by Glenn Langenburg

Dear Colleagues,

Last month, Kasey Wertheim wrote a summary of Minnesota v. Hull, a case in which there was a Frye-Mack hearing for the admissibility of fingerprint evidence (and also forensic document examination). The case has just completed its trial phase and the defendant was found guilty (fairly overwhelming physical evidence and witness statements). In Minnesota there is no death penalty, so cases such as this, where the sentence will be "life with no chance of parole", are considered the most extreme and will automatically be reviewed for appeal. Therefore I will limit my commentary on the case as a whole. The upside to this is that we knew this going in, and we attempted to prepare a very solid record for the superior courts' reviews.

I had been through a similar Frye-Mack experience two years prior in Minnesota v. Columbus. Even though the fingerprint evidence was found to be admissible, I was not personally satisfied with the presentation of the evidence or the strategy of that particular case. In a Frye-Mack hearing you attempt to satisfy the following: the method is generally accepted to be reliable within the relevant scientific community (Frye, Mack prong 1), the procedures were followed in this specific case (Mack 2), and the analyst applying them was qualified to do so (Mack 3). In Columbus, the defense team argued that "yes they followed ACE-V and other ASCLD-LAB/SWGFAST guidelines", but "the ACE-V method has never been demonstrated to be reliable". Simon Cole testified for defense in Columbus, supported that statement, and noted the lack of a validation study to demonstrate reliability. We did not present any data or studies in Columbus. This was the prosecutor's decision to simply stick to the Mack prongs and say "yes they are generally accepted, yes we followed them, and yes the analyst is qualified". Boom, we're done. Of course, I was vigorously cross-examined and I felt that although I had data and useful research to support my position, I could not use it because it had not been disclosed to the defense team. I left that hearing feeling like I had done the best I could, given the prosecutor's strategic decisions. In the meantime, I believed there was research I could do to help address this reliability issue for the next hearing.

Fast forward two years to MN v. Hull and we have the same situation. An initial examiner (Joshua Bergeron), Simon Cole, PhD, for defense, and rebuttal (myself). This time I convinced the prosecutor to take a much more aggressive approach, especially on the heels of the MD v. Rose case. We also introduced another rebuttal witness, Cedric Neumann, PhD, (Univeristy of Lausanne, Switzerland), from the Forensic Science Service in the U.K.

Our strategy was the following:
1) Joshua would testify to the fact he followed ACE-V, SWGFAST, and ASCLD-LAB accredited Minnesota Bureau of Criminal Apprehension (BCA) SOPs. They are generally accepted, meet current standards, and he is qualified to administer these exams.
2) Prosecution having met it's legal burden on Frye and Mack would rest. Defense would make their arguments.
3) We assumed Dr. Cole would say what he had been saying in previous cases (e.g. no validation study, no way to determine if this is reliable, no standard practices, etc.).
4) Rebuttal: My goal was to specifically demonstrate reliability. Here is the research XYZ that demonstrates that ACE-V can be applied reliably. My focus was to present every bit of research where the ground truth was known, and present those results for the court demonstrating, that while errors do occur, they are relatively rare, are conditioned (i.e. affected by training/experience, difficulty of the exam, and other factors), and verification as a quality assurance mechanism can further reduce chances of error.
5) Rebuttal: Dr. Neumann essentially was the heavy stroke for us. Here is a bona fide research scientist and well respected academic/forensic scientist that sees the practical aspects of our decision making but he is also actively working to improve practices through the introduction of tools that will make the process more transparent. His model could also aid in the decision making process by providing the examiner with more information and provide more objective evaluations of corresponding L1D and L2D. And given current practices and technology, fingerprint examiners aren't "doing it wrong", in fact it's quite acceptable, but practices could be improved with new technology and advances in the field. Additionally his research specifically would address a) that ridge details are highly discriminating and the more corresponding features, the less chance there is that someone else could share those features and b) the reason why there is no minimum point standard as an arbitrary decision making threshold (his studies, along with Champod et al (1995)) demonstrate that "not all points are created equal". Ten minutiae in one configuration could have quite a different weight than ten minutiae in a different arrangement.

What has caused a bit of controversy and confusion for some people was the manner in which we expressed these concepts. We certainly had a distinct interpretation of these concepts that some examiners are struggling with, simply because our approach sounds different than how many examiners may have learned to approach them.

