this is connected to Bev Harris in some way!!!
????

It's pretty impossible to talk to them and get them to understand it's not ---
A lot of people here were evidently very hurt by the scam she pulled and it's
difficult to reason with them!!!

Thanks for you help --- !!!

[Mr_Jefferson_24]

51. I glanced through it quickly and bookmarked it...

...it looks interesting and I will get to it. Thanks for the info about it.

Sometimes I feel like the more I learn about our country's political workings, the less I want to know.

[Kittycat]

2. Don't let the Clintonistas see this. Don't you know we don't question this anymore?

:sarcasm:

[K Gardner]

3. Hey.. you forgot where we are ! We're not in GD or GD-P.. we can

talk about it !

This is actually quite fascinating. Thank you for posting :-)

[Kittycat]

36. I was being sarcastic. It seems any time this question is raised, it's attacked.

I agree that this is legitimate, regardless of who is ahead or not. I think it's important for democracy that we continue to work on finding the truth about what has not only happened in the past, what's currently taking place, but also to prevent problems and ensure public support for the legitimacy of the future.

[K Gardner]

40. I am SO sorry !! I didn't open to see the sarcasm tag ! Apologies, truly !

[defendandprotect]

4. Nader didn't actually get a recount . . . see this . . .

Thank you for this info . . . !!!! So Nader didn't actually get a recount --- !!!

"We have no control over the ballot chain of custody and we have learned the pain from the 2004 Nader recount, in which only 11 districts were counted, chosen by a highly questionable person, and then nothing showed up," she wrote recently. "Now all we hear is how the Nader recount validated the machines. A candidate asking for a recount may well be a tool used to 'prove" everything was okay and then that candidate will be further discredited," she warned.

And we just had a scandal recently where two women were "organizing" the votes that they
were selecting to be counted --- in order to forestall a full recount!!!

Where was that---?





http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=102&topic_id=3133525&mesg_id=3133860

[Usrename]

16. Do you mean these two convicts?

http://www.freepress.org/departments/display/19/2007/2379

[defendandprotect]

48. No -- there was a strory like last week posted at DU ---

Two women election workers --- and I forget their location ---
were trying to stop a recount --- so they were collecting votes for the "test"
which would make it seem that everything was on the up and up ---

I'm sure others here saw the story ---

it came and went quickly ---

[lvx35]

7. They are not added up in a central tabulator

according to an article I just read, they are not connected at all. Also the machines are tied to certain types of districts. Really the recount should give this all peace...I haven't heard that the exit polls were off, so I'm not expecting much but people have reason to be paranoid after what we've seen in the past.

[drm604]

9. If that's the case

then it's hard to see this as anything but a coincidence. Rigging every counting machine (or at least the vast majority of them) would be a lot more difficult then just swapping totals in the central tabulator.

But it certainly is a head-scratching coincidence.

In reference to the exit polls, to be fair there are claims that the raw exit polls are off. I don't know how credible those claims are but they are being made.
http://www.bradblog.com/?p=5535

[lvx35]

10. check the math in the premise as well.

Divide up places based on some other criteria rather than voting machine use, and see if they both have the same or very close percentages. I think you will find that the results are much less homogenous than being the same down to a hundredth of a percent, which is the premise underlying the idea that this is fishy. If on the other hand they are very close, you have a strong case.

[drm604]

12. I'm really not sure that I have a case at all.

I'm not claiming anything here, just trying to understand some odd numbers. I have an analytical mind and it demands explanations for these kinds of things. There may very well be a statistical explanation for this. In fact there probably is. Swapping the totals so obviously seems like a stupid thing to do.

For the record, I don't have any bone to pick here. I support Edwards and he was obviously a distant third however you look at it.

[OnTheOtherHand]

34. I will try to post a scatterplot

Clinton/Obama margin is substantially correlated with Kerry/Dean margin in both hand-count and op-scan jurisdictions.

Maybe for fun I can flip the Clinton/Obama numbers in every (ETA: opscan) jurisdiction just to show how different that would look.

