« Huck Uses "The Guy Who Laid Them Off" Line In New Michigan Ad | Home | Team Romney: We've Got "A Clear Path To Victory" »

ARG: We Had Hillary Closing The Gap — To Nine Points

One particular pollster, New Hampshire-based American Research Group, has perhaps the most interesting explanation for how they got their state's Democratic result wrong. ARG put up a statement up on their site, saying the following:

While we missed the final number that Clinton would make in New Hampshire, our polling was one of only two daily polls that showed Clinton regaining support following her drop in New Hampshire the day after the Iowa Democratic caucus. Clinton was moving up in the final days and hours before the primary, and our polls and the Rasmussen polls were the only daily polls to catch Clinton's rebound.

...

Our polls missed the final Clinton number, but we did not miss the strong swing back among women reacting favorably to Clinton that started after the debate and continued with her comments in Portsmouth. We did not have a polling problem, we just ran out of time.

It is true, actually that ARG showed Hillary closing the gap. Their second-to-last poll had her down by 11 points — while the final one had her down by 9. Keep in mind that these were not composites of samples over three days, as in the case of John Zogby, but were instead conducted over 24-hour periods.

Congratulations, ARG.


21 Comments

| Leave a comment
user-pic

Both ARG and Zogby's statements tell me that clinton's internals told her it was a tight neck and neck race. Gee, I wonder why the clinton people were spewing about a big loss and campaign shake up and they were 11 points down, etc, etc. Politics as usual.

user-pic

RE: Voter Fraud in New Hampshire

It looks like voter fraud may have been responsible for the Clinton win in NH. The Obama camp has been approached by a number of parties that are asking him to request a recount of the votes.

There is a statistically significant discrepancy between exit polls, actual votes counted by hand in the various voting districts, and those tabulated by Diebold tabulator machines. The exit polls and hand counted votes indicate Obama should have won NH by at least 2% to 5%, but the votes tabulated by the Diebold machines (the ones fingered in the 2004 election fraud) had the result flipped to favor Clinton. In addition, several pre-election polls were right on the money for the results of ALL the other candidates in BOTH the GOP and democratic race – except the Obama/Clinton match-up and Ron Paul.

Also, NONE of the Ron Paul votes were counted. As of right now, NH is still reporting that Ron Paul got NO votes, which is not true. He was polling at 5% to 10% before votes were cast.

The Centre for Research on Globalization:
http://www.globalresearch.ca/index.php?context=va&aid=7753

Boston Now:
http://www.bostonnow.com/blogs/boston911truthorg/2008/01/09/major-allegations-of-vote-fraud-in-new-hampshire

The Ben Mosely Blog:
http://benmoseley.blogspot.com/2008/01/do-nh-primary-statistics-show-election.html

Product Reviews
http://www.product-reviews.net/2008/01/09/new-hampshire-vote-fraud-confirmed-ron-paul-votes-not-counted/

Malta Star:
http://www.maltastar.com/pages/msFullArt.asp?an=17896

user-pic

Too bad you ran out of time. I am sure by Wed. Morning you would have gotten it right. LOL

user-pic

ARG was off by a mile in Iowa.

user-pic

"Too bad you ran out of time. I am sure by Wed. Morning you would have gotten it right. LOL"

I think the right thing to do would have been to have done a much larger sampling on Monday and then get it right. This sort of post-election whining isn't attractive. But ARG isn't alone.

user-pic

Nice to see the pollsters, media pundits, and blogs patting themselve on their glory-covered backs for their stellar performance over the last five days.

user-pic

It is nice to see the media with egg on their faces.

However, this changes nothing. Obama is still my candidate and will be until the nominee is decided.

user-pic

Updated: 5:05 AM (EST) - Results tallied for 209 out of 236 of the municipalities.

