Kerry Won!!! Statistical Tools Everyone Can Use

Sunday, 25 December 2005, 7:43 am
Opinion: Michael Collins

USEFUL RELATED LINKS:

The Law of Large Numbers & Central Limit Theorem: A Polling Simulation

Excel Polling Simulation Model

2004 Election Model Projection; Exit Poll Collection; Excel Interactive Election Simulation; Other links.

The Kerry concession speech on November 3, 2004 marked the beginning, not the end of the controversy over the 2004 election. Just hours before the speech, Vice Presidential Candidate John Edwards emerged and said that, “John Kerry and I made a promise to the American people that in this election every vote would count and every vote will be counted."

Democrats were in a state of shock. 2004 was a banner year for new registrations, party financial support, and activism. Reported new registrations favored Democrats all over the country. Democrats were well ahead of Republicans in new registrations in Ohio. South Florida, the “scene of the crime” in 2000, saw major Democratic efforts and a lackluster Republican response.

Democrats matched and exceeded Republicans in funds raised. For the first time, the internet proved to be a highly potent form of fund raising. The Democrats collected $10 million a month for the Kerry Campaign on the Internet alone. Other groups supporting the Democrats raised substantial funds. MoveOn.Org and New Democratic Network ran parallel campaign commercials and provided other support with the $25 million they raised during the election cycle.

Activism was at an all time high. People who had never worked in elections volunteered in large numbers and local Democratic parties throughout the country saw a surge in citizen participation.

While Kerry may have conceded the election at 2:14 p.m., Nov. 3, a large portion of the population failed to accept the final results. They knew something was wrong.
As one Virginia activist said, “This is simply not possible, the national results or here in Virginia.”

Captured truth—the four National Exit Polls

Just as Election Day 2004 was winding down, the server at Edison – Mitofsky (EM) crashed. EM was the firm conducting the National and State Exit Polls for a media consortium consisting of the four major networks, CNN, the Washington Post and other major media concerns. In an historical article, "Scoop" Independent Media outlined the process that allowed Jonathan Simon to capture elements of the National Exit Polls that were never intended for public consumption. Between 11 p.m. election night and 1:33 a.m. November 3, 2004, EM’s servers were out and so was a major portion of the truth.

Using the national and state exit polls as a basis, TruthIsAll began a detailed analysis of the exit polls that demonstrated again and again the extremely high odds that Bush could have won the popular or electoral vote without a 24 hour suspension of the laws of mathematics and statistics.

The Debate Begins

The Internet hosted the first series of debates on the disparity between exit poll results and the final vote count. The Exit Polls were of major interest. In 2004, his final exit poll, 12:25 am showed Kerry winning by 3% (with a 1% Margin of Error). Thirteen hours later, the media consortium and Mitofsky revised the final poll by adjusting poll results with actual election results. To no one’s surprise, that poll matched the Bush victory margin, a reversal of the day-long live polling of 13,047 respondents’ nationwide.

The capture of unintended polling results and the clear discrepancy between polls (which have a paper trail to this day), the recorded election day votes (many of which were without a paper trail), prompted vigorous challenges to the election results. For the first time in history, many claimed that the public opinion and exit polls proved the election results wrong, rather than the other way around.

TruthIsAll’s new tools for election analysis for the politically motivated but math challenged: the Interactive 2004 Election Simulation and Monte Carlo Polling Simulation.

TruthIsAll is the username of a prodigious internet poster who has focused on bringing applied mathematics to current events. But the raging debate on the relevance of the National and State Exit Polls and, to a lesser degree, the pre-election preference polls, has been his main focus.

TIA shows the details of his assumptions and calculations and encourages others to review them. Many of the debates relating to the analyses appeared on democraticunderground.com. The rhetorical flourish was intense, unlike more moderate academic debates, and the statistical concepts were difficult to grasp for many. This did not prevent an avid following of his work from developing. TIA is responsible for over 150 separate statistical analyses on the exit polls and other matters. His work has appeared in hundreds of Internet news sources and blogs over the past 13 months.

In response to the growing movement which questions the Bush legitimacy and the desire of more and more users to understand the basic mathematical formulae upon which the analysis is based, TIA just released a revised version of his Excel-based Election 2004 Interactive Election Simulation. The model allows users to do a variety of calculations which show how unlikely it is to think that Bush won. It contains a wide variety of data including pre-election polls, the National Exit Polls, the State Exit polls, and other critical datasets. Users can adjust demographic variables (2000 and new voter turnout, gender, party-id, etc.) which proved so critical in the 2004 election results.

