Tuesday, January 04, 2005

Improving the elections

First, my apologies for irregular posting. The reason comes down to three words: "Work, work, work."

Not long ago, I spoke to a recount volunteer in Ohio who asserted that the real conspiracy may be to keep us all working so hard just to pay sky-high rents that we have no time to think about larger issues. Conversely (or perhaps as illustration of that very point), a former friend recently let me know that anyone in today's society who allows himself a few hours each day for non-paid activity -- such as, say, running a blog -- must be a Seutonian decadent. As I said, he's now a former friend.

Forgive the bitching. That's not why you came here.

Right now, I'd like you to meet one Larry English, whose qualifications you will soon read. He has cobbled together an open letter on the subject of our troubled vote.

I'm not sure I agree with any analysis which argues that the problem is one of process, as opposed to criminal intent. And his suggestions for improvement bring us, as always, to the question of how to bell the cat.

Still, he makes some good points. So I thought I'd pass along his words.

To Concerned Citizens:

I am a professional in the area of information quality. I am very concerned that the election processes apparently are still severely broken even after Congress appropriated $3.86 Billion for Election "upgrades" in 2002. The evidence is clear from continued issues observed in the press and online discussions.

There are several things we understand in our field that are not well understood at large or by those who have the ability to "reform" the election processes. I would like to briefly speak to some of these points because they are particularly relevant at present.

Please let me share briefly some of my credentials to speak to this matter. I have been called one of the top authorities in the world in assessing and improving the reliability and accuracy of information processes. My book, "Improving Data Warehouse and Business Information Quality," was called "the Information Bible for the Information Age" by Masaaki Imai, creator of the Kaizen quality system, used by many world-class firms. It has been translated into Japanese by the first information services organization to win the Deming Prize for Quality. I was featured on the cover of the American Society for Quality’s "Quality Progress" Jan 2000 issue, as "One of 21 Voices for the 21st Century."

I conducted a major analysis of the 2000 Presidential Election, "Information Quality Mandate for Election Reform," in "DM Review," a four-part series beginning in Oct 2001. A PDF file of the complete article is found at http://www.infoimpact.com/newspdf/DMR_10.01IQLessonsFromElection2000-w-Links.pdf.

Problems with 2004 Electoral Information Processes:

Despite the considerable attention and concern on the part of both the government and concerned citizens for the issue of the reliability of electoral information processes since 2000, there are still numerous, recurring problems that have been observed in the 2004 election. A very small sample includes:

* Washington State has had multiple recount(s), including belatedly discovered ballots

* 46,000 voters were discovered registered in both Florida and New York

* 58,000 absentee ballots "disappeared" in Florida

* 38,400+ Election 2004 incident reports have been filed according to Verifiedvoting.org

* Numerous irregularities and controversy in Ohio, including write-in votes allegedly incorrectly defaulting to one candidate when run through the voting machine

* Voter registration problems persist

* Provisional ballot problems, including rejection of two thirds of Florida’s provisional ballots

* Electronic voting machine malfunctions

* Electronic voting machine failures occurred this year in Georgia, Maryland, California and other states, but the companies that certify the machines refuse to discuss the flaws.

Some Important, Relevant Principles:

Assuring the reliability of electronic technology is not a mysterious art; it's an established field of practice in use by many information professionals. As I examine these reports and many others, it strikes me that it may be helpful to present a few key principles that seem particularly relevant at present:

First, automating a process doesn't assure accuracy. Machines or computer software cannot guarantee accuracy in themselves. Assessing the accuracy of election processes requires human observers, since only human observers comparing against a real world entity (ballots, in this case) can play that role. Many software applications implement "edit and validation" rules in a way that causes errors. Furthermore, producing accurate data requires processes and data collection devices that are clearly defined and controlled, as well as clear guidance for the information producers (the voters, in this case).

Second, while assuring the integrity of capture of the individual ballots (and their count) is the central issue in any election, assuring accurate vote capture and counts requires assessing the reliability of the electronic voting devices and processes, not just individual ballots. Inaccurate votes are produced by defective processes. Not focusing on the process can lead to bias in conclusions. For example, some of the apparent "errors" we're hearing about may be caused by sampling error, such as an apparent lack of random selection of districts that's been reported for the 3% test recounts in Ohio.

Third, before use of electronic voting devices processing election information in private, we had defined election processes that could be audited (recounts). Apart from voter anonymity, all steps of the process were subject to examination. When electronic voting devices are used, they entail the need for rigorous assessment of all components, including hardware, software, backup and recovery, against various types of failure (hardware, data storage, software, power outage, etc.) integrity against tampering, chain of custody control, and ability to audit the results for recount purposes.

My Recommendation:

My recommendation is that there should be a vigorous investigation into the systemic and pervasive problems in the integrity and reliability of the election processes, including but not limited to voter registration, early voting, provisional voting, election-day voting, vote counting, vote recount, electronic voting, chain of custody, etc.

As an American citizen, I am very concerned about the persistent problems.

As an expert in applying quality management principles to information processes, I would be pleased to share my insights or provide expert testimony to the problems, root causes, and to the real reforms (process improvements) required to error-proof this most- sacred privilege Americans have, to select our representatives in government. Some requirements for real and sustainable election reform include:

* Accessible, easy and open registration processes

* Voter-friendly ballot design (regardless of voting technology)

* Well-defined voting processes that are error-proofed, reliable and controlled and auditable

* Chain of custody controls to prevent tampering

* Training of election workers so they can error proof their processes

* Training of voters (as to types of problems with the voting devices & how to prevent errors and disqualified votes)

Thank you for allowing me to help improve one of the most important, but one of the most broken processes in our democratic society.

Sincerely,

Larry English, President INFORMATION IMPACT International, Inc. Larry.English@infoimpact.com

1 comment:

Anonymous said...

I think it's too late to talk about "improving" the way we do elections. The problem isn't mainly that the mechanics are all broken; the problem is more that we are being ruled by people who wish to break our elections. They want to break just about everything else, too.

No, I don't have any good ideas. The best I can come up with is to emigrate somewhere where you can get your citizenship and then have your votes actually counted. Of course Bush will be knocking on the door anyway to remind you that you are with him or you are with strawmen. It's your choice.

-- chemoelectric.org