By Percentage

Method Hillary Clinton Barack Obama
Diebold Machines 53.23% 46.77%
Hand Count 47.47% 52.53%

By Votes

Method Hillary Clinton Barack Obama
Diebold Machines 82860 72807
Hand Count 18898 20912

By Number of Municipalities Won

Method Hillary Clinton Barack Obama
Diebold Machines 54 33
Hand Count 43 77

About 81% of the votes will be "counted" by the Diebold machines.

user-pic

RE: New Hampshire Voter Fraud

Obama won the hand counted votes, but lost the machine tabulated votes. Clinton lost the hand counted votes but won the machine tabulated votes. The machine tabulated votes were flipped to favor Clinton.

http://www.opednews.com/articles/opedne_ron_corv_080109_new_hampshire_electi.htm

user-pic

The head clerk of the New Hampshire town of Sutton has been forced to admit that Ron Paul received 31 votes yet when the final amount was transferred to a summary sheet and sent out to the media, the total was listed as zero. The fiasco throws the entire primary into doubt and could lead to a re-count.

As we reported earlier today, an entire family voted for Ron Paul in Sutton, yet when the voting map on the Politico website was posted, the total votes for Ron Paul were zero.

Vote fraud expert Bev Harris contacted the head clerk in Sutton, Jennifer Call, who was forced to admit that the 31 votes Ron Paul received were completely omitted from the final report sheet, claiming "human error" was responsible for the mistake.

Two or three votes not counted could be a plausible mistake - but 31 votes for one candidate?

"The classic method for rigging a hand count is to write the wrong number on the form," Harris told the Alex Jones Show.

"They are counting everything in public real nice, they fill out a form in public real nice and then they transfer it to another form and they call that a summary sheet and then that is the one they send in," explained Harris.

"What happened is she said they did not transfer the number correctly and put zero instead of 31 - that is unacceptable as an answer."

With 100% of precincts now reporting, the map originally listed zero votes for Ron Paul as you can see below. It has now been updated to reflect the 31 votes Paul actually received.

The remainder of the 31 people in Sutton who voted for Ron Paul need to go public immediately with the charge of vote fraud and make it known that they were cheated out of their right to vote.

Harris estimates that it could cost the Ron Paul campaign as much as $67,000 dollars for a recount, but such a move could throw the entire primary into doubt, especially in light of the fact that Barack Obama appears to have been cheated out of a win by Hillary Clinton.

user-pic

Has everyone seen Krugman's take on last night's news? Worth checking out if you have not.

user-pic
user-pic

Not this Diebold crap again...

Hey! I heard that a missile hit the Pentagon on 9/11! And the moon landings are fake too!

user-pic

This is actually true as far as it goes. Rasmussen had the same thing with his two last polls, the earlier of which had Obama ahead by 10 and the last one, 7.

There were aspects of other polls had gave some whiff of Obama leveling off at least, if not falling back a little -- I was watching the percentages of independents planning on voting in the Democratic vs. Republican primaries in the UNH polls in particular. But it was all within the MoE's except for the Indies opting for the Dem race, which jumped back to 60% on Monday, so I tended to dismiss it as statistical jitter and perhaps a little wishful thinking. I also figured they really were picking up on something real, it was likely to be too little, too late for Clinton.

user-pic

RE: New Hampshire Primary Voter Fraud

Hillary LOST the paper ballot count but WON the optical scan ballot count. Obama WON the paper ballot count but LOST the optical scan ballot count. The machine tabulated votes were flipped to favor Clinton. See for yourself.

http://www.opednews.com/articles/opedne_ron_corv_080109_new_hampshire_electi.htm


http://ronrox.com/paulstats.php?party=DEMOCRATS


Voter fraud.

user-pic

ARG is an IVR poll, right? I'm pretty sure they are. If so, that means the two IVR polls, ARG and Rasmussen, both had Clinton closing the gap. This points to a mild "Wilder effect" in live-interview outfits. People lie to computers less than they lie to live interviews, as there is no social pressure. Probably a 2% Wilder effect in New Hampshire.

user-pic

Take a deep breath

I have no way whatsoever to know if massive fraud did or didn't happen last night in NH. But I do know that the idea that the results were so at variance with the last predictive polls prior to the election, that there must have ben fraud, is absolutely without foundation.

