Surprise, Surprise!!! Exits polls once again "biased" for Democrats!

[mhatrw]

In other words, the voting machines were biased against Democrats.

http://www.usnews.com/usnews/opinion/baroneblog/archives/061108/thoughts_in_the.htm

The exit poll.

The Edison/Mitofsky Research exit poll proved somewhat misleading, as past exit polls have. The Fox News decision desk personnel decided to abandon the exit poll entirely as a guide to calling winners on the grounds that it overstated Democratic percentages by 6 to 8 percent. They made this decision based on information from EMR that the actual tallies in the precincts tested showed Democrats winning 6 to 8 percent fewer votes than the exit pollsters from those precincts reported. Exit polls from 1992 on have consistently overstated Democratic percentages, most notably in the high turnout elections of 1992 and 2004.

The late Warren Mitofsky, who created the first exit poll for CBS News in 1968, went back and examined the 2004 results and found that the biggest discrepancies between actual precinct votes and the exit pollsters' results occurred in precincts where the exit poll personnel were female graduate students. All those discrepancies overstated the Democratic vote. Joe Lenski, the current EMR boss, tried to hire fewer young women as exit poll interviewers. (Mitofsky died suddenly in December; he had been a friend since 1974, and I had lunch with him in Mexico City last June 30, two days before the Mexican election, in which his Televisa Mitofsky exit poll produced results that were spot on.) The fault may be with the interviewers (do they approach only the voters they think simpatico), but it may also be with the respondents. Mitofsky has told me that almost everyone approached to fill out an exit poll questionnaire in countries like Mexico and Russia does so, while about half the people approached in the United States refuse.

Perhaps Republicans are more likely than Democrats to refuse, especially when the interviewer is a young woman whose appearance signals she is some kind of Bush hater. (Of course, this must be it! Republicans only like young men!) In any case, Lenski did not succeed in solving the problem. It's eminently solvable, but only at enormous cost: If EMR brought together all its interviewers on airline flights for centralized training, if it paid much higher fees for their work, etc. But would EMR's clients–ABC, CBS, CNN, Fox News, NBC–pay something like five to 10 times the amount for their services (which is what it would require to take these steps)? Doubtful. So we must recognize that the exit poll is an imperfect instrument and use it for such insights as it can reasonably be expected to deliver and ignore it otherwise. That is what the Fox News decision desk decided to do on election night.

By the way, exit poll information was sequestered tightly until 5 p.m. ET. None of us knew any numbers until then. Fox News had two people sent to a central location in New York, who were allowed to have no access to cellphones, land lines, or E-mail. This proved to be a much better way to handle information with an inherently misleadingly wide range of error margin than the procedures used in the past, when large numbers of people from the various news organizations got the first tranche of numbers a little after noon ET. I remember seeing the first tranche of exit poll numbers for the various seriously contested states in 2000, a little after noon. They were all very close. I had a two-word reaction, of which I will share with you only the first word, which was "Oh."

[NMMNG]

1. LOL

Perhaps Republicans are more likely than Democrats to refuse, especially when the interviewer is a young woman whose appearance signals she is some kind of Bush hater.


It must be the lack of a pinched face and "holier than thou" look that gives people the impression that the female interviewer is "some kind of a Bush hater". :rofl:

[BootinUp]

2. If half the people approached refuse

per the piece, then what is surprising about the fact the interviewed population does not represent the voters very well?

These exit polls are shit.

[mhatrw]

4. And unauditable fraud-made-easy voting machines are the tits?

Sorry, but I'd rather put my faith in exit polls. There is no hard evidence that respondents self-select in a partisan manner. It's just a bunch of people grasping at straws to explain why polls that were always 100% accurate before have suddenly begun to "favor Dems" with the advent of unauditable fraud-made-easy voting machines.

[BootinUp]

6. Actually you are wrong

There have been studies that show it. There was an article floating around here about one such study. I don't have time to look it up today. Here is a short piece in wikipedia that explains what a non-response bias is and how it can affect results of polls. Your claim that exit polls were always 100% accurate strikes me as humorous and not in need of strenuous rebuttal.

Your problem with electronic voting machines is very understandable, but it does not improve the efficacy of exit polls.

http://en.wikipedia.org/wiki/Opinion_poll#Nonresponse_bias

[johnaries]

8. That article is not about Exit Polls.

It's about telephone surveys.

Also, after the "low Bush responders" theory regarding the 2004 Exit Polls it was shown that response was actually better in predominately Republican precincts than predominately Democratic precincts.

So, if there is a "non-response bias" in Exit Polls it would favor republicans, not Democrats.

[BootinUp]

11. This is pointless

Keep your opinion without further assault from me. :hi:

[mhatrw]

12. It is only pointless because you don't seem to have any evidence

for your point.

