Topic: Presidential Campaign 2008
Diebold Vote Error in New Hampshire < 00.948% A recount prompted by fears of massive vote errors or counting machine flaws had no effect whatever on the outcome.by William Westmiller
(Libertarian)
Friday, January 18, 2008
There are certainly reasons to be cautious about electronic vote counting, but the fears about massive errors in January's New Hampshire Presidential Primary appear to be unfounded. There were errors in the tabulation, some of them corrected the next morning, but they had little or nothing to do with the Diebold Optical Scanners that were used as part of the counting process.
With a few scattered reports of errors, Democratic presidential candidate Dennis Kucinich paid $27,000 for a recount of Democratic ballots in the largest population county in the state, Manchester, which uses the Diebold scanners. With only a few small wards left to be counted , Kucinich's share of the vote dropped by less than two hundredths of one percent (-00.0166%), a total of three votes. That error apparently occurred in one ward, where a single precinct was accidentally counted twice or confused with vice-presidential votes. Those 106 corrections reduced John Edward's tally for Ward 5 by about one percent, but offsetting errors in other townships resulted in a total loss of only 12 county votes for Edwards (-00.0284%).
Ignoring that apparent human error in one ward, and assuming that all the remaining miscounts were a result of Diebold faults, the total number of errors in the county was less than one percent overall, including all the over and under-counts. However, there's no reason to believe that those were machine errors. Voter and poll handling errors could easily have accounted for the 400 mistakes among 42,210 votes tabulated. The New Hampshire Secretary of State will be reviewing the errors and may recommend changes in the counting procedures.
This was only a recount of the Democratic ballots in one county, not the entire state. A requested recount of all the Republican ballots in the state was submitted by GOP candidate Albert Howard, but the check for $56,000 could not be redeemed before the deadline. His unofficial vote count was 44 statewide. [See Update3 below]
Bear in mind that the reported results immediately after the election are not official. The Secretary of State has thirty days to check for errors before reporting the final, official tallies. That process may have found the reported errors in any case.
The New Hampshire election may not have been the ideal venue for those who fear vote manipulation, since all the ballots are cast on paper and preserved. The Diebold machines only scan the paper form for proper marks, which can be verified against a hand count. Most of the small, rural counties in New Hampshire manually count the ballots. There was a statistical variation of one or two percent among the candidate results in those counties, but there was no reason to believe that those were not simply the variable preferences of the voters in those areas. Representative Ron Paul decided against a recount after considering all of those factors.
Those who fear machine fraud or vote manipulation are going to have to search elsewhere for some hard evidence that voters need to be concerned about whether their ballots are not being counted. In New Hampshire, there are 106 voters who might be pleased to learn that their votes were counted twice.
...
Note: Excel spreadsheet links provide all the raw and calculated data. A PDF summary of the primary vote calculations can be found here.
Update1: The recount is not yet complete, with Weare township still outstanding, but the only significant errors were in hand-counts, not machine counts. Nasua Ward 5 errors are being misrepresented, by failing to note that poll workers incorrectly double-counted by hand the correct totals tabulated by machine. See comments on poll worker errors in the comments below.
Update2: The recount of Hillsborough County was completed on January 22nd and the recount of Rockingham Count was abandoned due to insufficient funds. The final error result was 00.5469% with a total vote variation among all the major candidates of -00.0724% and a loss of 5 votes for the sponsor, Dennis Kucinich.
Update3: The NH Secretary of State extended the deadline and agreed to a recount of all Republican votes, at a price of about $59,000, mainly contributed by Ron Paul supporters. With major counties completed, Paul gained a total of 16 votes in Rockingham County and lost 13 votes in Hillsborough County. All of the errors were a result of human transcription oversights.
Did you like this article? If you did, Thumb It! 11 thumbs so far
2008 William Westmiller, all rights reserved.
Published: Friday, January 18, 2008
Last modified: Tuesday, February 5, 2008
The views expressed in this
article are those of William Westmiller only and do not represent
the views of Nolan Chart, LLC or its affiliates. William Westmiller is
solely responsible for the contents of this article and is not an
employee or otherwise affiliated with Nolan Chart, LLC in his/her role as a columnist.