Predictive polls from the major polling outfits have a track record of pretty good accuracy for one-on-one, Democratic-on-Republican races, but of wide variability and unreliability for races that don't meet these criteria. NH was both intraparty, and had more than two serious contenders.

General elections easier to poll more accurately than primaries because the subset of registered voters who vote in general elections is much more stable and predictable year to year than the subset that votes in primaries. When you do a predictive poll, you have to come up with another subset of the registered voters, the likely voters who can be reached by telephone and who will respond to your poll questions. Making this subset perform the same, or at least close to, the subset of registered voters who will actually vote, requires some correction and tweaking of the raw data that the pollsters know how to do tolerably well when the target they are modeling is the predictable general election voters, but which they have to make huge guesses when they are trying to model the much less predictable profile of the voters who will decide to turn up this year in a given primary.

Races with more than two viable candidates suffer from tactical voting and tactical poll responses. In a two person race, people both vote and respond to polls for the candidate they prefer over the other candidate. But if there are more than two candidates, they have to consider, besides which one they'ld like to see win, which one of their top choices has a chance to win. If you rank the candidates in order of preference as A, then B, then C; you might very well vote B if you have reason to believe that A has little chance of winning, and you don't want to waste your vote on a losing candidate, and lose your chance to keep the hated C out of office. The same dynamic can apply to answering pollsters, or the voter could reverse the dynamic, saying he will vote for his favorite A, because A is his favorite, even though he knows that he will vote tactically for B.

user-pic

SLKRR wrote on January 9, 2008 3:57 PM:

Not this Diebold crap again...

Well the fact is that there has already been one mistake corrected! And to claim that there could not be others is to ignore the CBS report:


(CBS) We’ve had four years to straighten out the voting systems that led to the election debacle of 2000. And many counties have rushed to replace those infamous punch-card systems with what’s supposed to be the wave of the future — electronic voting.

In fact, there are now so many of these computer voting terminals it's estimated that up to 45 million voters will cast ballots on them next week.

With no less than the presidency at stake, you might think the software behind this technology has been perfected. But, as Correspondent Scott Pelley reports, there’s concern about whether bugs in the software could make the computers vulnerable to error and tampering.

More than $300 million is being spent on new machines, but are we any better off than we were four years ago?

Remember 2000? In Florida, angry partisans invaded a county election office. They were forced back in a scene that looked like a country having its first election. No election official in the nation wants to hear this at their door.

On Tuesday, election officials hope to convince voters in much of the country to switch from fists to fingertips and light up their choices on a touch-screen computer.
How are they working so far? "The voters love them. The surveys we’ve had, as well as other jurisdictions around the country, voters flock to them. They’re big print, easy to understand, easy to use," says Conny McCormack, who runs elections in Los Angeles County. She's been using
the terminals in early voting for four years.

McCormack showed 60 Minutes Wednesday some of the advantages. "In Los Angeles, you have a choice of languages [Chinese, Vietnamese, Korean, Spanish]," she says.

There is also a screen that shows the voter his ballot and lets him correct mistakes at the end.

There are about a dozen manufacturers of electronic voting machines, and they all have one thing in common: they record the vote in the computer’s memory. But the drawback is there’s no way for the voter to verify on paper that his vote was recorded accurately -- there’s no paper trail, which worries critics.

The machines are being set up in 29 states, and nearly a third of the nation’s voters are expected to use them on Election Day. One of the places they’ll find them is Palm Beach County, Fla., which was, in the last election, the home of the hanging chad.

"I think it's going to have a bigger problem than it had in 2000," says Charlotte Danciu, a lawyer whose father, a long-time politician, ran for city council in 2002. It was one of the first elections in which Palm Beach County used the new machines. He lost in an upset, but what upset Danciu even more were the calls she got from voters after Election Day.

