What happened in New Hampshire? Nader says the POLLS were wrong? TIA??

[helderheid]

Someone here asks:

http://www.inthesetimes.com/site/main/article/1970/P240/

“More than 2,000 people and organizations begged Nader to request a recount after a statistical analysis posted on the Internet showed some New Hampshire precincts using the Accuvote machines gave President Bush up to 15 percent more votes than expected, based on exit polls and the 2000 presidential vote.” AP Nov 19th

http://www.seacoastonline.com/2004news/rock/11192004/ne ews/49585.htm

“It looks like a pretty accurate count here in New Hampshire,” said Michael Richardson, Nader’s representative.

Nader campaign officials have said the recount could expand to other precincts, or even other states, based on the results. But no candidate’s tallies have changed enough to affect overall percentages so far.

“There doesn’t seem to be any error due to fraud based on what we’re seeing here today,” Richardson said. AP, Nov 30

http://www.nashuatelegraph.com/apps/pbcs.dll/article?AI ID=/20041130/NEWS02/111300050/-1/news

New Hampshire has a paper ballot trail and the recount showed that the exit polls were wrong there (According to the exit polls, Kerry should have won by a much greater margin than he did in New Hampshire.)

Does anyone know why the exit polls were wrong in New Hampshire?

Posted by JA on February 19, 2005 at 3:29 AM

TIA or anyone, do you know?

[Stevepol]

1. How can anybody say the exit polls were wrong, when

only 11 precincts were recounted by hand? They might have been wrong, but if they were, that just raises other questions about how the polls in NH were administered as opposed to the way they were carried out in other states and internationally. The exit polls are almost never wrong. There could be many reasons that the exit polls in NH do not match the so-called actual results. There are many ways to jigger the numbers when you're using central tabulators to figure the final results. Among many other possibiliites, there could have been collusion between some of the precinct officials and some of the "technicians" handling the scanners or the tabulators. Why assume the exit polls were wrong? In general, exit polls are better now, more accurate, than they ever have been. I feel 99% certain that Kerry won the election by about 5,000,000 votes as the exit polls indicate. The demographic evidence and just common sense would lead to the same conclusion.

[Mistwell]

7. The exit polls are ALWAYS wrong

"The exit polls are almost never wrong"

HA! Good one.

Raw exit poll data, for the prior four Presidential elections (which is, by the way, since there WERE exit polls), have always overstated the Democratic vote, sometimes by more than they overstated for this election.

http://www.washingtonmonthly.com/archives/individual/2004_11/005178.php

Year / Exit Poll / Results / Dem Lead / Dem Actual
1988 / Dukakis: 50.3% Bush: 49.7% / +0.6% / -7.7%
1992 / Clinton: 46% Bush: 33.2% / +12.8% / +5.6%
1996 / Clinton: 52.2% Dole: 37.5% / +14.7% / +8.5%
2000 / Gore: 48.5% Bush: 46.2% / +2.3% / +0.5%

"As you can see, the raw exit poll results always overstate the Democratic vote, sometimes by as much as eight percentage points. So the fact that the raw results this year overstated Kerry's actual vote tally is hardly cause for alarm."

The longer this myth that "The exit polls are almost never wrong" is perpetuated, the longer it will be until we spend our time and resources on proof of actual fraud rather than easily-disproven theoretical fraud.

[Wilms]

18. I was breach birth, so sometimes I get things backwards.

Your post leaves me "alarmed" 2000/2004 were not the only stolen elections. :shrug:

[Mistwell]

25. Dukkakis won?

Oh come on. You have to believe the following to believe all prior elections were stolen:

1) In 1988 there was no common software source for counting machines, and NO fully automated electronic voting systems. Indeed, the counting machines often doubled as school test scoring machines during the rest of the year, and came in MANY models (things were more competative back then). So, you have to believe a massive conspiracy took place, to reprogram thousands of vote counting machines, to believe this.

2) You have to believe that, with thousands of people in on the giant conspiracy, NOBODY talked for over 15 years about the issue, even when the topic became hot this year and in 2000.

3) You have to believe that thousands of people were willing to risk prison sentences for felony vote tampering on faith that everyone else in the conspiracy would do what they were told as well.

4) You have to believe that, despite this monumental effort, they FAILED in their goal EVERY SINGLE TIME. In 1988, they failed in that they would have had the electoral college locked even without the tampering. And then, they failed again 1992, and again in 1996, and again in 2000, when despite their massive secret efforts they failed to capture the popular vote through this election tampering in each of those years.

