Tomorrow, MP heads off on his annual family vacation. This will be MP’s first real vacation since starting this humble blog almost a year ago, and oh my, do we need it!
Thanks to all who continue to check in to this space regularly and email with thoughts and comments. I’ll be back on August 29…or perhaps earlier with a photo or two from the Mystery Vacation..
Mark Blumenthal is the principal at MysteryPollster, LLC. With decades of experience in polling using traditional and innovative online methods, he is uniquely positioned to advise survey researchers, progressive organizations and candidates and the public at-large on how to adapt to polling’s ongoing reinvention. He was previously head of election polling at SurveyMonkey, senior polling editor for The Huffington Post, co-founder of Pollster.com and a long-time campaign consultant who conducted and analyzed political polls and focus groups for Democratic party candidates.
14 thoughts on “The Mystery Vacation”
Mark,
Enjoy your vacation.
Hmm it is tempting to use this post as a type of open thread with your readers. Any objections?
I’ve been wanting to put a couple of questions out there on past topics, but the topics had not come back up.
I’ll wait a bit before posting.
So I basically had one question:
Feeble stated that nonrespondent bias (conservative’s not participating in exit polls) has been accepted and factored into English exit polls. Can anyone point to the literature on this UK procedure?
Exit poll bias causes so much convoluted arguing in America that we could probably benefit from the knowledge England has developed on the subject.
Alex, you can check out a Jurgen Hofricther’s chapter, “Exit Polls and Election Campaigns” in the Handbook of Political Marketing, ed. Bruce I Newman, 1999. He discusses the “Shy Tory” effect documented in the 1992 election.
The exit polls actually turned out a little better than the pre-election polling. Hofrichter cites three works by Worcester in 1997 and 1998 that discusses this issue in greater detail. I didn’t bother to dig up those citations.
You can also read about the exuberant liberal voter effect turning up in Ontario Canada exit polls in 2003. Google “You’re Leaving? Great! Exit Polling in the 2003 Ontario General Election.” Their poll significantly overstated the Liberal vote and cite differential response as a possible explanation. “Either voters misreported their vote to our interviewers or Liberal voters were simply more willing to reveal their preference to us.” Pg. 9 They admitted that evidence of this theory was “hard to come by.”
BTW – Differential nonresponse has been the suspect cause of most exit poll bias for over 20 years. Plissner and Mitofsky (1982, p. 15) first raised the possibility of nonresponse bias explaining that there “is no guarantee that voters who respond to the poll are like those who refuse to answer.” In reference to the 1984 presidential election exit polls, Mitofsky and Waksberg (1985, p. 5) noted that one of the problems with projections based on exit polls that did not exist when projections were based on vote returns is the bias introduced by nonresponse, although no specific evidence of bias was presented by the authors.
Mitofsky (1991 p. 96) attributed the bias observed in the 1988 presidential election exit poll to “people who refuse to be interviewed.” Partisan differential response was fingered for problems with the 1992 presidential election exit polls as well (Mitofsky and Edelman 1995, p. 92).
Warren Mitofsky is quoted as saying, “The only thing it can be (consistent Democratic bias) is a difference in response rates of supporters of one candidate versus the other. There’s no possibility for it to be anything else and be this systematic” (The Roper Center for Public Opinion Research 1993).
The Research Triangle Institute (RTI) report on the 2000 election network coverage problems noted that exit poll nonresponse had become a “serious problem” (Konner, Risser, and Wattenberg 2001, p. 21) and Edison Media Research and Mitofsky International (2005, p. 3) cited differential nonresponse as the likely source of the exit poll discrepancies of election 2004.
Back in 1969, Warren Mitofsky and Murray Edelman found that deviation from the sampling interval was the greatest contributor to the exit poll error. That is, as interviewers did not stick to their assigned interviewing rate, they tended to “self select” or avoid certain types of voters, introducing bias. This may not be a problem if interviewer characteristics were randomly distributed (and therefore so would be their natural biases), but they clearly are not.
References:
Edison Media Research and Mitofsky International. 2005. “Evaluation of Edison/Mitofsky Election System 2004” Retrieved from http://www.exit-poll.net/election-night/EvaluationJan192005.pdf. Last accessed August 1, 2005.
