Sunday, November 04, 2007
Brennan Center and Verified Voting learn old lesson the hard way
The field of election audits is one that is attracting a lot of academic attention. It's a good field for academia where learned papers, funding, thesis topics and studies are aplenty.
It seems though that the academics involved are squabbling over the honey pot instead of focusing on the bigger social goals and picture.
The public view is centered on the need to have election auditing that is above reproach and using mathematically sound and agreed methods to give people a high level of confidence in the outcomes.
Of course exacerbating the situation in the US is the counting software and records available for auditing that are woeful at best. Work has begun - such as by the California SOS - to use open public recording standards for publishing election results. In the meantime statisticians have only bare information to work with. Nevertheless good strides have been made in the past couple of years in the area of election audit mathematics.
Enter the Brennan Center and Verified Voting who together sponsored an October event in Minnesota to showcase election auditing techniques and capabilities.
The first hint that this may be problematic for them is that they wished to keep this somewhat exclusive and to pick and choose who came and who presented. Lesson #1 - if you are espousing open public election auditing its a pretty bad idea to restrict attendance to only those people you think should be there.
Next up Lesson #2 - this is a field that demands peer-review and ideas that can be universally verified as sound and in-line with the latest state-of-the-art and consensus around that (see here). It looks really bad if you exclude people who are critical of your central ideas and finding flaws that you appear not to be able to counter or otherwise refute except by heavy handedness.
Lesson #3 - hiding does not get the job done. Using hotel staff to harass would be participants and exploiting arcane local legal bye-laws to circumvent basic free-speech rights looks downright nasty and mean spirited. It also puts you in jeopardy for future events where government attendees may see your prior actions as falling foul of OMB restrictions on their event attendance. In short if your ideas cannot stand up to public scrutiny and fair
open comment there is no way you can have any credibility in this field. It's also not reality in the age of the Internet
and collaborative communities online that demand openness. Using
face-to-face opportunities to drive consensus building is the perfect opportunity - nothing like the lobby of a hotel or
the bar to really allow people to align their ideas and have frank and open talks.
Lesson #4 - just because you have a wall of PhD's does not mean you can brow beat people into submission - your ideas must be credible and work in the real world and stand up to formal peer review and public debate. You will not win contracts from election authorities to do auditing through exclusion of potential critics or competitors. You must demonstrate both that you have a solid team, and also that the ideas and techniques have passed open public peer review. Obviously it looks really bad if you deliberately exclude certain people to avoid such open debate occurring, for whatever reasons or justifications you may pro-offer up after the fact.
Lesson #5 - it not about your own personal Nobel Prize Award stupid! Academics will squabble over who's paper did what in the
field, and who first developed what algorithms. No place for this in a field that demands open frank and fair collaboration. Do your citations, do your attributing, and just get the job done here please. Just check your egos in with the cloakroom attendant at the door; you can collect those later.
Lesson #6 - can we PLEASE get back to solving the real needs in society here? A formal road-map involving combining good practices in building audit records, audit support techniques and then a suite of mathematical tools to operate on and verify the election outcomes are what is needed urgently. Showing how these align with EAC VVSG what steps vendors should be taking also will help build the case. Showing how using open public standards such as OASIS Election Markup Language (EML) can fundamentally empower election auditing is also essential.
Unfortunately instead the Brennan Center and Verified Voting seem to want to instead engage in a "three monkeys" exercise putting out statements like this one.
Hopefully the lessons learned are that they need to be fully inclusive at their future events; that they need an open road map that everyone in the field can objectively and fully contribute to; and then need open public standards and methods to be embraced at the core of that, not just a handpicked inner elite from academia circles.
It seems though that the academics involved are squabbling over the honey pot instead of focusing on the bigger social goals and picture.
The public view is centered on the need to have election auditing that is above reproach and using mathematically sound and agreed methods to give people a high level of confidence in the outcomes.
Of course exacerbating the situation in the US is the counting software and records available for auditing that are woeful at best. Work has begun - such as by the California SOS - to use open public recording standards for publishing election results. In the meantime statisticians have only bare information to work with. Nevertheless good strides have been made in the past couple of years in the area of election audit mathematics.
Enter the Brennan Center and Verified Voting who together sponsored an October event in Minnesota to showcase election auditing techniques and capabilities.
The first hint that this may be problematic for them is that they wished to keep this somewhat exclusive and to pick and choose who came and who presented. Lesson #1 - if you are espousing open public election auditing its a pretty bad idea to restrict attendance to only those people you think should be there.
Next up Lesson #2 - this is a field that demands peer-review and ideas that can be universally verified as sound and in-line with the latest state-of-the-art and consensus around that (see here). It looks really bad if you exclude people who are critical of your central ideas and finding flaws that you appear not to be able to counter or otherwise refute except by heavy handedness.
Lesson #3 - hiding does not get the job done. Using hotel staff to harass would be participants and exploiting arcane local legal bye-laws to circumvent basic free-speech rights looks downright nasty and mean spirited. It also puts you in jeopardy for future events where government attendees may see your prior actions as falling foul of OMB restrictions on their event attendance. In short if your ideas cannot stand up to public scrutiny and fair
open comment there is no way you can have any credibility in this field. It's also not reality in the age of the Internet
and collaborative communities online that demand openness. Using
face-to-face opportunities to drive consensus building is the perfect opportunity - nothing like the lobby of a hotel or
the bar to really allow people to align their ideas and have frank and open talks.
Lesson #4 - just because you have a wall of PhD's does not mean you can brow beat people into submission - your ideas must be credible and work in the real world and stand up to formal peer review and public debate. You will not win contracts from election authorities to do auditing through exclusion of potential critics or competitors. You must demonstrate both that you have a solid team, and also that the ideas and techniques have passed open public peer review. Obviously it looks really bad if you deliberately exclude certain people to avoid such open debate occurring, for whatever reasons or justifications you may pro-offer up after the fact.
Lesson #5 - it not about your own personal Nobel Prize Award stupid! Academics will squabble over who's paper did what in the
field, and who first developed what algorithms. No place for this in a field that demands open frank and fair collaboration. Do your citations, do your attributing, and just get the job done here please. Just check your egos in with the cloakroom attendant at the door; you can collect those later.
Lesson #6 - can we PLEASE get back to solving the real needs in society here? A formal road-map involving combining good practices in building audit records, audit support techniques and then a suite of mathematical tools to operate on and verify the election outcomes are what is needed urgently. Showing how these align with EAC VVSG what steps vendors should be taking also will help build the case. Showing how using open public standards such as OASIS Election Markup Language (EML) can fundamentally empower election auditing is also essential.
Unfortunately instead the Brennan Center and Verified Voting seem to want to instead engage in a "three monkeys" exercise putting out statements like this one.
Hopefully the lessons learned are that they need to be fully inclusive at their future events; that they need an open road map that everyone in the field can objectively and fully contribute to; and then need open public standards and methods to be embraced at the core of that, not just a handpicked inner elite from academia circles.
