Knight Of The Mind

I'll do my best to present a philosophical and generally conservative look at current events and life, the universe and everything. Readers are invited to take all that's posted herein with a grain of salt. or if they prefer, a grain of salt, a slice of lime and a shot of tequila.

Location: Alexandria, Virginia, United States

Greetings and welcome. My name is Steve, I'm 35 years old and I work for the US Army as an Operations Research Analyst. Hence my blog title Knight Of The Mind.

Tuesday, January 04, 2005

Exit Left: The 2004 Exit Poll Failure

Mickey Kaus continues to cement his reputation as a Democrat with an actual brain. Today he links to an excellent takedown of Edison/Mitofsky, the firm that conducted the Exit Polls for the 2004 Presidential Elections, by the blogger Mystery Pollster. Kaus points out the following:

As late as 7:33 P.M. on Election Day, Mitofsky and Lenski were apparently telling their clients (NBC, CBS, CNN, AP, etc.) that after "weighting" Kerry was beating Bush by 9 points among women and losing by only 4 among men.

He backs his claim with this Edison/Mitofsky actual report that was forwarded to the National Pool on election Night. The real results, contained in the Edison/Mitofsky "revised report", nearly mirror the final actual election totals.

On the surface, this just appears to be another example of gullible wishful thinking by liberal members of the media. They all got "pantsed" as badly as Dan Rather got it from Bill Burkett. They bought from a snake oil merchant who sold them the lie they wished was true. Another day, another line of media Bravo Sierra, so what?

Well actually exit polls serve an important and worthwhile purpose when dome properly. They can also serve the cause of mayhem when they are conducted the way Arthur Anderson used to audit Enron. An honest broker conducting an exit poll is Democracy's best check and balance for fraud on election day. If done fairly and competantly, exit polls are the insurance policy that guarantees a fair counting of the votes.

The recent overturn of a fraudulent election in the Ukraine has been partially credit to the work of exit polsters who smelled a rat. Perhaps no one would have known the real outcome if the exit polls had not been taken. The international pressure that led to a revote would never have been applied absent solid data from the polls.

The exit polls in Venezualia showed Hugo Chavez getting metaphorically treated the same way he physically treated leaders of his country's opposition parties. It took hours of careful "inspection and certification" by Former US Preseident Jimmy Carter to turn that election into a mandate for Chavez. It's a wonder Carter hasn't called for a recount of the results from all 50 states in 1980. So a set of exit polls can't always guarantee the correct outcome, but they do make it harder to explain away a broken process.

But what happens when the exit poll process itself fails to effectively work. We saw this in 2002, when the computers crashed and the networks got no data. We saw the process break in a different way in 2004. Mystery Pollster suggests things were not done in accordance with good polling ettiquete. His description of the polling process follows below:

This year's NEP exit poll interviewers were trained via telephone and most worked for just one day without supervision.

This slipshod methodology leads to doubts concerning the ultimate validity of the 2004 election. The academic papers concerning the supposed miscounting of the votes in 2004 are already amounting to mathematical propaganda. The sentence below is undoubtably true:

To carry our analysis further, we can employ a normal distribution curve (see
Figure 1) to determine—again assuming proper poll methodology, no discriminatory voter
suppression 24, and an accurate and honest popular vote count—that the probability that
Kerry would have received his reported popular vote total of 48.1%, or less is one in959,000—a virtual statistical impossibility.

If you doubt me, run the numbers on SAS or SPSS. These guys appear to be talking the talk only after they've walked the walk. The numbers run; just like they tell you. Just looking at those numbers would make the theoretical 50/50 undecided voter mad as all get-out that Bush got all those electoral votes. This is the rather obvious intent the authors had in mind before they ever put finger to keyboard.

Here's where they perhaps accidently mislead. They assume apriori that the exit polls are gospel. They totally ignore the evidence idicting the exit polls and focus instead on the prior resumes of the lead data managers. It's only when you overwhelmingly assume proper poll methodology that the result of the current election falls into doubt.

This is why the exit polling messup is more than just a minor problem. The exit poll results become an historical document that can be innacurately or dishonestly interpreted the same way Michael Bellesiles incorrectly read historical documents to present false history when writing The Arming Of America.

Update I: Ace Of Spades discusses Kaus' post as an exoneration of bloggers who were duped by the false exit polling data.

Update II: It seems the exit polling results from the Presidential race aren't the only funny results floating around. Cold Hearted Truth claims Seattle, WA had more ballots than voters. At least the exit pollsters can quite manage to pull of that one.


As you may or may not already be aware, members of the Watcher's Council hold a vote every week on what they consider to be the most link-worthy pieces of writing around... per the Watcher's instructions, I am submitting one of my own posts for consideration in the upcoming nominations process.
Here is the most recent winning council post, here is the most recent winning non-council post, here is the list of results for the latest vote, and here is the initial posting of all the nominees that were voted on.

Weblog Commenting and Trackback by