This is getting a bit ridiculous...

Someone please remind Mitofsky about these "little details" concerning the scientific method:

"The scientific method has four steps:

1. Observation and description of a phenomenon or group of phenomena.

2. Formulation of an hypothesis to explain the phenomena. In physics, the hypothesis often takes the form of a causal mechanism or a mathematical relation.

3. Use of the hypothesis to predict the existence of other phenomena, or to predict quantitatively the results of new observations.

4. Performance of experimental tests of the predictions by several independent experimenters and properly performed experiments.

If the experiments bear out the hypothesis it may come to be regarded as a theory or law of nature (more on the concepts of hypothesis, model, theory and law below). If the experiments do not bear out the hypothesis, it must be rejected or modified. What is key in the description of the scientific method just given is the predictive power (the ability to get more out of the theory than you put in; see Barrow, 1991) of the hypothesis or theory, as tested by experiment. It is often said in science that theories can never be proved, only disproved. There is always the possibility that a new observation or a new experiment will conflict with a long-standing theory."

Particularly number four. Exit polling is a SCIENTIFIC PROCESS he "claims."

Also a reminder for those giving credibility to Mitofsky's fairy tales.

http://teacher.nsrl.rochester.edu/phy_labs/AppendixE/AppendixE.html#Heading3

...seem to get backwards:

"If the experiments do not bear out the hypothesis, it must be rejected or modified."

The "it" in this statement refers to "the hypothesis", not the experiment. Far too often, the researcher is forced to modify the experiment until the resuts bear out the hypothesis rather than admit that the hypothesis is what needs to be altered or thrown out totally. Also, it is often the experiment which is challenged when the results fail to meant what the client seeks rather than the hypothesis. This is what drug companies, big tobacco, political pundits and major news organizations are doing all the time. They change the methodology, alter the survey questions, twist the results, until it tells the story they want told. Professional researchers I believe are fundementally honest and operate with integrity. Unfortunately, the researchers are also hired and paid by the people who commission the work, so that becomes the researchers main vulnerability.

Mitofsky's "hypothesis" fits the MEDIA WISHES. Therefore is INVALID AT BIRTH.

Thanks!

It is taught in Junior High Science class, if not Elementary school.

Sometimes the results don't match the prediction my hypothesis would make. Ha. "Sometimes"? Frequently.

I consider the experiment. I consider my hypothesis. I consider how I took my measurements.

Sometimes just one is wrong; sometimes two are wrong; and, yes, sometimes all three are wrong.

If my hypothesis presumes that people would look at any one of 4 pictures on the computer screen (I do eyetracking experiments) starting from a center point with equal likelihood, anything that violates that assumption makes the experiment invalid. If some pictures are more brightly colored, or larger, or there's too long a time between when I "force" them to look at the center point and when they hear the stimuli so that their gaze has moved ... my data's toast. Once the data said I was wrong; then analysis showed they strongly preferred the lower right hand corner, regardless of the stimuli or picture. Then I remembered the was a window with cars and kids to the right of the computer. Lower the blinds ... voila!--good data.

Or if I measure things differently (one part of the data is when their eye moves to the quadrant, another part is when they actually gaze at the object), my data's toast.

Getting experiments right is damned difficult. Making sure you're measuring what you think you're measuring is damned difficult.

Ah, yes. It certainly is.

was both paid off AND threatened.

Clearly, he went off the deep end and was squeezed from all sides.

Edited to add the right letter:

January 20, 2005
Warren Mitofsky
Mitofsky International
1776 Broadway - Suite 1708
New York, NY 10019

Larry Rosin
President
Edison Media Research
6 W. Cliff St.
Somerville, NJ 08876


Dear Mr. Mitofsky and Mr. Rosin:

I have reviewed the internal report you issued yesterday concerning your exit polling in the 2004 election, and, unfortunately, it has not caused my concerns and questions regarding the significant discrepancies between your polling data and the final electoral results to diminish.