The numerous ballot boxes that were found to have been tampered with, along with the missing memory cards. are primia facie evidence of voter fraud. The rebuttal to this entire column is manifest in the below thread:
The most obvious motive for slitting open the boxes and compromising the recount, is that vote fraud occurred, and the perpetrators had to cover their tracks. Or as summarized by Mike Rivero, "the New Hampshire primary (is) discredited and unrecoverable. That the ballot boxes have been opened proves the recount is rigged. That the New Hampshire officials could make such a stupid comment as to say that slits in the boxes don't matter as long as the seal on top is intact proves complicity. Prior history has shown how easily the seals can be removed and replaced, or in some cases simply faked. In one infamous case a driver delivering ballot boxes was caught with a pile of brand new unused ballot box seals in his truck, which allowed him to remove the original seals, alter the contents of the ballot boxes, then put new seals on them prior to delivery at the counting center."
There was a lot more to the recount as I feel sure you are aware. The ease with which the Diebold machines can be compromised has been so well documented even on mass media that there is no reason to repeat it here. It is also a "filmed" fact that the machine counted Republican ballots delivered to the site of the recount arrived with seals broken and any number of other inconsistencies with "established ways of doing things." You choose only to tell one slanted part of the story. I can't help but wonder why.
Anticipating your next article, the chaos in Nevada with wrong areas listed on precincts, followed by several other last minute changes and even differences in the regulations as far as when you registered throughout the state are also documented.
Yes, I have a favored candidate; however, unlike you I can write about my feelings of despair as far as the impossibility of fair elections being conducted in this country without mentioning the candidate. It has nothing to do with him. If any campaign can bring fairness in elections to the forefront and make any headway in correcting this travesty, I would vote for that candidate for that reason alone.
You are far too intelligent to not know that the past two Presidential elections have been compromised and I cannot for the life of me understand why anyone who proclaims to be a libertarian would not be concerned. Beyond that your articles have been both a source of confusion and a blessing to me. I thank you for the blessing part.
First, the confusion. Your article a few weeks back where you laid out a totally ficticious reason for the reason Fox black-balled one candidate from a debate caused me to spread wrong information to a few sites. Pollyanna, here, read your credentials and was impressed. After all, you were a co-founder of the Canadian Libertarian Party among other impressive affiliations and accomplishments. I not only posted your articles, I defended them based upon your "credentials." I was wrong. Plain and simple. Fox chose not to include one candidate who should have been included for the American people to hear.
Then I read your apology yesterday. I saw no sincerity and I saw no real apology. The only reason I have returned to this site is because the site owner took the time to write to someone he did not know; who is not a "player" in this whole debacle. I'm just an American who wants a better life for my grandchildren. Walt's nice note to me pushed me to look into why Nolan Chart was formed. I believe that a site that can truly help people like me better understand all the divisions amongst these "collective groups" (of special interests who like to call themselves political parties) in order to better understand how we might someday work toward unity and a better life for everyone is a lofty, but important goal. I truly support Nolan Chart and what it stands for. However, you do not help me better understand anything and that is where the "blessing" comes in. I better understand now that you and others coming from "agendas" and "vendettas" , perhaps, have little to offer me as far as my continuing education in trying to better understand what can and should be done. Perhaps I do not need to understand Democrats, Republicans, Conservatives, Liberals, Constitutionalists, Green Party, and particularly Libertarians who seem right now to me to be the most infighting of all groups. Partially thanks to you, I'm all done with labels. I'm back to individuals trying to be better people and caring about others. Maybe I am the true libertarian and do deserve to be one of those gold stars at the top of the diagram.
After this latest article by you, I have to tell you that it appears to me that all of your background...all of your intellect...all of your energy is expended on "spin and slant" about what you are against. Sir, I would be more interested in knowing what you are for. Are you for anything, because as far as I am concerned you are nowhere on the Nolan Chart right now based upon your articles. Saying whom you are *not for* in cleverly cloaked, semi-articulate articles hardly educates me about the political divisions in society today. Some of us care not a whit about long-standing, petty arguments or power plays. This hardly addresses the very serious problems we are faced with today. In this article, you out did yourself. In fact for me, you outted yourself. I once was truly impressed by your credentials. Thanks for the eyes wide open blessing!
"Those who fear machine fraud or vote manipulation are going to have to search elsewhere for some hard evidence that voters need to be concerned about their ballots not being counted."
Um... I don't know about you, but I'd be more comfortable if the votes were on physical paper that, you know, can actually be seen. Having the votes counted on invisible microchips that are proven to be easily hackable just troubles me. Surely you'd want your vote on something that ANYBODY could count as opposed to something only "Diebold Customer Service Technicians" could confirm.
This article does not explain where the overcount and undercount numbers (66 and 440) come from. Those numbers simply do not appear anywhere else in the spreadsheet, and appear to have been pulled from a hat. I get 405 total errors for Hillsborough County. Discounting the 152 errors for Machester Ward 5 I get an error percentage of 0.6192%.