Konner, J., Risser, J., and Wattenberg, B. 2001. “Television’s Performance on Election Night 2000: A Report for CNN.” Retrieved from http://www.cnn.com/2001/ALLPOLITICS/ stories/02/02/cnn.report/cnn.pdf
Last accessed, August 1, 2005.
Mitofsky, Warren J. 1991. “A Short History of Exit Polls.” In “Polling and Presidential Election Coverage.” Eds. Paul J. Lavrakas and Jack K. Holley. Sage Publications: Newbury Park.
Mitofsky, Warren J. and Murray Edelman. 1995. “A Review of the 1992 VRS Exit Polls.” In “Presidential Polls and the News Media” Eds. Paul J. Lavrakas, Michael W. Traugott, and Peter V. Miller. Westview Press; Boulder, CO.
Mitofsky, Warren J. and Joseph Waksberg. 1985. “Election Estimation from Incomplete Sample Data.” Paper presented at the Public Opinion Surveys and Electoral Forecasting conference, Sevilla, Spain, September 25-27.
Plissner, Martin and Warren Mitofsky. 1982. “Voting Twice on Election Day.” Public Opinion, August/September.
The Roper Center for Public Opinion Research. 1993. “The Polls and the 1992 Elections: A Public Perspective Symposium; Problems in Exit Polling; Interviews with Warren J. Mitofsky and John Brennan.” The Public Perspective. January, 1993/February, 1993. Vol. 4, No. 2; pg. 19
Rick, thanks for the literature review. I will try to dig up the Worcester works.
But, Rick, can you tell how the debate in England proceeded when they decided to institutionalize, rather weight by, respondent bias in exit poll calculations? I’m thinking more in a policy way than strictly scholarly way.
Also, there is an important reason why the British data and scholarly work is so important. At the time Feeble made the important point that the British felt confident in looking for errors within the exit polling, because the vote counts were so trusted, i.e. hand counts of paper ballots conducted at the precinct by two people from opposing parties, if I’m not mistaken.
If you take the British trust in the vote count and the decision to include non-respondent bias in exit poll calculations, British academics must have reached some strong conclusions. Much stronger than anything we should expect in America becuase of our laughable vote count processes, if not laughable vote counts themselves.
With all due respect to Mitofsky, et al, any analysis of American exit polling is clouded by uncertainty in specific vote results, IMO.
So I hope the Worcester papers shed some light. If anyone can provide anymore data on British exit polling specifically, please post.
Thank you!
Alex in Los Angeles
Alex, great questions about the British poll process. Unfortunately, I haven’t had time to look into it, but I did understand that they were weighting for the affect. However, I’m not sure that they did for the recent election. Quite unsure actually.
True, febble says the poll results were immediately cast into question because their elections are transparent. I don’t know much about their process. However, the poll results have been thrown into question in the US since the first projection was made back in 1982, and I’m not sure it had anything to do with the transparency of the process here.
Something also to consider about the British exits. They only ask one question. Very simple. All you do is check a box and drop your slip. Our forms are two-pages long! Consider that Democrats are more likely to be women, college students, unemployed than Republicans, they could have a bit more time on their hands than the average Republican. Again, speaking in terms of averages.
Another potentially interesting issue. The vote count totals used in the exit poll analysis may be “polluted” with absentee/early votes. As I understand it, some counties count these votes separately, others lump them in the home precincts to be counted with all the election day ballots. When the E-M worker obtains the final vote count result from the precinct, they send it in. They don’t know if their totals are only for election day voters or are mixed. Who knows if this is a significant problem or not? But, given that Republicans are much more likely to vote early or absentee, if the problem was widespread, it could “look” like ballot stuffing because the exit poll is not sampling the same population it is intended to represent.
The biggest clue is the fact that bias increased in 2004 as sampling rate increased. That is, as interviewers have more freedom to deviate from the scientific sample, the bias increases, suggesting that interviewers gravitated toward more willing and/or more “like” voters. This is the same thing Mitofsky and Edelman found in 1969. However, when febble’s alpha was used instead of Mosteller’s measure 5 (WPE), the relationship between sampling interval and bias was striking. The bias steadily increased from almost non-existent when every voter was supposed to be approached, to off the charts when every 10th voter was to be approached (Mitofsky presented this chart in Miami). That’s pretty solid evidence of interviewer induced differential non-response.