In particular, I would note that there are a number of concerns with the explanations you posited in your internal report that do not credibly account for the unprecedented five point differential between your exit polls and the reported results. As I am sure you know, Professor Steven Freeman of the University of Pennsylvania has determined that such a differential was of a less than 1 in a 1000 likelihood - virtually impossible as a statistical matter.

To be frank, blaming such factors as distant restrictions on polling places, weather conditions, the age of exit poll workers, and the fact that multiple precincts were contained at the same polling place, as your report does, does not come close to explaining why the exit polls overstated support for the Kerry/Edwards ticket in 26 states and support for the Bush/Cheney ticket in only 4 states. Many of the factors you point to appear to merely be random characteristics of the election and your exit polling, rather than quantifiable and justifiable explanations. Nor can I believe that the massive discrepancies can credibly be written off to eagerness of Kerry voters to participate in the exit polls.

As a result, I would like to reiterate my request to receive the actual raw exit poll data that you obtained. I would also like to obtain copies of all internal deliberations, memos and other materials of your employees and consultants concerning or seeking to explain the discrepancies. To the extent you have concerns regarding releasing propriety information, I am willing to work with you to either receive this information on a confidential basis, or alternatively to bring in a neutral, outside expert to review these materials.

The stakes for our democracy are simply too high for us to allow this matter to pass without a serious and substantive review of the exit poll data. While the election is over, there is significant bipartisan sentiment in Congress and around the nation for voting reform. A complete and full release of the exit poll information will therefore not only help to resolve lingering doubts regarding irregularities in the 2004 election, it will also go a long way towards helping Congress understand how to best craft these reforms. I am hopeful that the media companies that contract for your services will also understand and support the importance of providing full, complete, and transparent information in this matter.

I would appreciate your responding to my office through my Judiciary Committee staff, Perry Apelbaum and Ted Kalo, 2142 Rayburn House Office Building (tel. 202-225-6504, fax 202-225-4423), by January 27th. Thank you.

Sincerely,

John Conyers, Jr.
Ranking Member
House Judiciary Committee

cc: Hon. F. James Sensenbrenner, Jr.
Chairman, House Judiciary Committee



http://www.freepress.org/departments/display/19/2005/1111

Conyers is a man with a clear mind, indeed!

Here is what was written in that response letter just posted "unprecedented five point differential between your exit polls and the reported results"

That claim is laughable. Not only is it not unprecedented, but EVERY FRIGGEN YEAR, EVERY SINGLE UNWEIGHTED EXIT POLL HAD AROUND THAT LEVEL OF DEVIATION FROM THE REPORTED RESULTS, AND SOMETIMES MORE THAN THAT LEVEL OF DEVIATION.

What the hell is wrong with people around here these days. Do they think that reported this lie over and over again will somehow make it the truth? Are they just so friggen blind to the truth that it becomes invisible on the screen when the data is posted over and over and over again from the prior years?

Come on guys, DEAL WITH THE TRUTH ALREADY.

I think there was fraud in this last election, but I know for sure that the exit polls are not useful for proving that fraud. The exit polls are more flawed than the actual poll, and therefore useless for proving fraud at the polls. How many years in a row do exit polls have to be off by more than the margin of error in the unweighted release of the data before you admit to yourself that maybe, just maybe, this claim that they are never off by this much isn't quite accurate?

"MOVE ON"

"GET OVER IT!"

You failed in your efforts!

Time to "cash in" now...

You obviously think I am some agenda other than my stated agenda, which is to get people to focus more on actual fraud events than on theoretical fraud events based on exit polling.

If you think I have another agenda, then lets hear it. Otherwise, my cause is as just as yours, to pursue fraud in the most effective way possible.