At least the method used here is correct. Others seem to focus only on how much the recount shifted the vote, not caring about the degree of randomness of the count.
The official margin for Florida 2000 was 0.009%, by the way. The actual margin is smaller, depending on which recount you use. So despite the fact that NH has probably one of the better election systems overall, if we had another close race we simply could not conclude with reasonable confidence who the winner was. I am thankful the two candidates had the guts to ask for a recount just because it needed to be done.
The article begins with the statement that there are reasons for caution. The paperless voting machines have properly been rejected by several states, though they will still be used in some counties in Florida. That concerns me, of course, but it wasn't the case in New Hampshire.
Yes, there does seem to be "chaos in Nevada," but it has nothing to do with compromised voting machines. As long as humans are involved in the process, there will be intentional and accidental errors.
There was prolific misinformation regarding the Fox debate. See "Ron Paul: Fair and Balanced" for a summary. I only reported the facts as I found them.
Every human being past puberty has an "agenda" of things that are important to them. I try to make my positions clear and present verified facts and civil arguments. Sometimes I fail, as noted in "An Apology To Readers."
Given my "credentials", it should be fairly easy for you to find out my agenda, which is simply stated: individual liberty. If you think a "serious problem" needs to be addressed, become a Nolan Chart author. You are certainly articulate and have your own "agenda" and passions.
Mike LaBonte: "This article does not explain where the overcount and undercount numbers (66 and 440) come from. Those numbers simply do not appear anywhere else in the spreadsheet ..."
The spreadsheet "works around" the hard data. If you load the spreadsheet into Excel, look at the far right columns in the raw data [Sheet 1] for totals and the summing formulas.
At this time, the recount is not complete. When the SoS publishes the final totals, I'll run the spreadsheet again and provide a link. Of course, the SoS is interested in whether errors actually change the outcome, not the small number of random human errors that occur in every election.
My question is, what is the statistical probability, that across 20 different areas, with variable under and over counts in each, across the three front-leaders, that a result would be obtained as follows:
Clinton: 71-59 = 12+
Obama: 46-34 = 12+
Edwards: 38-26 = 12+
The amazing symmetry in these balances, thus nullifying any movement between the candidates, to me seems improbable given the diverse base from which these figures had to come together to balance.
When people are absolutely convinced that the election is going to be stolen, nothing will ever convince them otherwise.
Anything and everything becomes evidence of fraud - including the absence of actual evidence of fraud. After all, that just proves that the conspiracy worked.
It's justifiable paranoia when it comes to the Diebold Voter machine no matter what candidate you're for. Americans want their vote counted accurately....statiticians can play with numbers to suit their purpose all they want but the real question is this...are voting machines safe. Why don't we leave that up to the experts like Computer Scientists who aren't getting their paycheck from a large corporation and not trying to justify a voter count with statistics produced from machines that can be tampered with. If you're truely scientists...then how can you rationalize that?
Call me paranoid, but I work with statistics and spend a lot of my time looking at patterns in datasets. The exact same overall balances for all three candidates, given the variability of locations and that more candidates than these three were on the ticket just seems like an improbable result. I would expect to see some variation, if only a little - not exact balancing in all instances. It just seems 'tidier' than random variability.
I believe your paranoia is well-founded given that these electronic voter machines are being contested through-out the United States in their ability to give an accurate account and missing memory chips add to the complexity and scrutiny of such a system.  Â
It's those types of stastical analysis that have prompted expert computer scientists to evaluate and finally withdraw their approval of Diebold Voter Machines. Â Perhaps the old saying is correct..."if it looks too good to be true, etc..." Â Â Let's allow real scientists to work out the bugs...not money mogels. Â
Yes, but the Diebold machines are one thing. However, these exact balances have come about from the hand recount in the midst of people talking about slits in the sides of boxes. The way the numbers are falling to say 'hey, yeah there was some counting error, but it didn't impact on the results' is 100% precise - not mostly precise - 100% precise. If you look at the numbers in each row they look variable enough - but when you come to subtract the overcounts from undercounts it's like an accountant has balanced profit and loss to come out with zero tax to pay. The variability in the rows might distract you, but the net effect - that's the interesting bit.
Yes, the hand ballot boxes have a peel back label on top that "can be" peeled back and replaced without speculation and then there is the question about slits in the sides of...
Well...we can put a man on the moon but we can't get this right. What's wrong with this picture? What's preventing this government from creating a fool-proof voting process? I think we have to raise these questions and insist on answers to solve this problem. The bottom line is that if we don't ... we'll be seen internationally as just another third world country when it comes to the electoral process ... if we're not already, in this regard.   If integrity has to be engaged by hiring a private, not-for-profit entity to create a precise, bug-ridden and hacker secure voting machine, tamper-proof voting boxes and reputable counters...it's a small price to pay for democracy.