"There was this elderly woman on the phone and she’s like, 'Miss Danciu, you don’t know me, but I kept trying to vote for your dad last night. I kept pushing his name on the screen and the opponent’s name kept registering,'" says Danciu. "And they would call over a poll worker, who would do such things as punch the machine, hit the machine, unplug the machine."

Not only that, but in a small runoff election in the same county, with only one race on the ballot, officials came up 78 votes short. Palm Beach County says 78 people who signed in decided not to vote.

"I don't think people would take off from their day and go there with one choice and make no choice," says Danciu. "I think that that's ridiculous. I think these computers did not record the vote."

After the election, Danciu brought in an expert to inspect the machines and she brought a photographer along. When the expert pressed two names at the same time, the computer lit up a third name in the middle. The expert said the machine should "detect that I'm pressing two things."

Election Supervisor Theresa Lapore, who spent $14 million on the machines, complained it was a trick: "You're just trying to trick the machine. Normal persons wouldn't do that."

But the expert was trying to trick the machine. And she did. The maker of these machines, Sequoia Voting Systems, declined to talk with 60 Minutes Wednesday on camera. But the company said it knows about this trait in its system, and the voter can correct any mistakes he makes before recording the vote.

Still, many computer experts are concerned about the software used in electronic voting machines.

"It is not possible for humans to design software that is that complex and is perfect," says David Jefferson, a scientist with the Lawrence Livermore National Laboratory. He's the technical adviser on electronic voting for California.

"Every software application you have ever used in your ordinary life using a PC, you run into bugs all the time," says Jefferson. "Either the software will do something wrong or it will just plain crash. … They [voting systems] have received quite a bit less scrutiny than most of the software that you run every day on your PC."

They have received less scrutiny because the computer makers consider their election software to be a trade secret. Danciu asked to inspect the software from her father’s election to see if there were bugs, but Palm Beach County said no.

"They even took the position that if they were to disclose any of it to us, that they would be breaking the law and committing a felony," says Danciu.

"The situation that we are in today is that the voting system software is basically secret, basically a very dark, deep, dark secret," says Jefferson. "That is to say only a very tiny number of people in the United States have been allowed to inspect this software."

The software in electronic voting machines is unbelievably complex, and it's supposed to meet federal standards. But the computer manufacturers don’t let just anyone look at it. They pay testing labs to certify the software, and neither election officials nor the public typically gets to see the thousands of lines of computer code that make up the software itself.

That is, until last year, when one version of a voting program leaked out into the public by mistake. Diebold Election Systems, one of the largest manufacturers of electronic voting machines, accidentally put its software on the Internet briefly and Avi Rubin got hold of it.

"Within one hour, we had found some unbelievable security problems, and we knew at that point, our hearts started beating fast, and we knew we were sitting on some kind of time bomb," says Rubin, who is technical director of the Johns Hopkins University Information Security Institute.

Rubin, also a professor with a Ph.D. in computer science and engineering, said the program he saw last year used security software that was obsolete.
How would he characterize the software? "It was far below state-of-the-art. I would say it was just, they had absolutely no idea what security is about," says Rubin.

How difficult would it be to tamper with one of these machines? "I don't think it would take much sophistication," says Rubin. "I think a teenager who meddles with computers and has some programming experience could do it."

Diebold says that the program Rubin saw was two years old, that it's a work in progress, and that it's never been used in an election. Still, Rubin’s study was the first public analysis of voter software, and it sent a shock through the states.

As a result, the state of Maryland did its own studies and a Maryland expert was able to pick the lock on a machine, attach a keyboard, and get to the point of changing the votes. The Maryland study concluded there are "considerable security risks that can cause moderate-to-severe disruption in an election."

Maryland decided to go ahead with the machines after Diebold promised security improvements recommended by the state experts.