But, bottom line – E-M never said that their wasn’t fraud or that the vote count was accurate. Only that the exit poll data is more consistent with a differential non-response hypothesis than a fraud hypothesis. That is a very sound conclusion given the available data.
Hi Rick,
Excellent points on the absentee reporting uncertainty and simplicity of British elections. Exactly what I meant by laughable American vote count processes. I probably should have said any analysis of American exit polling is clouded by the complexity, not necessarily uncertainty, of our vote results.
I personally don’t doubt the existence of non-respondent bias, but I would quibble with you slightly in that I think the non-transparency of American elections, and their complexity, absolutely delayed, and still delays, a clear and sufficient understanding of nonrespondent bias in American exit polls. If British exit pollsters feel comfortable incorporating the bias into their exit polls, assuming Feeble was correct that they do, I can only imagine they have reached that level of understanding of exit poll bias within the British context. Hence my interest.
And there’s always the German context! I don’t know, but maybe, they don’t even need to worry about nonrespondent bias. I’m willing to admit this could be for reasons unique to their election/exit polling processes and context.
Just curious really on what we can learn. And I still hope to get into the Worcester stuff and anything else anyone can recommend on the apparent British resolution of their exit poll bias.
Oops, I think my post does not acknowledge the great progress made this year in uncovering and finally starting to explain to the public the causes of exit poll/vote result bias. My quibble was with the historical slowness, since 1982 according to your note on Mitofsky work, in coming to grips finally with exit polling bias.
Also, I think more progress has been made than I was aware of. I didn’t realize that Feeble’s alpha was applied to the sampling interval.
This is great as it may finally spur investment in shoring up interviewer training and pairing of interviewers. Frankly, it was the weakness of the exit polling interview methodology that gave the non-respondent bias theory immediate apparent validity in my book.
Was Feeble’s alpha applied to other variable’s with notable results? I guess Mitofsky would be smart to redo his whole analysis using the FA, or something better.
Thanks for the response, Rick.
Alex, my research has, quite frankly, been a major pain in the ass. Assembling my list of sources took months and I’ve probably annoyed many in the process. I’m pretty darn frustrated at how little research has been conducted on exit polling and exit poll error. The list of references, which is comprised largely of conference papers, that deal specifically with analysis of presidential election exit poll data is about a page.
Your points above are well heeded. I don’t know much about the German polls, other than that they were so accurate in 1978, there was an outcry because people thought it depressed voter turnout and as a result exit polls weren’t conducted in the 1980s. Fantastic huh? But, MPs early posts with quotes from a German pollster sums it up I think. Much different ballgame in Germany. They have an easier time. But that doesn’t excuse anything.
Hey, I’m about to start posting articles summarizing different aspects of my research over at the Election Audit Institute, a site recently established by Bruce O’Dell, former Vice President of USCV. http://www.electionauditinstitute.org/
Register and partake in the discussion!
Psychological assault forces Malkin to reveal Crowdist roots
Populist profiteer who infiltrated right unmasked
3:17 PM 8/21/2005
ANUS News
San Francisco (NNN) – A two week campaing of information terror culminated early yesterday with the Michelle Malkin award-winning blog disconnecting its trackback and response features as a consequence of ANUS activist antagonism. By taking refuge behind social pretense, and claiming to be victim, Malkin is validating claims that despite her right-wing pedigree, she is in fact an instrument of class/tribal revolt much as her leftist counterparts are.
Modern “conservatives” most commonly claim to be a bastion against a cluster of related mentalities, including Marxism and pacifism and rabble revolt, which are propelled by the ego-drama necessary to compensate for lack of superior traits in any form. The undifferentiated crowd will always be slaves to the machine, and they know it, and hate anyone who is not condemned to such a life. While Malkin insists that her “conservative” agenda would reverse this, her agenda in fact supports the very same mass revolt. ANUS activists were determined to prove this.
Using a scientifically-calculated mixture of gay pornography, racist innuendos and scatological/morbid humor, ANUS activists deliberately pushed Malkin and her blog maintainers past the tolerance threshold for socially offensive material. At this point, as psychological profiles such as the Stockholm study suggest, Malkin and staff violently reversed their positive outlook into a negative, paranoid one. Victory was achieved with a notice posted on the Malkin blog.