EDA's quality control has tanked.

(ETA: Gotta decide where to post this, but sure enough, it looks pretty funny if one flips the op-scan numbers. That did not happen.)

The original results (based on what I got from MSNBC), with places with under 100 votes omitted :



If we flip the Clinton and Obama results in op-scan jurisdictions:

In both cases, the black line is an Ordinary Least Squares best-fit line. If you look closely at the second, you'll see that a line for just the hand-count places would slope up, and a line for just the op-scan places would slope down. As it is, they roughly cancel out. (This is OLS, so every place counts equally although the op-scan places tend to have more votes.)

[drm604]

42. I'm not sure that I follow what you're saying.

I'm also not sure why you chose to compare to Kerry vs Dean. What was your reasoning for that?

[OnTheOtherHand]

43. historical context

You've probably noticed some arguments that the more densely populated southern tier tends to favor establishment Democrats, while other parts of the state favor mavericks.

I can't make (or even test) the geographical aspect of that point, because I haven't gotten my GIS program to play nicely with my data (and I have to work on other things). But I can tell you that just as the hand-count areas broke Obama over Clinton, they broke Dean over Kerry, and they broke Bradley over Gore.

[drm604]

44. Maybe that's works, maybe not.

I have no way to judge.

I've never heard those arguments, but even if they're correct, weren't there other "establishment Democrats" in the race? Some of those were probably, at the time, considered to be more establishment than Kerry considering his past (the Vietnam protests).

[OnTheOtherHand]

46. reckon it as you will

It's an observable empirical fact that the Kerry/Dean margin is rather strongly correlated with the Clinton/Obama margin. We can ponder the reason(s).

[HughMoran]

11. Exit polls - Obama +1, Actual Results Clinton +3 - they were off by 4%

I'm not sure if this is significant based on expected margin of error. :shrug:

[drm604]

14. Where do you get your figures?

4% doesn't strike me as very significant but I'm not a statistician.

[HughMoran]

17. Here...

I posted this as a separate post, but at the live blog here the fellow reveals the exit poll numbers @ 8PM http://www.washingtonpost.com/wp-dyn/content/discussion/2008/01/07/DI2008010702131.html?nav=hcmodule

Note that all of the references to "Brad Blog" or "Chris Matthews" state that he was up in the polls an average of 8% before the voting, but that the exit polls still show him winning. 1% is still winning - I think the 8% number is getting conflated with "winning" in people's heads as "exit polls show Obama winning by 8%". This appears to be a bad interpretation of the quote by Chris Matthews.

[Usrename]

19. That's not the raw data.

E/M, who I guess did the exit polling again, doesn't release any raw data.

Tweety said "raw data."

[HughMoran]

20. I think the +1% for Obama would still hold

since the data he was reporting was @ 8PM - before it could have been adjusted for bad sampling of the different groups who actually voted.

[Usrename]

24. Why not?

I think returns were coming in since 7:00.

Of course that data would have been weighted, or else they would not have released it.

They don't release their raw data to the public.

From their site:

Subscribers will receive:

Statewide race crosstab reports. (First report in the late afternoon and updates throughout election night.)
Detailed demographics include age, gender, income, education, employment, and other key factors.
Voter responses to current issue and candidate quality questions.
The ability to view all data in horizontal and vertical cross tab format
Printer friendly PDF view of results
The ability to chat with your subscriber contact via telephone and a Live Chat Messenger System on Election Day.
No projections of election outcomes will be made until all polls close in a state.

After the polls close, if the results indicate a clear outcome, Edison Mitofsky will project the winner. If a race can not be called at poll closing, we will remain in communication throughout election night and inform you when a projection will be made.

http://www.exit-poll.net/subscriptions.html

From what Tweety said, he somehow had info about the raw data, not the stuff you are talking about.

[HughMoran]

38. Because he seemed to state that the numbers would be adjusted after hard returns came in

Robert G. Kaiser: One of the problems with exit polls is that the proportions of voters they show from various demographic categories are never accurate. Through the evening the exit poll numbers will be adjusted to reflect the evidence provided by the hard returns.