Come on...is all of that really believable? Is it more believable than the simple fact, which all electoral experts agree on, including the self-professed left-wing inventor of the exit polls, who ALL say that the exit polling process is just not good enough to predict a winner? And they DO all say that. They will all tell you (except those non-electoral experts who are instead just statisticians in other fields with a clear agenda when talking about the elections industry) that an insufficient amount of data is gathered to actually predict how the vote will come out, and that the exit polling is ONLY useful when used to describe demographic data after the vote, after being poured through a formula that includes adjustment based on the actual counted vote. That disclaimer is, in fact, right there in all the exit polls.

Face it, the exit polls just don't work for the thing we are trying to use them for. Exit polls in the US don't function to detect fraud.

I think there WAS fraud in this last election (though I am unconvinced it was enough to change the result, so far). However, I just don't think our current system of exit polls could possibly detect that fraud. It wasn't meant to, and it is always wrong when tested for the "how people voted" data.

[Wilms]

29. Actually, I don't "believe" much.

Certainly not that it would take thousands of people to rig an election. Or that anyone so involved would be able to finger any/everyone else involved. Or that they'd not interview about it.

Further, I don't "believe" that election tampering would be used exclusive of insuring/padding a victory. Or that it always worked. Or that it was always attempted, or never was.

I'm assuming Mitofsky is the "self-professed left-wing inventor of the exit polls" to which you refer, but I'm not certain of his motivations. I don't find "all electoral experts agree" persuasive when many experts lined up to question the 2004 Poll.

I don't even "believe" the exit poll debacle proves fraud.

I don't "believe" a damn thing I can't verify. So I'm free to wonder and ask questions.

I do "believe" that research using these data can detect anomalies. That they can inspire others to wonder. That it can support those who would call for a recount. That it would lead to folk wondering why we are increasingly using equipment (ballot-less DRE's) that don't provide for meaningful recounts.

I "believe" TIA's and other's work to be among the pillars of the Election Reform Movement.

And I agree with you that there is other important work to be done, as well.

And I do believe we can support all these efforts as constructive contributions toward the overall aim.

[southwood]

31. Oh no, not again!

Our insights have increased since November 17 (date of your link), not in the least because of Edison/Mitofsky's own evaluation of their exit polls:

http://www.exit-poll.net/election-night/EvaluationJan19...

See page 34.

For Within Precinct Error (WPE), which is what matters here because that could potentially point at "irregularities", differences between exit poll and tally are small (2% in favor of GOP)for the years 1988, 1996 and 2000. This difference can roughly be explained from the greater probability for a Democratic vote to be spoiled. (This is why Greg Palast could claim before the election that Kerry was already 1 million votes behind.)

For 1992 (Clinton vs Bush Sr.) the Democratic "overstatement" in the exit polls was 5%, and for 2004 6.5%.

The authors conclude: "What this means is that the errors in 1996 and 2000 were more random, while the errors in 2004 were much more in one direction". Favoring Bush Jr., that is.

The rest of the differences between exit polls and apparently lies in the imperfect choice of polled precincts in some years (but not in 2004. Last year the error could be attributed to WPE only.)

Edison/Mitofsky have no serious explanation for their WPE in 2004. They tried the "Reluctant Bush Responder" theory, which was quickly shot down by Freeman on statistical grounds derived on data from Mitofsky's own report:

http://uscountvotes.org/ucvAnalysis/US/USCountVotes_Re_...

Making things worse, Edison/Mitofsky have released their data on the 2004 exit poll, but not down to identifiable precincts, thus precluding independent researchers to verify their analysis, to find alternative explanations for the difference in WPE or to actually compare precinct tallies to precinct exit poll results. The latter would have been great fun, because that way "hot spots of irregularities" at the precinct level might have been found.

[Amaryllis]

17. I was told by someone who knows how central tabulator manipulation works,

that if they only counted precincts, and not counties (and they did only count precints) that they wouldn't find the problems. He said they have to count entire counties to find the discrepancies, not just individual precincts, and if they had done this, the result could have been quite different.

[seaclyr]

21. Amaryllis, do you have more info on how that would work?

Or a link, perhaps.

[Amaryllis]

23. Kip Humphrey, who posts on DU, told me, and I just asked him to post

more about this. I don't understand it well enough to explain further, but he is an IT guy and has worked with this stuff for several years now.