If you think I am going to stand by while good people continue to chase a red herring, while actual evidence of fraud slips away because too many people were following that distraction, you got another thing coming. You are hurting our cause by repeating a lie so easy to disprove that it taints the entire anti-fraud effort. I'm sure you and others have good intentions, but plenty of good intentions have hurt the thing they were trying to help throughout history. This is yet another example of good people not knowing the harm they are doing.

And those who will respond to your "leadership."

Good luck!

You haven't disproved anything!

Your words are hollow, your "facts" are non-existent. Take a clue from Conyers, and ask that all the data be released so that a full investigation can take place.

You, mistwell, seem to be jumping to conclusions, and your jumping leads to questions about your veracity, and introduces doubts concerning your agenda.

Conyers is focusing on ACTUAL FRAUD. He basically dropped the entire exit poll thing. PLEASE follow his lead!

It's easily disproven once one actually looks at the data from prior years and sees that unweighted data always overstated the Democratic vote. Once that fact came out, y'all were stuck with either claiming that ALL prior elections were stolen (which is silly, looking at the results of those elections), or that the exit poll data isn't useful for proving fraud in a close election.

Focus on the over 40,000 documented election "incidents".

We know 86 of 88 voting machines turned Kerry votes to Bush.
Focus on that.

And we all know of the need for a verifiable paper ballot on the touchscreens.
Tell us why you think the Repubs fought to keep it out of HAVA.

The mathematical circumstantial mathematics is good basis and motivation for you and others to focus on the direct evidence.

Like the toucshscreens and HAVA,

Glad to hear our cause is your cause.

make sense.

Of course the data had to be weighted

I do not understand your point

The reported results were from the weighted after being put through the model data.

exit polls are only off in Bush presidential runs - does that suggest a problem?

The stuff that indicates that Kerry won is ALL not fully weighted. They had tossed in some of the weigting items, but not the bulk of them. Only the final poll was fully weighted. And, the final poll matches the actual election count.

The exit polls are off in ALL Presidential runs, prior to being fully weighted. That is a fact.

1988 / Dukakis: 50.3% Bush: 49.7% / +0.6% / -7.7%

Of course, that was pre-weighting. And any even casual glance at the pre-weighting shows they interviewed WAY too many Democrats that year. Which is why the weighted results adjusted for the poll watcher data on the political affiliation of who actually voted.

Even with it being off, the electoral college vote also showed a Bush senior win however, so I suppose in one respect you are correct. But I thought we were talking more about the popular vote.

Exit polls is the topic.

That was directly from the pre-weighted exit poll for 1988. It mistakenly showed Dukakis as the winner, before weighting happened. That IS the F'ing topic....are you just not following?

You want me to repost it?

Taking your pills lately?

THE ONLY SOURCE YOU QUOTE ALL THE TIME IS THE "WASHINGTON MONTHLY" GARBAGE.

Game over!

Except that is plainly not true.

1) The "washington monthly crap" isn;t even the source. They quote the PDF of a foundation that did the research, and I politely do not cite directly to the PDf because some folks don't want to click on a link that opens their Adobe.

2) I quote the Duke study (whic IS a PDF so please be warned) that explains why the exit polls tend to

http://www.duke.edu/~mms16/non_response2000.pdf

3) And I quote this one:

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

4) I quotes lots of stuff on the Ukraine exit poll, showing how their exit poll for the second election, the most carefully watched election in the history of the F'ing planet, also was off. In fact, all three of them were off, by more than the margin of error, dispelling the myth that other countries use the exit poll as a good indicator of fraud.

http://www.myrtlebeachonline.com/mld/myrtlebeachonline/10499250.htm

and Official results: http://tinyurl.com/3ufxp

So seriously, you need to chill out. If you think the data from the prior US exit polls is false, prove it. It's in regular newspapers from those years.

If all prior pre-weighted exit Presidential exit polls in the US were flawed to deviate in favor of Democrats, and the studies explain why that is, and the creator of exit polls says they tend to do that (which he does, see http://mayflowerhill.blogspot.com/2004/11/mayflower-hill-exclusive-warren.html ), then damnit why the hell can you not see that perhaps this exit poll nonsense is a red herring, and we should be focusing our efforts on evidence of actual fraud rather than this horse shit ment to prop up TIA's ego!

