IAI Conference in Reno

[Dr. Borracho]

Not being in attendance, I am curious if there are any interesting new topics or hot-button issues being discussed at the IAI this week.

[SConner]

Once again, on the last day of the conference there will be a business meeting that unless you're able to attend you don't get to participate in. I'm personally annoyed at how the IAI conducts some of their "business" and limits voting on issues and electing officers to those who choose to attend the business meeting. The organization has something around 6000 members, you get around 1500 attendees at a conference, and you get maybe 300 or so at the business meetings I've attended. So around 5% of the membership is making decisions and electing officers. Somebody challenge my math, tell me off, I'll accept it.

My point is, every company that I own stock in somehow manages to send out a proxy vote packet for major corporate issues and elections for the Board of Directors. If I don't vote, that's on me, but at least the company tried to include me. The IAI doesn't seem to want to even attempt to include members that can't physically attend the business meeting to be able to participate. There's technology out there and available to make this happen.

Start trying to include more members in the decision making process of what is supposedly OUR organization.

[Dr. Borracho]

Good point. One IAI Division to which I belong sends out an email to all eligible members. A link takes you to written position statements from each candidate for each office. Then you cast your votes online. Seems like a lot more "democratic" system for managing the organization.

[NRivera]

Business meetings have always been on the Friday of the conference, but I would support a more inclusive approach to voting for officers as well.

I get that we want as many officers as possible to actually be in attendance at the conference, but I don't see why it should be a requirement to run when it is entirely reasonable for candidates to communicate with members electronically.

[josher89]

Heidi Eldridge presented on her black box study on palm prints. I can't remember the exact numbers (I haven't looked at the site yet) but they were similar to the FBI/Noblis for FNR and FPR.

[timbo]

0.67% false positive rate; 9.5% false negative rate

[josher89]

Once again, on the last day of the conference there will be a business meeting that unless you're able to attend you don't get to participate in. I'm personally annoyed at how the IAI conducts some of their "business" and limits voting on issues and electing officers to those who choose to attend the business meeting. The organization has something around 6000 members, you get around 1500 attendees at a conference, and you get maybe 300 or so at the business meetings I've attended. So around 5% of the membership is making decisions and electing officers. Somebody challenge my math, tell me off, I'll accept it.

My point is, every company that I own stock in somehow manages to send out a proxy vote packet for major corporate issues and elections for the Board of Directors. If I don't vote, that's on me, but at least the company tried to include me. The IAI doesn't seem to want to even attempt to include members that can't physically attend the business meeting to be able to participate. There's technology out there and available to make this happen.

Start trying to include more members in the decision making process of what is supposedly OUR organization.

Since they allow for nominations from the floor during the business meeting, there's no way they can hold a vote of the entire membership for those offices.

A change to the bylaws would correct this and allow for all members to vote but the issue would be ensuring that the votes were from actual voting members and not random people.

If you can come up with a solution that would address those issues, I'm sure the IAI Board of Directors would at least listen to it. Right now, though, I see it as a tall hurdle to overcome to ensure accurate counting of votes.

My division, on the other hand, requires those wanting to run for office to get vetted prior to the vote. That keeps someone from nominating their coworker, who isn't at the conference, to a position that they may not want or are able to hold.

[josher89]

0.67% false positive rate; 9.5% false negative rate

If you participated in the study, get your user # ready and click on the following link:

https://cchampod.shinyapps.io/Results_BBStudy/

Use the side options to find your number and see how you did. Or, look at all of the data and sort appropriately.

[SConner]

Since they allow for nominations from the floor during the business meeting, there's no way they can hold a vote of the entire membership for those offices.

A change to the bylaws would correct this and allow for all members to vote but the issue would be ensuring that the votes were from actual voting members and not random people.

If you can come up with a solution that would address those issues, I'm sure the IAI Board of Directors would at least listen to it. Right now, though, I see it as a tall hurdle to overcome to ensure accurate counting of votes.

My division, on the other hand, requires those wanting to run for office to get vetted prior to the vote. That keeps someone from nominating their coworker, who isn't at the conference, to a position that they may not want or are able to hold.

They could allow nominations from the floor at the business meeting and then hold voting for a period after the conference closes, it's not that difficult. The business meeting could be live streamed or made available via conference call, that's what major corporations do for their stockholders.

I remember one particular business meeting I attended, I think it was in Providence where the issue of the year was the position statement on likelihood ratio software, if I'm remembering correctly. I seem to remember the voice vote being taken and it was close enough that they went to hand raising. The hand raising was still too close to call so they resorted to having to hand count the vote. At the same time that this vote was occurring, several members that had an interest in the outcome of the vote were observing the process of counting ballots for Officers and Board of Director positions, so they were being obviously excluded from participating in that vote because the attitude seemed to be the business meeting must go on! No, you need to wait and make sure that those people serving as proctors for the vote counting get an opportunity to vote on this issue also, especially when the vote was so close. And more importantly, why was my vote as just a crime scene person even being considered when the discipline affected by that decision was the latent print community? Why would the vote of a bloodstain pattern person, a forensic art person, or any other discipline other than latent prints be invited to vote on what should be their topic alone to decide?