Odd, in the raw data in excel file, there are Republican recount columns, for data shown R Paul comes in 3rd. Is a recount of Republican numbers coming automatically out of the Dems recount?
edie: "... I believe your paranoia is well-founded ..."
Note that there are two different electronic systems: voting machines and vote tabulators. Diebold (and others) offer both to states and the election officials make the choice.
The optical scanners used in New Hampshire only count (proper) marks on paper ballots, which are preserved for audits and recounts. The all-digital voting machines have no such physical record (just electonic bits) and are subject to all kinds of intentional and accidental errors.
Kleeb: "... seems like an improbable result. "
If the human errors are random, then the probable result is that the correct counts will match the original distribution. The precise corrections aren't known, because recount results aren't yet finalized (probably not until Monday), so don't take the literal numeric match as improbable.
ron: ".. is a recount of Republican numbers coming ..."
As noted on the SoS website, these are write-in votes for Republican candidates on Democratic ballots. All votes are counted. There is no recount of the Republican ballots.
Kleeb, trust your instincts and your experience. Westmiller exhibits an average American mindset. If you go to the trouble to actually review the CNN election count as it appeared LIVE on CNN, you can see that Hillary's lead was constant and steady; there was no apparent variation. You can actually see Kucinich and Ron Paul's count go UP AND DOWN!!!!! It's a matter of videotaped record. I didn't once see Hillary's or Obama's count EVER GO UP AND DOWN. Hillary's margin remained steady ALL NIGHT LONG! Do you expect me to believe that voters were voting in such a symmetrical fashion????
'splain that!'splain that!'splain that!'splain that!'splain that!'splain that!'splain that!'splain that!'splain that!
Hmmm curious - as are the comments made by a woman named Bev - see this link, particularly the comments at the bottom of the page - where a chap shares what she posted on a forum:
The article reports the facts about the recount and whether it showed any significant errors from the optical scanners. I don't see any evidence of those kind of errors. The recount showed that the Diebold scanners were accurate when they were used correctly.
What the article never attempted to discuss is all of the possible human errors. For example, some cardboard storage boxes were delivered to the polling places in strapped bundles. Some of the straps were sliced with a box cutter and produced superficial "slits" in the top boxes. That's human error, not evidence of malicious vote tampering.
edie: "... and the point is....?"
The intention of the recount was to discover whether there were significant errors in the machine-scanned count. None were discovered that had any significant affect on the candidate results. Some human errors were found, but the fears about massive errors or manipulation by the Diebold machines were not justiied. End of story.
"The memory cards used in the Diebold optical-scanners are apparently "missing in action". Some have reportedly been returned to LHS, the company responsible for maintaining the Diebold machines in New Hampshire, and may have been erased already (17th Jan). Federal law requires all materials from elections be preserved for 22 months after the election. "
http://whatreallyhappened.wiki.com
Human error or malicious voter tampering ? Does it matter?
What matters to many Americans is that it "can" happen. If they can eventually piece together this mess and then claim it was 'fairly' accurate..that just isn't good enough.
What it now shows is that there are more holes in our election process (AGAIN) and that probabilities and outcomes 'can' be manipulated...
Perhaps you can enlighten me and assimulate the statistical outcome with the missing chips...
I'll anoint you a god....because I'm sure you're going to try.
We're all human and humans make mistakes. Bear in mind that an election is a once or twice a year "job" for poll workers and it would be foolish to expect every one of them to do everything exactly according to instructions. That's why results are sampled and audited. The Demos study indicates that election fraud allegations drop with the introduction of more technology:
The memory cards are irrelevant after the machine is initialized. They only contain the candidate names for each office, not vote results. Of course, data security and validation is important, but missing memory cards after the fact have no bearing on the vote results.
The recount shows that the tiny number of vote errors were human (damaged ballots, wrong pens, manual transcriptions), not a result of any equipment or program defects. Of course, I'll defer to the NH SoS audit of the process, which may result in revisions to their tabulation procedures.
We must all demand 100% publicly hand-counted paper ballots nationwide immediately! There is no other way to insure that our elections are free and fair.
Want to comment on this
article? Leave your comment here. Your email address is
required to track your comment. However, we will neither
publish your email address nor distribute it to other
organizations or persons. The only reason we might use
it would be if we needed to contact you regarding your
comment. All comments are subject to our
terms of use policy.
2008 William Westmiller, all rights reserved.