“We’ve had to turn the trackbacks feature off while Michelle is away. There was some nasty porno spam in the trackbacks and some other problems. I believe that Michelle’s kind assistants are working on the problem and hope to turn it back on when she returns. I’m sorry for the inconvenience,” said the statement, attributed to someone named “Betsy” (probably a bot, since even the conservatives prefer minkier names for their women these days).
Chuckling, ANUS activist Iconoclast, back from temporary confinement in the FMP room of the ANUS compound’s dank and fetid basement, conjectured about the psychology of Malkin’s staff. “I like how they said ‘some other problems,'” he said. “When the enemy gets vague, you know they’ve been forced into hiding something. What they’re hiding here is that they’re afraid to cross the taboo line and admit that most people are idiots, and that the idea of a Filipina conservative married to a Jew is ludicrous cover for their actual Crowdist activities.”
ANUS CTO (Chief Trolling Officer) Penisbird voiced his own sympathy for the situation. “People try to distract us with labels and orientations,” he said. “The truth is that politics is a question of personal integrity. Those who can lead do so, and the rest form masses to tear those others down because they hate them for being beautiful and smart and capable, and we call this mass revolt ‘Crowdism.’ Malkin is clearly a Crowdist, no matter what she insists.”
Apparently, the onslaught of sodomy, feces and burning crosses caused profound psychological change in the fifth columnist “conservative” Crowdist. Dr. Siegfried Freund, a goyish psychologist from Eastern Philadelphia, summed up his conclusion this way: “Burroughs says in Naked Lunch that the greatest hazard for an agent is merging with his cover story. Subversives like Malkin and her Jew penis, Craig Newmark, sublimate their actual intent, but when they’re forced up against the thread of Crowd disapproval, they snap back to the fight and in doing so, reverse polarity.”
Since the assaults began, Malkin’s comments have veered away from her traditional reactionary brain-in-a-bucket conservatism and have become increasingly like cocktail party small talk or the rhyming sentiments written in greeting cards. Dr. Freund predicts that within 48 hours she will own a pair of Birkenstocks and support al-Qaeda. In the meantime, ANUS infoterror group chalks up another success for its mixture of politics and antagonism that treats Net publishing the way it was meant to be treated: like a masochistic, incontinent, HIV positive gimp.
About ANUS
The American Nihilist Underground Society advocates nihilism, or a removal of interpretive layers from our perception of physical reality, as a means of transcending neurotic crowdism and thus achieving adaptive success. It has been online since 1995 and attracts thousands of readers daily with articles about philosophy, politics, music and culture. Every major internet filtering service bans anus.com, and many “anti-hate” organizations decry it as an anti-crowdist site which must be censored and its perpetrators bankrupted. http://www.anus.com/
About Nihilism
Nihilism is the belief that nothing we perceive has Absolute value; reality exists, but beyond its inherent meaning to us as the physical container of our existence, it has no significance outside of what we perceive. “The world is my representation,” indeed. When we strip away all of the values projected onto physical reality and its outcomes, we are left only with personal ideal and natural ideal, and bringing the former into adaptation with the latter is the lifetime task to which nihilism is a gateway. http://www.nihil.org/
About Michelle Malkin
Nationally-syndicated columnist Michelle Malkin began life by slandering liberals for the Los Angeles times, but was quickly promoted when it was discovered that her conservatism infused with leftist core values was a convenient propaganda talking point for the NWO, which supports Crowdist revolt in all its forms as heck, it’s good for business. Currently Malkin churns out dreary editorials endorsing the Same Old Failing “Solutions” to problems that never go away, and when she’s not busy with that, posts dramatic crap on her blog that people link so they don’t have to read it. She is married to Craig Newmark, a fat Jew with one Catholic grandparent. http://michellemalkin.com/
That is excellent. I imagine the electionauditinstitute.org is where the exit poll dialogue continued.
“election audit institute” sounds exactly like my view of election reform..well except, that I think in the end American’s deserve transparent elections in the British Mode, paper based and hand counted. I doubt any respectable American institution would promote that view.
I will definitely register. Thanks for the lead.
That is excellent. I imagine the electionauditinstitute.org is where the exit poll dialogue continued.