[Dr. Borracho]

And more importantly, why was my vote as just a crime scene person even being considered when the discipline affected by that decision was the latent print community? Why would the vote of a bloodstain pattern person, a forensic art person, or any other discipline other than latent prints be invited to vote on what should be their topic alone to decide?

I am rusty on my understanding of the internal development of policy decisions in the IAI, but I believe a policy decision such as that on statistical evaluation of a latent print identification has to begin in the appropriate subcommittee, then go before the whole body for final approval. Fine tuning the policy would be done by the relevant community within the IAI, but the vote of the whole body supposedly reduces the chance of internal conflict, say, between the AFIS/10-print folks and the latent print examiners. Conceivably, even Crime Scene Analysts may have overlapping concerns with a policy they write, when on review by Latent Print Examiners or AFIS specialists, hypothetically, a new development policy that makes the CSAs job faster and easier doesn't produce crisp, high contrast latents for AFIS or comparison purposes.

But to your point of close call votes being decided by a only a handful of votes, when only a very small percentage of members are present and voting, I couldn't agree more. In today's electronic environment, a secure system should be easy to devise that would allow all members in good standing to vote electronically on candidates for office, new policy positions, changes to the Constitution and Bylaws, or any other matter requiring a vote. Let the conference attendees vote on all the traditional commendations for excellent service and contributions to the conference itself, but put votes of interest to all members before all members for balloting.

[josher89]

If the business meeting is Friday morning, and the installation of officers is Friday evening, I think that opening up the vote to the entire membership may still prove somewhat difficult given that it would have to be done in a day (considering international members may have to vote during their sleep cycle or whatever). I'm all for including the entirety of the membership to vote on those matters but I still think it's an extremely challenging endeavor to accomplish.

If there is some sort of electronic voting process that would ensure that members are voting, I think the IAI would have to take a look but until then, this is probably the best we'll get.

I've helped count ballots a few times now and realized that while I may be missing out on some portion of the meeting, I was willing to overlook that to help count ballots. It's a choice I made to participate. They usually ask for first-timers and that might be the last group that should be asked to help so they can actually see how the meeting is conducted.

When it comes to close votes, there are those members who are only there to vote and then leave promptly after the ballots are collected. The issue with that is if there's a close vote between three or more people and a new ballot must be cast to decide a final candidate. We would now have less people voting on the more important vote than the one before. Again, no easy way to do it to be fair to all.

[Boyd Baumgartner]

If there is some sort of electronic voting process that would ensure that members are voting, I think the IAI would have to take a look but until then, this is probably the best we'll get.

I'm a cryptocurrency nerd and wouldn't you know, there's a token just about to launch for that purpose. Give every active member a wallet with a percentage of dues being translated into tokens and let the members vote on the issues that matter to them. This solution would also act as a disincentive to the good ol' boy network that move to table important items during the meetings as well.


youtu.be/4jqPODCSyLo

[josher89]

This solution would also act as a disincentive to the good ol' boy network that move to table important items during the meetings as well.

And therein lies the rub...

[Dr. Borracho]

This solution would also act as a disincentive to the good ol' boy network that move to table important items during the meetings as well.

And therein lies the rub...

Ah, yes. And then there's that.

[ER]

Back to the Palm Black Box (Which is AWESOME, because you can download all the raw data right now. I've been in heaven.)

It's better to look at False Discovery Rates in our context then FPR and FNR. Essentially, it's the difference between these two questions:

"When an examiner is comparing different source prints, how often was an Erroneous ID made?"
vs.
"When an examiner reaches an ID decision, how often was an Erroneous ID made?"

When presenting in court, the second question better answers the situation. The court knows that I reached an ID, so how many of those are wrong in a research study. Since the court can't know when I come in with a different source comparison, the FPR doesn't really fit.

Anyway...

The False Positive Discovery Rate for palm black box is 0.2% (exact same as fingers)
The False Negative Discovery Rate for palm black box is about 24% (vs. 14% for fingers, when considering Vcomp and all 611 errors)

However, Discovery Rates can skew based on the prevalence of samples (ground truth same vs. different source). The palm study had a higher prevalence of same source than the finger black box did. This would tend to skew the FNDR higher, so it's probably a combo of palms being harder, larger area to consider, orientation not being known, and the prevalence difference.

Anyway again... I'm super excited about the whole study and have loved the chance to look at all the numbers. I can't wait to read the final paper(s) and see what else can be gleaned from even more info that isn't in the basic Confusion Matrix data that's available now.