[helderheid]

26. Thanks!

Please keep us posted :)

[hereinmissouri]

2. So, what is the theory?

Were the 11 precincts chosen at random? I've been wondering about NH ever since the recount. I hope someone can answer this.

[hedda_foil]

3. No, they weren't chosen at random. They were the most skewed ones.

DUer IdaBriggs did the statistical work to ID the most suspect wards ... the ones with the greatest red shift. Ida, members of the NH Ballot Integrity Project, Democracy for New Hampshire, New Hampshire Greens and Democrats all observed every moment of the hand recount of all the wards, along with the Nader people. The results of the hand recounts in all 11 cases were within a couple of the original vote counts. And every ballot in every one of those wards was recounted entirely by hand. There was no machine recount.

Does this mean that there might not have been fraud in other parts of New Hampshire? Of course not. But the selection, recounting and observation processes were clean.

[hereinmissouri]

4. So......

if I understand correctly, this means the exit polls were wrong in these 11 precincts? If this is so, it makes me question the exit polls in general. Any theories on this?

[IdaBriggs]

9. Go Here for the Story

It was my public report to DU: http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=97506

You can also read the stuff I put up at www.invisibleida.com for more details.

:) Best, Ida

[Stevepol]

19. Ida, if you see this, could you answer a question for me?

Did you, or did the people you were associated with in doing the recount in NH, use the exit polls as a criterion for your choice of which precincts to recount? My memory of the process is that you used other criteria besides the exit polls to settle on these 11 precincts, primarily the way the precinct voted in 2000. Am I right? If I am, do you know what the exit polls showed for those precincts you recounted?

I would be interested to know this since I believe the exit polls should be the best way to determine when miscounting occurs. All the guess work about previous voting patterns and sampling the populations that will actually turn out on election day is taken out of the equation with the exit polls. It's about the only trustworthy check available at the moment when the machines are either un-auditable (the touchscreens) or almost never recounted (the scanners and the other machines that use computer programmed machines to count the votes but which do have a paper trail) and when the results run counter to both mathematical modeling and common sense. Clearly, the smaller the population (as in NH), the greater the margin of error. But the margin of error is always a factor.

Just curious. I remain in awe of your patriotism and dedication and initiative in helping to get these recounts.

[skids]

20. The exit polls don't show individual precincts...

Even with the raw data, which Ida did not have, we still do not know which precincts were sampled by the exit polls. I can pretty much guarantee that Ida didn't use exit polls to pick the precincts.

The odds are against any of the 11 recounted precincts being the same precincts that were used in the exit polls.

We may be able to figure out which precincts are which, but it is going to require that NH residents who saw the exit pollers at work chirp up and tell us where they saw them.

[sacxtra]

5. BYPASS NADER, IT'l NEVER BE RECOUNTED, BAN DIGITIZED DATA!

THey will NEVER have a FULL recount.

NEVER.


No amount of money or resources will ever bring about a 100% recount.

And if electronics
or
digitized data
or
insecure networks
were used, then a FELONY was commited,
because your vote when converted to digital, traveled at the speed of light and disappeared. And no human can see it anymore. No human can audit that format in it's enviornment either.

An electronic component could have burned out, a wire broken.
You'll never know because you can not see it, and you can not be in all geographic locations simultanously (even if you could see it.)


You now need to fill out an electio/n fraud complaint Form.

Your right to vote was denied.
Because your vote was not counted.
Whoever changed it to digitized data is guilty of a FELONY.

Just like if you would have smashed the machine! And denied others the right to vote. JUst like if you tapped into the telephone wires. Just like if you tapped in on a switch, hub, or keyboard.

Tell them electronics, digitized data, and insecure networks have no PLACE in the United States Of America.

[hereinmissouri]

6. How does this answer the question? n/t

[IdaBriggs]

8. Go Here for the Story

It was my public report to DU: http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=97506

You can also read the stuff I put up at www.invisibleida.com for more details.

:) Best, Ida

[hereinmissouri]

10. Thank you. Do you know

if there have been any studies done on the NH exit polls to see if they differ from the other states with exit poll/vote discrepancies? Am I understanding, correctly, that you believe NH vote count was right? If so, do you have any theories about the exit polls?

[davidgmills]

11. According to TIA's numbers

the exit poll of NH only involved 373 individuals. That has a significant margin of error. 1000 people has an error rate of about 3% and this would be much higher.