It's possible that the data was weighted based on some of the results that came in from the early closing polling stations, but you don't know that any more than I know it wasn't weighted. It sure seems to me that he thinks this is unweighted data (see above).

[jbnow]

31. No, Matthews mentioned it several times

He did say just what you said he did a couple of times. But later wanting to prove his point about voters evidently lying to pollsters he wanted to clarify it was exit polls
Hardball transcript

I want to start with you and try—I‘m wide open on this. I have no pre-conception. Like everyone else, I was stunned at 5:30ish last night. I was passed a piece of paper just for guidance that told me that Barack Obama was going to win a significant victory. This was based upon the polling of people going—sorry, coming out of the booth, having voted.

On the 11pm show he said

The exit polling showed, I got to tell you, I got it late in the evening yesterday that showed Hillary behind. And they were telling us it was going to get to be a bigger behind. She was going to fall further behind, so explain that...

So until the final exit polls that are weighted they were definitely saying something else/

[HughMoran]

37. Well, OK - so it's unclear

I thought it was a very telling piece of information posted on the blog without comment at 8PM.

Since I am not able to have access to what Matthews was looking at or know whether the 5:30 data he was looking at was the same as the 8:00 data, there is no way I can discuss this further. I thought it was significant, your mileage may vary :shrug:

[jbnow]

30. That's good but the software could be

the problem, the way it is programmed. That's where problems with diebold tend to be shown...they can make mistakes or be rigged to give one person's vote to another candidate.
Exit polls
will never be off when we get them. (They learned after 2004 not to make raw data piblic)

"Corrected" Exit Poll Tabulations - Once the actual results have been counted in the wee hours of election night, NEP re-weights the results of each exit poll so that the vote preference on the poll matches the actual count. They then release new cross-tabular tables for each state to the general public. In theory, weighting to match the vote preference to actual results makes the complete exit poll more accurate.

That only sounds paranoid but it is just how it's done.

All the networks and AP use the same polling firm and get results from them.

As secret as raw data was supposed to be Matthews on Hardball wanted to prove his point about voters lying to pollsters and revealed that he'd been given a note that evening from the raw data that Obama had a big victory. He repeated it several times. (He wanted to know why they lied to exit poll workers).

[politicallore]

13. Go Ron Paul go!!

Ron really does deserve 3rd... I live in NH... NO WAY he only got 8%

[chknltl]

33. Welcome to the DU Politicallore

um.... we have a rule around here about who we are allowed to show support for. Yeah I know it's pretty fascist but then as you may have heard elsewhere, our beloved Admin are secret supporters of a dead Italian dictator and they feel if they aren't allowed to openly show support for him we can't show support for the gal who quit the party to run against the Speaker of The House. As you are well aware, shit rolls downhill so you won't be allowed to show support for any republican candidates, even if they share the same initials as Ross Perot. Rules are rules

All that said, you are very welcome to be here and share in our :popcorn: I am fairly confident most of the crack was cleaned out of it before Satin blessed it.


btw: I have secretly hoped that your candidate would look into possible vote tamperings on that other side of the fence. I further hope he continues to do so at each phase of this process. Who knows, if he does that, he might wind up with the gratitude of a nation should he discover any shenaningans.

[RufusTFirefly]

15. I think there's abundant evidence that 2000 and 2004 were stolen. But...

... I think there's a remarkably simple explanation for the divergence in hand-counted and optical scan ballots in this latest New Hampshire primary. I'm not a resident of the Granite State, but I know that DUers who live there have made similar assertions:

The demographics of the areas where hand-counted ballots are used are markedly different from those where optical scan is used. The voters for the two voting formats don't come from the same cross-section. It is the difference in demographics, not the difference in formats that explains the alleged discrepancy. Voters in hand-count regions tend to be rural, affluent, and anti-status quo, while voters in the optical scan regions are more urban, working class, and establishment.