Still, McCormack says tampering isn’t likely in the real world: "Don't you think a poll worker might notice if someone came in and started trying to put, plug in a piece of equipment? I think that would be pretty noticeable in a polling location."

In Los Angeles County, and many other places, they put tamper-evident tape on the machines. They test the terminals to make sure they’re working properly and never connect them to the Internet.

But Jefferson, California’s state expert, told 60 Minutes Wednesday he worries not so much about hacking into one machine, but the possibility that a rogue programmer could tamper with software and corrupt the votes in thousands of the terminals.

"The electoral weapon of mass destruction, if you will, would be if you were able to modify the software at the source, where the vendors write it, and insert either a bug or some piece of malicious logic that was distributed to every state that used electronic voting machines," says Jefferson. He calls voting a national security issue, and says voting machine programmers should have security clearances.

Would you know if somebody modified the software?

"No. I think the important point is that if you were interested in doing this, you would design it in such a way that it would not show," says Jefferson.

McCormack, however, says that while security could be improved, much of the concern is overblown. "There's no question the security could be enhanced. And I’m not trying to say that it couldn’t be and I think that’s true in all aspects of our life," says McCormack.

Does she believe that security enhancements on these machines are called for? "I had no problems with them. And so, everything in life could be more secure. Of course it could be," says McCormack. "But the measure to me is whether or not there's been a problem with the accuracy. Have votes been stolen? Have there been any proof of evidence? There hasn’t been evidence."

But critics say that's the problem. With electronic voting, there might not be any evidence of tampering. On election night, the computer reports hard numbers, but the question is did anything strange happen in the complex software that lies between the fingertip and the memory chip?

Danciu said there's no way to be sure, because she believes there's no way to do a real recount.

"The Palm Beach supervisor’s position was, 'Well, when you push this button, the computer will recount.' Well, it just retabulates and spits out in a nanosecond what it said the nanosecond before. There is no recount. There’s no physical evidence to recount."

"You're essentially running the same data through the same software on the same computer, you’re gonna get the same answer every time," says Pelley.

"You are gonna get the same answer every time," says McCormack.

Is that a recount? "Oh, I think it's a recount. And you know, do people really want to get a different answer? What we saw four years in Florida was a recount that was done, where people got a different answer, chad fell out and the numbers were different," says McCormack.

"This shocked people. The recount doesn't match, and yet, in electronic, the recount matches and everybody's critical of that. So I don't think there's anything in between. It's either gonna match, or it's not gonna match. There's criticism at either level."

With the election predicted to be close, and up to 30 percent of the vote being cast on electronic terminals, that recount issue may make the computer chip this year's hanging chad.

As a precaution, Nevada is equipping its computer voting terminals with a paper record so that voters can verify their ballots.

Diebold declined an interview, but the company told 60 Minutes Wednesday its equipment is the most accurate and secure on the market, and greatly reduces the errors of older voting machines. The company says it fixed the bugs that were found by Rubin and the State of Maryland. Still, at least in California, there will be old-fashioned paper ballots available for anyone who doesn’t want to use the touch-screen machines.

As to the ad-hominen insinuation:

Hey! I heard that a missile hit the Pentagon on 9/11! And the moon landings are fake too!

I heard that you are a dirty sock sucker, preferring to dive into a dirty laundry basket and vociferously suck a dirty sock, instead of course a clean one from the clothes dryer, but that of course has nothing to do with counting votes.

user-pic

Oh, crap! Where do these conspiracy nuts come from? It's one thing to rely on actual reports to go to bat on the voter fraud issues. It is quite another to whine and holler because your candidate lost. Give us all a break and go vote for Ron Paul, or something.

A couple of points that really ought to put this issue to rest. First, the "same Diebold computers" are not quite the same. The vote-flipping accusations were leveled against touch-screen machines. This is not what was available in NH. Some NH counties had optical scanners--something that most people believe is more reliable--but these were tabulated by Diebold computers (ones that have been previously identified as easily hackable) and run by a NE Diebold distributor that BradBlog pans as something between incompetent and criminal. Fine. Get some evidence, then raise the point--the problem is, it's only questions about the tabulators and the outfit that runs the machines, but no evidence of any kind, just suspicions.