Let the stats guys tell you how much.

Looks like the pre-election polls were very accurate though. Maybe they polled enough people.

[TruthIsAll]

12. NH: 1849 polled, MOE = 2.27% but....

There are many who are suspicious about the NH recount (method
and precincts chosen). Just look at the  Ohio recount farce.

And...I never trusted Nader. 
Not in 2000. 
Not in 2004. 
Not now.

Sorry, Mistwell. The exit polls were right. That has been
proven to everyone's satisfaction except you.  



		Poll			Kerry	2 party	Poll	Prob.	Dev/	Beyond
	St	Size	MOE	Stdev	Vote	Poll	Dev	Dev	MOE	MOE?
1	NH	1849	2.27%	1.16%	50.69%	55.50	-4.81%	0.00%	-2.12	Yes
2	NY	1452	2.47%	1.26%	59.29%	63.97	-4.68%	0.01%	-1.89	Yes
3	SC	1735	2.34%	1.20%	41.36%	45.79	-4.42%	0.01%	-1.89	Yes
4	NC	2167	2.10%	1.07%	43.76%	47.31	-3.55%	0.05%	-1.69	Yes
5	VT	685	3.56%	1.81%	60.30%	65.69	-5.38%	0.15%	-1.51	Yes

6	PA	1930	2.22%	1.13%	51.26%	54.41	-3.15%	0.27%	-1.42	Yes
7	OH	1963	2.21%	1.13%	48.94%	52.06	-3.12%	0.28%	-1.41	Yes
8	MN	2178	2.09%	1.07%	51.76%	54.61	-2.85%	0.38%	-1.36	Yes
9	FL	2846	1.84%	0.94%	47.48%	49.93	-2.45%	0.44%	-1.34	Yes
10	DE	770	3.48%	1.78%	53.83%	58.44	-4.61%	0.47%	-1.32	Yes

11	MA	889	3.10%	1.58%	62.68%	66.46	-3.79%	0.84%	-1.22	Yes
12	AL	730	3.57%	1.82%	37.10%	41.08	-3.98%	1.45%	-1.11	Yes
13	RI	809	3.30%	1.69%	60.58%	64.24	-3.66%	1.49%	-1.11	Yes
14	NJ	1520	2.49%	1.27%	53.40%	56.13	-2.73%	1.61%	-1.09	Yes
15	AK	910	3.18%	1.62%	36.77%	40.14	-3.37%	1.91%	-1.06	Yes

16	UT	798	3.18%	1.62%	26.65%	29.93	-3.28%	2.16%	-1.03	Yes
17	NE	785	3.37%	1.72%	33.15%	36.54	-3.39%	2.43%	-1.01	Yes
18	CT	872	3.27%	1.67%	55.28%	58.47	-3.20%	2.76%	-0.98	
19	NV	2116	2.13%	1.09%	48.68%	50.66	-1.98%	3.44%	-0.93	
20	AR	1402	2.61%	1.33%	44.72%	46.93	-2.21%	4.89%	-0.84	

21	VA	1431	2.59%	1.32%	45.87%	47.96	-2.09%	5.65%	-0.81	
22	MS	798	3.44%	1.75%	40.44%	43.20	-2.76%	5.77%	-0.80	
23	NM	1951	2.22%	1.13%	49.60%	51.34	-1.74%	6.17%	-0.79	
24	LA	1669	2.38%	1.22%	42.67%	44.50	-1.83%	6.67%	-0.77	
25	IL	1392	2.60%	1.33%	55.22%	57.13	-1.92%	7.44%	-0.74	

26	CO	2515	1.95%	1.00%	47.63%	49.07	-1.44%	7.45%	-0.74	
27	AZ	1859	2.27%	1.16%	45.00%	46.60	-1.60%	8.39%	-0.70	
28	ID	559	3.91%	1.99%	30.68%	33.33	-2.66%	9.14%	-0.68	
29	WA	2123	2.12%	1.08%	53.65%	55.07	-1.42%	9.43%	-0.67	
30	GA	1536	2.48%	1.26%	41.65%	43.11	-1.46%	12.33%	-0.59	

31	DC	795	1.92%	0.98%	90.52%	91.63	-1.11%	12.86%	-0.58	
32	MO	2158	2.11%	1.07%	46.38%	47.48	-1.09%	15.42%	-0.52	
33	IA	2502	1.96%	1.00%	49.66%	50.67	-1.01%	15.62%	-0.52	
34	IN	926	3.17%	1.62%	39.58%	40.97	-1.39%	19.43%	-0.44	
35	MI	2452	1.98%	1.01%	51.73%	52.55	-0.83%	20.66%	-0.42	