That said, I applaud Dennis Kucinich for demanding a recount. The difference between the pre-election polls and the actual voting is dramatic, and there may in fact be some problem that a recount will uncover. I just don't happen to believe that the numbers listed above say anything more than that people who vote in one region of the state have different preferences than people who vote in another region. If it were Upstate New York vs. New York City, no one would bat an eye.

That said, I think that election officials should always expect the demand for a recount. It will keep them and the process honest.



NO ONE expects an election recount!!

[drm604]

21. That's true.

The demographics are different and that would explain the discrepancy. I thought of that and I wouldn't have found a simple discrepancy to be very puzzling.

You're not addressing my post. Read it again. I wasn't talking about the discrepancy. I was talking about the exact swapping of percentages. I don't see how that can be explained by demographics.

[HughMoran]

22. That's how the math works out

I don't see the significance of a coincidence.

I did the math for the 47.05%

0.47049280650301364801514906588458

versus

0.47049416640389206721023469525654

The percentages are very close, though not identical.

[drm604]

25. I'm not sure that I get your point.

I did the math also. I don't see how that explains away the coincidence.

One thing I do notice is that the two percentages add up to 100%, so these numbers don't account for the other candidates. I'm not sure if that's relevant or not. Maybe I need to dig into the numbers myself.

[HughMoran]

39. Point - you made their numbers artificially = 100%

by isolating Clinton and Obama from the rest of the candidates (like you said). Just because the percentages work out the way they do when you divide one of the numbers by the total has no statistical significance whatsoever. I did joke that the accidentally flipped the results when they were adding them up, but that was just a joke :D

[drm604]

41. I didn't make their numbers do anything.

Just for the record, those are not my numbers. They're from the link in the OP. Those numbers don't prove anything, I agree. But they certainly are odd and they are what might result if (big if) the totals were swapped.

[RufusTFirefly]

23. That may be nothing more than a statistical anomaly

After all, it's not like the results in Texas a few years back (or perhaps it was the Max Cleland race in Georgia) where several of the percentages had the same repeating decimal.

Also, why are we just shown Clinton and Obama (and Paul, in the original posting)? Is there a similar discrepancy among other candidates?

I think a far more effective way of rigging the election might've been to programmatically siphon off a small percentage of votes from a certain underpublicized anti-corporate candidate and allocate them instead to a candidate who badly needed a victory.

[drm604]

26. I agree that this seems like a dumb way to rig an election.

It's just too stupidly blatant. But it still seems like an unlikely anomaly to me. I'm wondering if the omission of the other candidates has something to do with it.

[K Gardner]

27. To my understanding, the only discrepancies were in the

Clinton and Obama numbers. There is a graph somewhere, but I'll have to find it tomorrow. I'm cross-eyed tired :-)

[Usrename]

29. It's still a pretty weird demographic.

It seems to affect Clinton and Obama by way of an eight percent swing, with everyone else remaining a whole lot more constant.

http://checkthevotes.com/index.php?party=DEMOCRATS

Although there is a lot of variance between hand and machine on the Republican side, too.

http://checkthevotes.com/index.php?party=REPUBLICANS

[drm604]

35. Those totals differ from the ones I posted.

Maybe we need to ignore all of these second hand reports and go back to the original source.

[jbnow]

32. Not just demographics. Here is a break down

by town size and machine or hand count

http://checkthevotes.com/index.php?party=DEMOCRATS
For the Dems you can see overall Obama won by 4% in hand count but lost by 4% in machine counts. Maybe rural voters who usually have hand counts just liked Obama better. That's unusual as common wisdom would have the young, black upstart candidate do better in urban areas.

But it it seems more rural machines just don't like Obama, big swings.
Clinton beat him in small town machine votes by 2% and lost in hand count votes by 6%.
In medium towns she beat him by 6% on machines and lost by 4% in hand counts.

There was a small number of hand counted votes in large towns and Obama did get a lower percentage. It is too small a sample (2%), probably 1 town, but it might turn out that a lot more urban voters did like Clinton a lot more.

[cutlassmama]

28. Way to go Dennis!!! I hope the paper ballots don't "disappear" eom