[BootinUp]

14. I don't feel a need to prove anything to

you. Doing so won't do anything for me. I have provided enough for you to look at or think about and delve further into, if you so desire.

[ThomCat]

18. Non-response bias has never been shown

to be significant in exit polls. Non-respondents refuse for reasons Other than party affiliation, so they tend to be randomly distributed among republicans and democrats. They therefore don't bias the results.

If you have any legitimate peer-reviewed research that documents bias in exit polls then I'd love to see it.

[BootinUp]

30. you first. eom

[mhatrw]

37. What is this supposed to mean? You are claiming that exit polls that

have always been historically accurate are now not accurate due to some new, imaginary non-response bias that you assume to exist. Where is your evidence that this bias actually exists in any statistically significant quantity?

Remember that more voters responded in highly Republican precincts than highly Democratic precincts in the 2004 election. This supports the contention that any putative non-response bias should underestimate Democratic votes. So tell us, what conflicting evidence can you provide?

[BootinUp]

41. The poster I responded to

made some strong assertions not backed up with any "peer reviewed studies" but asked me to provide one for my position. I asked him to "go first" with his peer reviewed sources.

[w4rma]

31. Here is a study. Explain these results to me, please.

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=364&topic_id=2590865

[BootinUp]

32. I take neither your position nor the articles position

I simply question exit polls as an accurate means of auditing elections. Accurate is a loose term mind you, I would accept exit polls to characterize data with the proper qualifiers. I certainly would not accept them as 100% accurate as one of the previous posters spewed out.

I also wonder why anyone would think people refusing to take exit polls is a different phenomena than people refusing to take phone polls or any poll for that matter. I would be amazed to hear a logical explanation for that position.


edit: one word missing in 1st para, no change to meaning.

[w4rma]

33. Are you saying that we should do away with phone polls?

Or are you saying that there is a "Democratic bias" in phone polls?

Or are you saying that there isn't any fraud going on and everything is on the up and up?

I really don't understand, at all, the points you want to make on these types of threads. You say you hold no opinion, and yet you are always on these types of threads aggressively voicing some opinion. I really do not understand your point.

[BootinUp]

36. in answer to your question please re-read

my previous post.

[mhatrw]

35. From your own link.

http://en.wikipedia.org/wiki/Opinion_poll#Nonresponse_bias

In terms of election polls, studies suggest that (nonresponse) bias effects are small.

Do you have any evidence at all that Republicans are significantly (1% or greater) more non-responsive than Democrats when it comes to exit polls? Any at all?

[BootinUp]

38. I'll try to find the article

[BootinUp]

39. I believe this is the article I read before

I leave it to you to belittle its premise or claim that it does not apply to this discussion. But I am through debating it. I do not claim to have proved anything either. Hope you find it interesting.

http://people-press.org/reports/display.php3?ReportID=89

[BootinUp]

43. This aint the same article

I can't find the original article I read, but there are other sources that discuss non response to polls that may be of interest to this discussion. Here is another article abstract only:

PRE-ELECTION POLITICAL POLLING AND THE NON-RESPONSE BIAS ISSUE
Ronald G. Shaiko, Diana Dwyre, Mark O'Gorman, Jeffrey M. Stonecash and James Vike

Abstract

Non-response bias is of great concern in pre-election political telephone polls, where the number of callbacks is far less than in most academic and government studies. The fewer number of callbacks raises the issue of whether there might be a greater non-response bias problem in such surveys. Given the public importance of political polls, there is a need to know something about this. The general problem faced in trying to assess this issue is that it is very difficult, if not impossible, to obtain information on non-respondents. There is an additional problem with political polling in that there are few public studies about non-response in such polling. This study assesses non-response bias for five pre-election political polls which used four callbacks each. The study is unique in that information is available for all respondents and non-respondents. The initial list of individuals to be called was taken from a computerized voter registration list which contains the age and sex of each registrant. The non-response bias associated with not-at-homes, bad numbers, and refusals is assessed. The results indicate that bad numbers and not-at-homes have different bias effects from refusals, but they offset each other to some degree in a situation of limited callbacks. More callbacks may reduce bias associated with not-at-homes, but not the bias associated with bad numbers and refusals. We usually do not know how biased nonresponse is, but it is seldom a good assumption that nonresponse is unbiased (Flower, 1984, p. 51–2


http://ijpor.oxfordjournals.org/cgi/content/abstract/3/1/86

[ThomCat]

17. Exit polls have proven very accurate

time and time again. Until very recently it was very uncommon for an exit poll to vary from the actual results by more than 1%.

Even if some people refuse to participate, those people tend to be equally distributed among voters. So the remaining sample is large enough and random enough to generate very reliable results. Nobody has ever presented any research documenting a bias.

It is more likely that people really are voting for Dems more often than show up in official numbers. Irregularities in voting machines and in vote "spoiled" ballots are the more likely reason for the discrepency.