Second, there is an obvious and trivial explanation for the differences between the scanned results and the hand-counted ones--the demographics are completely different in different parts of NH. Look at the exit polls (yeah, yeah...). The larger the towns, the more likely they were to go to Clinton. Obama carried most of the rural precincts, was even in small towns and lost the large ones. And there was a clear split between Northern (more rural) and Southern (more urban--if such a concept even applies to NH--and more stacked with Massachusetts expats) NH. Obama did better in the North, Clinton in the South. The optical scanners were predominantly found in the South and in the larger towns (but not universally so).

Furthermore, if you look at the vote by county, town and precinct, you will see that there is no pattern--Obama won some large counts with optical scanners, but lost some small hand counts. It was a mix with no particular pattern. There is no evidence of any kind, unlike Florida in 2000 and Ohio in 2004. There are no reports of irregularities--except for Drudge and Fox generated ones that were debunked by even conspiracy-inclined BradBlog.

There is simply no reason to believe that there is any problem regarding the NH vote other than the outcome that a lot of people don't like. So STFU!

user-pic

Obama also has a better philosophical sense of what is right for America

"We do have a philosophical difference. John [Edwards] and yourself [ Hillary Clinton] believe that if we do not mandate care [...] if the government does not force taxpayers to buy healthcare, that we will penalise them in some fashion. I disagree with that because as I go around town hall meetings, I don't meet people who are trying to avoid getting healthcare; the problem is they can't afford it. The costs are too high. And so, as a consequence, we focus on reducing costs."
- Barack Obama,
Democratic Debate, 5 Jan 2008

Barack gives liberal policies the libertarian sensibility (i.e. not nanny-state but a fair leveling of playing field) that Americans can welcome.

user-pic

Obama also has a better philosophical sense of what is right for America

"We do have a philosophical difference. John [Edwards] and yourself [ Hillary Clinton] believe that if we do not mandate care [...] if the government does not force taxpayers to buy healthcare, that we will penalise them in some fashion. I disagree with that because as I go around town hall meetings, I don't meet people who are trying to avoid getting healthcare; the problem is they can't afford it. The costs are too high. And so, as a consequence, we focus on reducing costs."
- Barack Obama,
Democratic Debate, 5 Jan 2008

Barack gives liberal policies the libertarian sensibility (i.e. not nanny-state but a fair leveling of playing field) that Americans can welcome.

Leave a comment

Recommended Reader Posts

  • Unwritten...
    by stillidealistic
  • BABIES, RACCOONS AND HEALTH CARE
    by dickday
  • Two Dreamers, by Dorothea Lange
    by Rutabaga Ridgepole
  • Tsunami Wave: Will Wipe Out Republican Party
    by coonsey
  • OBVIOUSLY, YOU AREN'T A HUNTER.....
    by wvbiker
  • The Stupack Amendment played politics with women's lives and won.
    by J. Clarence
  • wooden projects
    by kubaser
  • holly colorado
    by blumun
  • boxes generator
    by boluwel
  • short stories
    by lumacer



  • Resources

    The Palin Effect

    GOP Map Of Sleaze

    Tire Swinging

    The Final Debate

    World of Sleaze

    All About Sarah

    The Presidential Debates: Round 2

    The Vice-Presidential Debates: Biden v. Palin

    Critic or Cheerleader The Definitive McCain Iraq Timeline

    The John McCain John Hagee Timeline

    Masthead

    Editor-in-Chief
    Josh Marshall

    Reporter-Bloggers
    Elana Schor
    Eric Kleefeld



    Subscribe to this blog's feed.

    Advertise Liberally
    Share
    Close Social Web Email

    "To" Email Address

    Your Name

    Your Email Address