“election audit institute” sounds exactly like my view of election reform..well except, that I think in the end American’s deserve transparent elections in the British Mode, paper based and hand counted. I doubt any respectable American institution would promote that view.
I will definitely register. Thanks for the lead.
Nice Thomas. So I can blame ANUS for my 150 truly nasty trackback spam pings per day. Thanks.
Mark,
Enjoy your vacation.
Hmm it is tempting to use this post as a type of open thread with your readers. Any objections?
I’ve been wanting to put a couple of questions out there on past topics, but the topics had not come back up.
I’ll wait a bit before posting.
So I basically had one question:
Feeble stated that nonrespondent bias (conservative’s not participating in exit polls) has been accepted and factored into English exit polls. Can anyone point to the literature on this UK procedure?
Exit poll bias causes so much convoluted arguing in America that we could probably benefit from the knowledge England has developed on the subject.
Alex, you can check out a Jurgen Hofricther’s chapter, “Exit Polls and Election Campaigns” in the Handbook of Political Marketing, ed. Bruce I Newman, 1999. He discusses the “Shy Tory” effect documented in the 1992 election.
The exit polls actually turned out a little better than the pre-election polling. Hofrichter cites three works by Worcester in 1997 and 1998 that discusses this issue in greater detail. I didn’t bother to dig up those citations.
You can also read about the exuberant liberal voter effect turning up in Ontario Canada exit polls in 2003. Google “You’re Leaving? Great! Exit Polling in the 2003 Ontario General Election.” Their poll significantly overstated the Liberal vote and cite differential response as a possible explanation. “Either voters misreported their vote to our interviewers or Liberal voters were simply more willing to reveal their preference to us.” Pg. 9 They admitted that evidence of this theory was “hard to come by.”
BTW – Differential nonresponse has been the suspect cause of most exit poll bias for over 20 years. Plissner and Mitofsky (1982, p. 15) first raised the possibility of nonresponse bias explaining that there “is no guarantee that voters who respond to the poll are like those who refuse to answer.” In reference to the 1984 presidential election exit polls, Mitofsky and Waksberg (1985, p. 5) noted that one of the problems with projections based on exit polls that did not exist when projections were based on vote returns is the bias introduced by nonresponse, although no specific evidence of bias was presented by the authors.
Mitofsky (1991 p. 96) attributed the bias observed in the 1988 presidential election exit poll to “people who refuse to be interviewed.” Partisan differential response was fingered for problems with the 1992 presidential election exit polls as well (Mitofsky and Edelman 1995, p. 92).
Warren Mitofsky is quoted as saying, “The only thing it can be (consistent Democratic bias) is a difference in response rates of supporters of one candidate versus the other. There’s no possibility for it to be anything else and be this systematic” (The Roper Center for Public Opinion Research 1993).
The Research Triangle Institute (RTI) report on the 2000 election network coverage problems noted that exit poll nonresponse had become a “serious problem” (Konner, Risser, and Wattenberg 2001, p. 21) and Edison Media Research and Mitofsky International (2005, p. 3) cited differential nonresponse as the likely source of the exit poll discrepancies of election 2004.
Back in 1969, Warren Mitofsky and Murray Edelman found that deviation from the sampling interval was the greatest contributor to the exit poll error. That is, as interviewers did not stick to their assigned interviewing rate, they tended to “self select” or avoid certain types of voters, introducing bias. This may not be a problem if interviewer characteristics were randomly distributed (and therefore so would be their natural biases), but they clearly are not.
References:
Edison Media Research and Mitofsky International. 2005. “Evaluation of Edison/Mitofsky Election System 2004” Retrieved from http://www.exit-poll.net/election-night/EvaluationJan192005.pdf. Last accessed August 1, 2005.
Konner, J., Risser, J., and Wattenberg, B. 2001. “Television’s Performance on Election Night 2000: A Report for CNN.” Retrieved from http://www.cnn.com/2001/ALLPOLITICS/ stories/02/02/cnn.report/cnn.pdf
Last accessed, August 1, 2005.
Mitofsky, Warren J. 1991. “A Short History of Exit Polls.” In “Polling and Presidential Election Coverage.” Eds. Paul J. Lavrakas and Jack K. Holley. Sage Publications: Newbury Park.