[Cessna Invesco Palin]

3. Heh.

I can imagine anybody voting for a Republican this time around not wanting to admit to it. :evilgrin:

[mhatrw]

5. It's probably just my imagination, but I can imagine that you don't

need much imagination to imagine this.

[Quantess]

10. I had the same thought.

Maybe some people didn't want to admit voting for an R.

[mhatrw]

13. In 1996, 1998, 2000, 2002 and 2004 as well?

Does that make any sense? Is there any actual evidence to back this assertion?

[electropop]

19. I certainly wouldn't admit being a puke, to an attractive young woman.

:evilgrin:

[readmoreoften]

7. So we won by 6-8% more than we even knew.

It makes sense. The country is 56%-45% dem in conservative areas, 70-30% dem in liberal areas.

We own this country. Fuck their fraud.

[helderheid]

9. Recommended this thread - we MUST take care of it before 2008!

[No DUplicitous DUpe]

16. That's my conclusion, as well..

And that means Kerry easily won Ohio, and Florida too.

Let's hold a hearing on that subject, new Congress!

[ThomCat]

20. I've been saying for a year that we needed to win with

at least 55% just to get by with a bare majority. I guess I was pretty accurate. It's not hard to believe that 6-8% of our votes have been taken, hidden, or lost given all the election fraud that we keep seeing.

[mirandapriestly]

22. Exactly! ExitPolls were ALWAYS accurate before bush

They must have set up to steal it and underestimated the margins that democrats would win by or something. OR, it would have been too obvious..Then there is there is the purging and people not being able to register, etc, those don't even show up on the exit polls...

[ComerPerro]

15. that's pathetic

[Xenotime]

21. Exit polls were accurate this time.

[mhatrw]

24. Not according to Fox News. Read the original post. n/t

[Xenotime]

28. Sorry...But everything I heard...we had it around 11am CST

Even Drudge was posting articles leaning that direction.

[mhatrw]

34. So why did it take until last night for MT and VA to be called?

And what about the 10-12 close House races that Repukes squeaked out? Did you see the exit polls on those as well?

[Sancho]

23. We gone around this many times....

...chasing down the gender effects of the pollsters is fun, but there is likely a relationship between the availability of female graduate students and the type of district...etc., etc. As we all know, correlation doesn't imply causality and ex post facto examination of data is also suspect.

There is NO way in this situation to determine if the variability is due to the exit poll being "wrong" or a subtle manipulation of the vote. There certainly are hints that weird undervotes and patterns occur in selected precincts and races.

SO....you don't need to pay to fly in people from all over...but you certainly need to go into the 2008 election with a planned experiment on some targeted districts and races that will determine if those expected "problem" areas have accurate poll data or not...

gee....what kind of rocket science does this take?

Anyone who didn't EXPECT the west coast of Florida to have issues after 2000 and 2004 and multiple claims of discrepencies would be pretty slack. I'd put some really, really accurate pollsters and plan carefully in targeted areas. You only have to prove or disprove "fraud vs poll error" in a couple precincts to have an idea what's happening in the district or state.

Why don't the absolute best and most experienced researchers observe and poll selected critical races while they are feeding stuff to the TV stations? You only have to catch the crook once, and IF there is a vote manipulation, exit polls could be excellent empirical evidence that something is wrong!

[mhatrw]

25. How about simply showing this supposed discrepancy in precincts that still

count votes manually?

Nobody has presented any hard scientific evidence that any exit polls show a true bias toward Democrats. The pollsters are simply grasping for any anecdotal explanation because they assume that the final vote count is far more accurate than their exit polling results.

[Sancho]

27. I think that assumption is the issue...a true researcher would

seek the source of the differences - not just assume one of two variables was true. That's part of the frustration of seeing the discrepencies in the polls. There likely has always been manipulation of all sorts on a small local level, but the possibility of hacking a close vote is real, and exit polls are a clue to large undervotes, flipping, or tabulator errors. Also, the demographic predictors of those voting can be tallied even though all don't answer an exit poll; more evidence if there is a pattern that doesn't make sense...

you get the idea...

[kenny blankenship]

26. Kinda like how SCIENCE is biased for EVOLUTION

see, once you surrender to facts and logic, that's it-the game's all over.
Your FAITH has got to be stronger than facts and logic combined.

[understandinglife]

29. Excellent.

:thumbsup:


Never, Ever Forget: George W. Bush willfully violated National Security to cover-up his willful launch of a war of aggression and illegal occupation of Iraq .... and, now he willfully provided nuke-making instructions to terrorists -- if you doubt it, just check 'the google' ....

[mvd]

40. Oh yeah, we still need reform

Our victory doesn't change that.

[leftofthedial]

42. it's because they don't interview Diebold

[mhatrw]

44. kick

Now is NOT the time to back off on fair, reliable and fully auditable voting systems.