The post explains the basic mathematical and statistical concepts of polling in plain English, allowing a much broader audience to understand the earlier Interactive Election Model. These educational and analytic tools are now available to all who choose to use them on the internet at the links posted above.

When was your first post on the 2004 election controversy?

Nov. 4, 2004: On DU “To Believe Bush Won the Election….” Will Pitt referenced the full post in an article he wrote a few days later and it spread all over the Net.

How many main posts on DemocraticUnderground (DU) and ProgressiveIndependent (PI) and other forums have you written on the election?

About 150 original posts.

Why all this energy devoted to the 2004 election?

I was appalled when the election was stolen from Gore in 2000. I had posted daily projections of my Election Model on DU in the four months leading up to the election. The projections were based on state and national polls. The final national model had Kerry 51.6 % of the two-party vote; the state model 51.8 %. The state model included a Monte Carlo simulation with a Kerry expected total of 337 electoral votes.

The initial exit polls and the Iowa Election Markets showed that Kerry was a 3% winner. When Bush came from nowhere after 9 p.m. to win, I had this feeling of Déjà vu. So I decided to confirm the doubts using mathematical probability analysis, based on the
preliminary exit poll data downloaded by Jonathan Simon.

Your posts on DU are like no others. They are intense, complex, and require careful reading? What do you hope to accomplish with your new work.

My goal is to present the analysis in such a way as to inform both those who may have limited training in mathematics and those more knowledgeable. At the same time I want to negate the straw man arguments of those who attempt to obfuscate and misrepresent the accuracy of the pre-election and exit polls.

Where can you be reached for questions on this work and other dialog?

I’m posting on PI (primarily in the “Elections & Voting Rights” forum – which contains a collection of TIA's most recent posts.)

From your first posts on DU, you have attracted a large group of avid supporters? What is it that draws these users to your work?

It’s probably because they are frustrated like I am that the truth has been kept hidden by the media, politicians, Democratic bloggers and professional naysayers. They appreciate that the analysis is comprehensive and honestly arrived at.

In addition to supporters, there have been smaller cadres of users highly critical of your work. How do you explain that?

I firmly believe a few have a clear agenda to debunk the exit polls. But how do they explain the pre-election polls? Others are in denial. They can’t accept that their votes don’t count. Many are math-phobic and miss the essential analysis. Others feel the need to impress colleagues with their conservative credentials by knocking a spread-sheet blogger. Others just think I’m just an arrogant wise-ass.

Are the criticisms you receive helpful in revising your thinking or redirecting your analysis? Examples?

Yes. They inspire me to do additional analysis. As an example, the UCSV simulation which debunked the reluctant Bush responder (rBr) hypothesis was attacked by the usual suspects. I decided to take a different approach than USCV and developed an Exit Poll Response Optimization model based on the same E/M 1250 precinct response and WPE
data. The results confirmed the USCV analysis. Later, I did a similar analysis based on state exit poll data which also confirmed USCV.

Who and what are the major influences on your work, both in applied mathematics and political advocacy?

I almost failed elementary 9th grade algebra. One day as I was entering class, I saw my mother speaking to the teacher. From that day forward, I studied hard and got a perfect score on my 12th grade advanced algebra final. I went on to get bachelors and two masters degrees in applied math. As a career, I have focused on applying mathematics in the development of engineering, investment and financial models.

As far as political influence, my favorites: JFK, Hubert Humphrey, George
McGovern and Harry Truman, Adlai Stevenson, Edward R. Morrow, Bill Moyers.

Anything you’d like to say in closing?

I’ll keep posting as long as I have something to say. If my work has in some small measure helped to raise awareness, then it’s all been worth it.

Thank you for taking the time to talk with me.

The miracle of timing that made the full set of national exit polls available to the world and the election fraud movement.

11/02/04

11:00 p.m. EM Server goes down. Exit Polls available on the Internet.

12:25 a.m. Final National Exit Poll captured by Jonathan Simon.
3047 Respondents: Kerry 51% - Bush 48%

11/03/04

1:33 a.m. Edison Mitofsky Servers come back online effectively ending
Internet access to the National Exit Polls.

2:30*a.m. Vice Presidential candidate John Edwards addresses the crowd and
and nation promising to “fight for every vote.” (time approximate)

1:25 p.m. National Exit Poll revised by EM to take final election results into
account 13,660 respondents: Kerry 48% - Bush 51%

2:14 p.m. Kerry concedes the election to Bush.

This remarkable sequence of events was not the beginning of the end but rather the end of the beginning. The remarkable intensity of research, analysis, and debate that energizes the modern day election fraud movement was born crying “foul!”

END

© Scoop Media