Mitofsky, Warren J. and Murray Edelman. 1995. “A Review of the 1992 VRS Exit Polls.” In “Presidential Polls and the News Media” Eds. Paul J. Lavrakas, Michael W. Traugott, and Peter V. Miller. Westview Press; Boulder, CO.
Mitofsky, Warren J. and Joseph Waksberg. 1985. “Election Estimation from Incomplete Sample Data.” Paper presented at the Public Opinion Surveys and Electoral Forecasting conference, Sevilla, Spain, September 25-27.
Plissner, Martin and Warren Mitofsky. 1982. “Voting Twice on Election Day.” Public Opinion, August/September.
The Roper Center for Public Opinion Research. 1993. “The Polls and the 1992 Elections: A Public Perspective Symposium; Problems in Exit Polling; Interviews with Warren J. Mitofsky and John Brennan.” The Public Perspective. January, 1993/February, 1993. Vol. 4, No. 2; pg. 19
Rick, thanks for the literature review. I will try to dig up the Worcester works.
But, Rick, can you tell how the debate in England proceeded when they decided to institutionalize, rather weight by, respondent bias in exit poll calculations? I’m thinking more in a policy way than strictly scholarly way.
Also, there is an important reason why the British data and scholarly work is so important. At the time Feeble made the important point that the British felt confident in looking for errors within the exit polling, because the vote counts were so trusted, i.e. hand counts of paper ballots conducted at the precinct by two people from opposing parties, if I’m not mistaken.
If you take the British trust in the vote count and the decision to include non-respondent bias in exit poll calculations, British academics must have reached some strong conclusions. Much stronger than anything we should expect in America becuase of our laughable vote count processes, if not laughable vote counts themselves.
With all due respect to Mitofsky, et al, any analysis of American exit polling is clouded by uncertainty in specific vote results, IMO.
So I hope the Worcester papers shed some light. If anyone can provide anymore data on British exit polling specifically, please post.
Thank you!
Alex in Los Angeles
Alex, great questions about the British poll process. Unfortunately, I haven’t had time to look into it, but I did understand that they were weighting for the affect. However, I’m not sure that they did for the recent election. Quite unsure actually.
True, febble says the poll results were immediately cast into question because their elections are transparent. I don’t know much about their process. However, the poll results have been thrown into question in the US since the first projection was made back in 1982, and I’m not sure it had anything to do with the transparency of the process here.
Something also to consider about the British exits. They only ask one question. Very simple. All you do is check a box and drop your slip. Our forms are two-pages long! Consider that Democrats are more likely to be women, college students, unemployed than Republicans, they could have a bit more time on their hands than the average Republican. Again, speaking in terms of averages.
Another potentially interesting issue. The vote count totals used in the exit poll analysis may be “polluted” with absentee/early votes. As I understand it, some counties count these votes separately, others lump them in the home precincts to be counted with all the election day ballots. When the E-M worker obtains the final vote count result from the precinct, they send it in. They don’t know if their totals are only for election day voters or are mixed. Who knows if this is a significant problem or not? But, given that Republicans are much more likely to vote early or absentee, if the problem was widespread, it could “look” like ballot stuffing because the exit poll is not sampling the same population it is intended to represent.
The biggest clue is the fact that bias increased in 2004 as sampling rate increased. That is, as interviewers have more freedom to deviate from the scientific sample, the bias increases, suggesting that interviewers gravitated toward more willing and/or more “like” voters. This is the same thing Mitofsky and Edelman found in 1969. However, when febble’s alpha was used instead of Mosteller’s measure 5 (WPE), the relationship between sampling interval and bias was striking. The bias steadily increased from almost non-existent when every voter was supposed to be approached, to off the charts when every 10th voter was to be approached (Mitofsky presented this chart in Miami). That’s pretty solid evidence of interviewer induced differential non-response.
But, bottom line – E-M never said that their wasn’t fraud or that the vote count was accurate. Only that the exit poll data is more consistent with a differential non-response hypothesis than a fraud hypothesis. That is a very sound conclusion given the available data.
Hi Rick,
Excellent points on the absentee reporting uncertainty and simplicity of British elections. Exactly what I meant by laughable American vote count processes. I probably should have said any analysis of American exit polling is clouded by the complexity, not necessarily uncertainty, of our vote results.