Arguments about the aggregate results aside, even the tiny squirt of new data coming from Mitofski's report has the potential to help home in on some of the smaller pockets of fraud/error. More might even go so far as to pinpoint a few.

We now know that a few precincts had to be thrown out due to huge discrepancies. We don't know which ones yet, but surely they deserve an investigation to see whether there is a non-fraud/error explanation or not.

We now know which groups of precincts, by five categories of final result, had more outliers.

To say that NEP results cannot be used as a fraud barometer in the aggregate is one thing. A debatable thing. But to say that getting the raw data would not be useful in the investigation to locating and prosecuting what fraud did occur is to advocate that we voluntarily ignore useful data on your say-so. That is not going to happen.

Conyer's assessment that Mitofski's offered explanation is shaky (because it is, just look at the nonresponse vs. unsigned WPE. The neat correlation there is not very compatible with the double-peak one would expect to see if there was a differential rate between two large and evenly split groups like Bush and Kerry voters) and especially his call for more internal poll data, are both right on target, whether or not you choose to believe that the magnitude of the fraud/error was a signifigant contributing factor to the aggregate results.

Not to mention, refusing such a request from a ranking senator is pretty disrespectful, and I might add, unpatriotic.

And as far as the differential, it is unprecedented in that it came from the WPE. The last time we had that level of error, it wasn't from the WPE, it was from methodological bias in other parts of the survey, and that mistake has since not been repeated.

I am ALL in favor of getting the raw data. I agree, it can be useful.

My only contention is that the current exit polls we have, the ones that are partially weighted and the final one that is fully weighted, are not themselves useful for fraud detection because both have empirically been shown to be flawed (either by always weighting to heavy Democratic in the pre-fully-weighted version, or being tailored to the actual vote in the fully-weighted version).

Conyer's is not a senator, by the way. Boxer has not made the request, as far as I know. I wish she would.

As for the difference between WPE and methodological bias, I would ask that you define WPE. I don't know what it is, and I have been following most exit poll threads. So, it would be polite to actually define the term, so I know if there is an actual difference or not.

"The exit polls are more flawed than the actual poll."

Please prove this statement.

You can't. You won't. You never attempt.

At least with even "flawed" polls we have some independent proof of their accuracy.

We have no such thing with the actual count.

Why don't you get on my band wagon and support a re-exit poll, one that by your own design is not "flawed" and will prove or disprove this hypothesis of yours. I have been suggesting this for over a month now. But it usually falls on deaf ears.

Pick your state or the country at large. I'm game.

"The exit polls are more flawed than the actual poll."

Please prove this statement.

You can't. You won't. You never attempt."

I have...over and over and over again.

Here, I will do it again. Perhaps some day, you will actually look at the links and consider it:

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%

That empirical data shows that the exit polls are ALWAYS very flawed, usually well beyond the margin of error. So either A) thoursands of Republicans secretly got together in a huge conspiracy to always commit massive election fraud every Presidential election year, even prior to electronic voting, and even though they lost the popular vote anyway 3 out of 4 times and would have won the 4th time even without fraud, all at great personal criminal risk for no gain at all, or B) The exit polls always skew Democratic.