36	CA	1919	2.22%	1.13%	55.04%	55.73	-0.69%	27.12%	-0.31	
37	KY	1034	3.00%	1.53%	39.99%	40.76	-0.76%	30.89%	-0.25	
38	MD	1000	3.07%	1.57%	56.57%	57.04	-0.47%	38.14%	-0.15	
39	OK	1539	2.38%	1.21%	34.43%	34.73	-0.30%	40.32%	-0.13	
40	ME	1968	2.20%	1.12%	54.58%	54.83	-0.25%	41.15%	-0.11	

41	WI	2223	2.08%	1.06%	50.19%	50.21	-0.02%	49.17%	-0.01	
42	MT	640	3.78%	1.93%	39.50%	39.28	0.22%	45.51%	0.06	
43	HI	499	4.38%	2.23%	54.40%	53.32	1.08%	31.38%	0.25	
44	OR	1064	3.00%	1.53%	52.11%	51.22	0.89%	28.17%	0.29	
45	SD	1495	2.45%	1.25%	39.09%	37.42	1.67%	9.11%	0.68	

46	WY	684	3.50%	1.78%	29.69%	32.07	2.38%	9.07%	0.68	
47	ND	649	3.63%	1.85%	36.09%	33.58	2.51%	8.76%	0.69	
48	KS	654	3.65%	1.86%	37.13%	34.60	2.53%	8.68%	0.69	
49	WV	1722	2.35%	1.20%	43.52%	45.19	1.67%	8.14%	0.71	
50	TX	1671	2.31%	1.18%	38.49%	36.84	1.65%	8.08%	0.71	
51	TN	1774	2.29%	1.17%	42.81%	41.15	1.66%	7.75%	0.73	
										
		73607	2.71%	1.38%	47.09%	48.84	-1.59%	10.27%	-0.62	

[helderheid]

16. Thank you TIA

I've come to depend on you and I really appreciate your hard work

[Igel]

24. Non-response rate for NH? n/t

[davidgmills]

13. obviously I was wrong about this

See TIA's #s above

[berniew1]

14. There wasn't use of touchscreens or indications of problems in N. Hampshir

as there was in over 20 other states as reported to the EIRS election hotline. so why did Nader pick one of the few states with few known problems to recount??? There were lots of states that would have found problems, based on the EIRS reports to direct you.

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

I don't understand the N. Hampshire recount. I'll try to follow it based on what people show here.

[berniew1]

15. I"ve been long time Nader supporter, but he seems to have been supporting

Repub agenda throughout this election campaign, knowingly or not.
I think not knowingly.
This may be a continuation of that. Doing a recount where its unlikely to find a problem(optiscan counties) in state with virtually no reported irregularities(unlike most other states) and thats the one he chooses to recount and then make noise about no problem. Seems like combatting major problem by choosing token area with no problem and focusing only on it.

[davidgmills]

22. If you look at TIA's #'s , New Hampshire was the most suspect

so it would have been a good choice.

But I assume that when picking precincts, after reading Skids' entry, that they went by registration and that we had some northern Dixiecrats switching over in certain precincts making the precincts look highly suspicious. So then when they do the recount, they discover that there really are northern dixiecrats.

But of course if these Dixiecrat precints were not exit polled, as skids suggests, then they would not be proof that the exit polls were wrong.

[garybeck]

28. this is why Nader picked NH

there were electronic voting machines with paper backups that could be checked.

there was a large deviation from the exit polls

randi rhodes told a bazillion people to send him a fax to do it.

[BushSpeak]

27. Some 96,000 same day voter registrations might be a lead

I remember reading at the time, that there was surprising number of same day voter registrations in NH - something like 96,000. The article found this unusual to say the least.

This would be something to check out, if the new voter registration data is easily accessible, especially by precinct.

The first thing to check would be to see if there is a correlation between the increased voter registration and the suspect counties/precincts (found on Ida's site).

With Bush & Rove's smoke and mirrors, it's hard to find the line between mirage and reality.

[garybeck]

30. Question: Why was the recount stopped for over a week?

Where were the ballots during that time?

Do we know for 100% certainty that they were secure and not tampered with?