I personally don’t doubt the existence of non-respondent bias, but I would quibble with you slightly in that I think the non-transparency of American elections, and their complexity, absolutely delayed, and still delays, a clear and sufficient understanding of nonrespondent bias in American exit polls. If British exit pollsters feel comfortable incorporating the bias into their exit polls, assuming Feeble was correct that they do, I can only imagine they have reached that level of understanding of exit poll bias within the British context. Hence my interest.
And there’s always the German context! I don’t know, but maybe, they don’t even need to worry about nonrespondent bias. I’m willing to admit this could be for reasons unique to their election/exit polling processes and context.
Just curious really on what we can learn. And I still hope to get into the Worcester stuff and anything else anyone can recommend on the apparent British resolution of their exit poll bias.
Oops, I think my post does not acknowledge the great progress made this year in uncovering and finally starting to explain to the public the causes of exit poll/vote result bias. My quibble was with the historical slowness, since 1982 according to your note on Mitofsky work, in coming to grips finally with exit polling bias.
Also, I think more progress has been made than I was aware of. I didn’t realize that Feeble’s alpha was applied to the sampling interval.
This is great as it may finally spur investment in shoring up interviewer training and pairing of interviewers. Frankly, it was the weakness of the exit polling interview methodology that gave the non-respondent bias theory immediate apparent validity in my book.
Was Feeble’s alpha applied to other variable’s with notable results? I guess Mitofsky would be smart to redo his whole analysis using the FA, or something better.
Thanks for the response, Rick.
FYI:
http://www.projectcensored.org/newsflash/voter_fraud.html
Dr. Loo so obviously did not dig deep into the current state of the exit poll dialogue. Of course, he’d have to read MP for that 😉
Alex, my research has, quite frankly, been a major pain in the ass. Assembling my list of sources took months and I’ve probably annoyed many in the process. I’m pretty darn frustrated at how little research has been conducted on exit polling and exit poll error. The list of references, which is comprised largely of conference papers, that deal specifically with analysis of presidential election exit poll data is about a page.
Your points above are well heeded. I don’t know much about the German polls, other than that they were so accurate in 1978, there was an outcry because people thought it depressed voter turnout and as a result exit polls weren’t conducted in the 1980s. Fantastic huh? But, MPs early posts with quotes from a German pollster sums it up I think. Much different ballgame in Germany. They have an easier time. But that doesn’t excuse anything.
Hey, I’m about to start posting articles summarizing different aspects of my research over at the Election Audit Institute, a site recently established by Bruce O’Dell, former Vice President of USCV.
http://www.electionauditinstitute.org/
Register and partake in the discussion!
Psychological assault forces Malkin to reveal Crowdist roots
Populist profiteer who infiltrated right unmasked
3:17 PM 8/21/2005
ANUS News
San Francisco (NNN) – A two week campaing of information terror culminated early yesterday with the Michelle Malkin award-winning blog disconnecting its trackback and response features as a consequence of ANUS activist antagonism. By taking refuge behind social pretense, and claiming to be victim, Malkin is validating claims that despite her right-wing pedigree, she is in fact an instrument of class/tribal revolt much as her leftist counterparts are.
Modern “conservatives” most commonly claim to be a bastion against a cluster of related mentalities, including Marxism and pacifism and rabble revolt, which are propelled by the ego-drama necessary to compensate for lack of superior traits in any form. The undifferentiated crowd will always be slaves to the machine, and they know it, and hate anyone who is not condemned to such a life. While Malkin insists that her “conservative” agenda would reverse this, her agenda in fact supports the very same mass revolt. ANUS activists were determined to prove this.
Using a scientifically-calculated mixture of gay pornography, racist innuendos and scatological/morbid humor, ANUS activists deliberately pushed Malkin and her blog maintainers past the tolerance threshold for socially offensive material. At this point, as psychological profiles such as the Stockholm study suggest, Malkin and staff violently reversed their positive outlook into a negative, paranoid one. Victory was achieved with a notice posted on the Malkin blog.
“We’ve had to turn the trackbacks feature off while Michelle is away. There was some nasty porno spam in the trackbacks and some other problems. I believe that Michelle’s kind assistants are working on the problem and hope to turn it back on when she returns. I’m sorry for the inconvenience,” said the statement, attributed to someone named “Betsy” (probably a bot, since even the conservatives prefer minkier names for their women these days).