Then see this study from Duke on how exit polls always skew Democratic, and why:

http://www.duke.edu/~mms16/non_response2000.pdf

You can see that the problem persists internationally as well with the SECOND Ukraine election, which was one of the most watched elections in the history of mankind:

http://www.myrtlebeachonline.com/mld/myrtlebeachonline/10499250.htm

Exit Poll 1: 58.1% to 38.4%, MOE=2
Exit Poll 2: 56.5% to 41.3%, no MOE given
Exit Poll 3: 56% to 41%, MOE=2

Official results: http://tinyurl.com/3ufxp

51.99% to 44.2%

As can be seen, the final results varied from all the exit polls by much more than the margin of error, showing once again that exit polls are just not all that relaiable for a close race, though useful for races that are not close.

then we have the father of exit polls, Mitofsky, a lifelong liberal who "apparently holds no brief for Bush", who did the polls himself, who says he does not believe the early raw exit poll data (the stuff showing a Kerry win) indicates fraud. Indeed, he thinks the early raw data is inaccurate, and not useful until it is fully weighted with his formula.

http://mayflowerhill.blogspot.com/2004/11/mayflower-hill-exclusive-warren.html

And after all that, if you still conclude pre-fully-weighted exit polls are a reliable way to call close elections, then I guess your head is permanantly buried in the sand.

I am all for any idea that calls for a better exit poll system for future elections. I am all for even having it run by an independant government board that pays to have a LOT more people interviewed, and publishes all data including the raw data and the full weighting formula so people can run their own tests on it. Count me 100% in. But as it is done now, the exit polls are simply not useful for calling a close race. That isn't their design, that isn't what the margin of error is based on, and it isn't useful in achieving the goal of detecting election fraud...nor has it ever been, the way we have done them in the US.

Differences between adjusted and unadjusted numbers will be small at the state level -- less than 1% to perhaps 4%. For a summary from the 1990 Census, see http://cmbp.gov/v3/index.shtml. Non-Hispanic whites were overcounted in several states by up to 1% in 1990. Blacks were undercounted by up to 8%, American Indians, up to 10%. In addition, children of all races were undercounted in most states by up to 6%. For smaller areas, the differences will be larger, especially in poor neighborhoods, both urban and rural. For poor neighborhoods, the undercount can range to 10%. Overcounts, mostly in wealthy suburbs, range up to 2%.

Differences between adjusted and unadjusted numbers will be small at the state level -- less than 1% to perhaps 4%. For a summary from the 1990 Census, see http://cmbp.gov/v3/index.shtml. Non-Hispanic whites were overcounted in several states by up to 1% in 1990. Blacks were undercounted by up to 8%, American Indians, up to 10%. In addition, children of all races were undercounted in most states by up to 6%. For smaller areas, the differences will be larger, especially in poor neighborhoods, both urban and rural. For poor neighborhoods, the undercount can range to 10%. Overcounts, mostly in wealthy suburbs, range up to 2%.


http://cronkite.pp.asu.edu/census/Compare1990_2000.doc


If this is how we count the population, I got real doubts about how we count the votes.

Take a good note of who is undercounted and overcounted.

why there was a deviation from the final election results.

But they do point to non-response rates. If non-response rates are sufficient explanation for somebody, then the deviation's largely explained. (Which reminds me, I have a pdf to read.)

I have yet to hear anybody posit a reason that interviewer experience or weather should correlate with fraud.

the voter is telling the truth, and since only about 1% or less of
the voters are actually polled, how does the pollster know if an
accurate cross section was obtained?

In the end, polling is for selling newspapers and TV commercials.
It is entertainment. If polling was accurate and completely reliable
then we can save Billions each election cycle on printing ballots,
setting up polling stations, staffing them with people, paying guards
to handle the ballots, voting machines (and those electronic machines
with paper printing devices are expensive as hell), personnel to
count the votes, personnel to oversee the vote counting, cost of
lawyers, judges, court personnel to handle lawsuits arising from
legal contesting of election, cost of recounts (sometimes TWO),
etc etc etc.

The ONLY practical use of exit polling is to verify that a LARGE
discrepancy (10% or more) did not occur with actual vote counts.
There is some validity to statistical extrapolations based on a 1%
or less sample. But since no one can read minds of people answering
pollsters, it is impossible to know how accurate the raw data is.

check on ballot fraud that can not be found on recount (goggle and see the discussion of the type of ballot fraud that occurred in 1960)

...voters be more likely to say they voted for Kerry then to tell the truth and just come out and admit they voted for Bush? I think just the opposite, they would tend to say the voted for Bush or they would just refuse to answer. Refusals are tracked in exit polls and I think that the level of refusals were not abnormally high, nor were they sufficient to distort results by the levels that would so over state the poll results compared to actual voting results.