Chuckling, ANUS activist Iconoclast, back from temporary confinement in the FMP room of the ANUS compound’s dank and fetid basement, conjectured about the psychology of Malkin’s staff. “I like how they said ‘some other problems,'” he said. “When the enemy gets vague, you know they’ve been forced into hiding something. What they’re hiding here is that they’re afraid to cross the taboo line and admit that most people are idiots, and that the idea of a Filipina conservative married to a Jew is ludicrous cover for their actual Crowdist activities.”
ANUS CTO (Chief Trolling Officer) Penisbird voiced his own sympathy for the situation. “People try to distract us with labels and orientations,” he said. “The truth is that politics is a question of personal integrity. Those who can lead do so, and the rest form masses to tear those others down because they hate them for being beautiful and smart and capable, and we call this mass revolt ‘Crowdism.’ Malkin is clearly a Crowdist, no matter what she insists.”
Apparently, the onslaught of sodomy, feces and burning crosses caused profound psychological change in the fifth columnist “conservative” Crowdist. Dr. Siegfried Freund, a goyish psychologist from Eastern Philadelphia, summed up his conclusion this way: “Burroughs says in Naked Lunch that the greatest hazard for an agent is merging with his cover story. Subversives like Malkin and her Jew penis, Craig Newmark, sublimate their actual intent, but when they’re forced up against the thread of Crowd disapproval, they snap back to the fight and in doing so, reverse polarity.”
Since the assaults began, Malkin’s comments have veered away from her traditional reactionary brain-in-a-bucket conservatism and have become increasingly like cocktail party small talk or the rhyming sentiments written in greeting cards. Dr. Freund predicts that within 48 hours she will own a pair of Birkenstocks and support al-Qaeda. In the meantime, ANUS infoterror group chalks up another success for its mixture of politics and antagonism that treats Net publishing the way it was meant to be treated: like a masochistic, incontinent, HIV positive gimp.
About ANUS
The American Nihilist Underground Society advocates nihilism, or a removal of interpretive layers from our perception of physical reality, as a means of transcending neurotic crowdism and thus achieving adaptive success. It has been online since 1995 and attracts thousands of readers daily with articles about philosophy, politics, music and culture. Every major internet filtering service bans anus.com, and many “anti-hate” organizations decry it as an anti-crowdist site which must be censored and its perpetrators bankrupted.
http://www.anus.com/
About Nihilism
Nihilism is the belief that nothing we perceive has Absolute value; reality exists, but beyond its inherent meaning to us as the physical container of our existence, it has no significance outside of what we perceive. “The world is my representation,” indeed. When we strip away all of the values projected onto physical reality and its outcomes, we are left only with personal ideal and natural ideal, and bringing the former into adaptation with the latter is the lifetime task to which nihilism is a gateway.
http://www.nihil.org/
About Michelle Malkin
Nationally-syndicated columnist Michelle Malkin began life by slandering liberals for the Los Angeles times, but was quickly promoted when it was discovered that her conservatism infused with leftist core values was a convenient propaganda talking point for the NWO, which supports Crowdist revolt in all its forms as heck, it’s good for business. Currently Malkin churns out dreary editorials endorsing the Same Old Failing “Solutions” to problems that never go away, and when she’s not busy with that, posts dramatic crap on her blog that people link so they don’t have to read it. She is married to Craig Newmark, a fat Jew with one Catholic grandparent.
http://michellemalkin.com/
That is excellent. I imagine the electionauditinstitute.org is where the exit poll dialogue continued.
“election audit institute” sounds exactly like my view of election reform..well except, that I think in the end American’s deserve transparent elections in the British Mode, paper based and hand counted. I doubt any respectable American institution would promote that view.
I will definitely register. Thanks for the lead.
That is excellent. I imagine the electionauditinstitute.org is where the exit poll dialogue continued.
“election audit institute” sounds exactly like my view of election reform..well except, that I think in the end American’s deserve transparent elections in the British Mode, paper based and hand counted. I doubt any respectable American institution would promote that view.
I will definitely register. Thanks for the lead.
Nice Thomas. So I can blame ANUS for my 150 truly nasty trackback spam pings per day. Thanks.