The exit polls accomplished exactly what they were designed to, they raise red flags that something significant is not right with the actual vote counts in a large number of states and precincts and therefore the whole voting process needs to be thoroughly examined for possible fraud.

He uses the four step fascist approach:


1. Ignore

2. Pretend

3. Cover Up

4. Lie

I'll bet he has plenty to say about it.

state exit polling. Zogby sounds honorable. Maybe we should hire him for the 2006 election.

trudyco

around $10M (see http://www.washingtonpost.com/ac2/wp-dyn?pagename=article&node=&contentId=A51692-2003Jan13¬Found=true about halfway down the page)

A German-style exit poll that would be suitable to detecting fraud would cost about $80 to $160 million - judging by the methodology used and comparing the sizes of Germany and US. (see http://www.mysterypollster.com/main/2004/12/what_about_thos.html for the description of differences between the two types of exit polls) and would still not be as accurate, due to the US high non-response rates and the fact that German Census bureau does official demographic surveys inside polling places and nothing like that exists in the US.

YOU "CONVINCED ME"...

I'm going to bed now.

I was merely suggesting that we ask Zogby his opinion on this.

After all, his firm suggested at the first Conyers hearing that a blue ribbon panel be assembled to study the exit polls.

http://www.wpherald.com/North_America/storyview.php?StoryID=20050120-011015-8076r

<snip>

But an independent pollster said yesterday that the firms still had not told the full story behind the error-filled polls and questioned whether exit polling should be used to project the final outcome of an election.

"I'm not sure we're at full disclosure yet. An awful lot still needs to be explained," said pollster John Zogby, who expressed deep skepticism about the report's explanations and excuses. "The sum total of what we got today is enough to suggest that there should never be exit polls again."

Among the explanations offered by Edison Media Research and Mitofsky International, the two polling firms that were hired by a consortium of news organizations: Mr. Kerry's supporters were more likely than Mr. Bush's voters to agree to participate in exit poll interviews.

Polling errors were found to be made more frequently by younger interviewers, about half of whom were 34 or younger.

Early preliminary polling data collected during Election Day that was not supposed to be made public was leaked to blogger Web sites and suggested a wider Kerry lead than he actually had. Interviewers in many precincts were kept 50 feet or more away from voting places, potentially skewing the results toward Kerry supporters who were eager to participate in exit polling interviews.

But pollsters like Mr. Zogby said they were not buying the polling firms' explanations.

"A lot of money and a lot of time is put into this. Ultimately, to blame all this on logistics and on the youth of the interviewers is enough to say, 'Can we ever believe an exit poll again?' " Mr. Zogby said.

"I think we need a healthy debate and discussion about this, and I, for one, do not think an exit poll should ever be used to make a projection because there can be imprecision, there can be leading of the public and we don't need the exit polls to do that," he added. "I'm always going to be skeptical when I see an exit poll. There are just some numbers that don't jive with reality."

Brian Jones, the Republican National Committee's communications director, also was skeptical about exit polls and their dependability.

"Clearly, there were methodological flaws with the exit polls, but that was something we were able to detect in our operation. We noticed after a few reports that some of the information did not match up with reality," he said.

<snip>

"The information (vote count) DID NOT MATCH REALITY."

Kerry is the legitimate President!

in Nov. Bush was at 44% so 1% FOr 3rd party votes--Kerry gets 55%.

112 million votes cast IIRC--- Now if TIA looks at the states-- figures the Kerry win---I 'll bet you a cup coffee that it jives with my prediction back last spring----

What ya say TIA---follow thru---what was the size of the Kerry wIn?