Specifically, where we deviated from typically strategies were the following issues:

1) We stayed away from zero error rate. Josh, Cedric, nor myself believe that the "zero methodology/all errors are up to examiner/and you can't hold me accountable for everyone else's error rate" is neither a helpful answer for the court nor an informative one. Forget if it's true or not...it just doesn't help the court (in our opinion). It doesn't answer the issue of "ok, so how often do mistakes occur, just in general, in fingerprint practices?".

2) We did not commit, as Dr. Cole refers to it, the "Examiner Fallacy". This is the statement that "BECAUSE friction ridge skin is unique, it directly follows that when I have found 'enough ridge detail in sequence to individualize' it is that person and could be no one else (and there is zero chance I have made an error since I followed the method)." We took what we believed to be a more defensible position: "We found X features in agreement. The more features you have in agreement, the less likely chance of reproduction in another individual".

3) Therefore, because of #2, we actually did not use the word "individualization" because it is clearly defined by SWGFAST as "determination that two friction ridge impressions originated from the same source to the EXCLUSION OF ALL OTHERS". We used the term "Identified" (which is still allowable by SWGFAST) and interpreted it in the following manner:

"We have identified the latent prints as having originated from the defendant. We did not compare these latent prints to every single human on the planet (past, present or future) therefore we cannot dismiss the theoretical possibility that someone out there may be able to produce a similar latent print. We don't believe that someone has IDENTICAL friction ridge skin to the defendant, but we can't dismiss the possibility that someone may have an area of friction ridge skin that may be similar enough to produce a similar looking latent, and that given the quantity, quality and specificity of the ridge detail, a human examiner applying the ACE-V method may not necessarily be able to exclude it as having come from the other source."

Essentially, we left the door open a crack to say "hey, it's theoretically possible to have a similar enough arrangement of friction ridge details such that I can't exclude everyone" and theoretically I could have made a mistake. At the end of the day, we are saying "These latents were made by the defendant, but there is a remote chance of being wrong.....and this chance will be dependent on the quantity, quality, and specificity of the ridge detail".

Of course, you are thinking, "Well what is this chance???" (which of course, we were asked by defense). And our answer was "At this time, given the present technology, there is no practical way to measure this; however we are currently and actively seeking tools that might answer that question, but we believe, based on the current studies and research (FSS and Univ. of Lausanne) that this chance is relatively small....small enough that it can be dismissed in this case". And it is practical and reasonable to dismiss this chance in order for us to make a decision.

The expression of "individualization...to the exclusion of all others" has been a matter of debate and consternation for me for some time. No doubt in the beginning of my career I said it, and often. I have over time been swayed by what I believe to be very compelling and strong arguments by Cole, Kaye, Stoney, Champod and others. Even going back to the writings of Kirk, Kingston, and Kwan (all from Berkely) we can see discomfort with cavalierly making such powerful claims. Recently in Maryland v. Johnson, a judge would not allow an examiner to make this claim in his courtroom (particularly citing the "exclusion of all others", not to mention then saying 100% certain and zero chance of making an error). And lastly we are already seeing that the National Academy of Science are having issues with this term. I expect they will address this topic in a report on their review of our discipline, due in February 2009.

To be clear, I believe and assert that friction ridge skin is unique. But in the realm of latent prints we are dealing with a different beast. I don't believe we are doing things "wrong" or even making significant number of erroneous ID's. I do believe we may be over-expressing the strength of the evidence and making claims that simply cannot be supported by the existing research.

I am strongly opposed to statements by some critics such as Saks, Cole, Habers, etc. that it does not meet the admissibility "guidelines" or should not be admissible at this time. However, once the evidence is in the courtroom, I disagree sharply with some examiners on how it should be presented to a jury. To me, this is the only real fight that the critics and practitioners should be debating: "What is the best and most appropriate way to present latent print evidence without overstepping the bounds of science, and while still being practical and simple enough for a jury to weigh and make a decision?" This to me is the only real debate and unfortunately we are all (critics and practitioners) spinning our wheels and wasting time and effort on "admissibility" rather than "weight and presentation of the evidence".

I have transcripts from this case that I am willing to share with examiners. But I have to be clear here, I am only willing to share them (for free) with EXAMINERS (for training purposes). Anyone else (commentators, academics, attorneys) that wants a copy, contact the court reporter, Linda Neske at (320) 983-8286. You can buy a copy for about $50. I promised Linda that anyone that would be using these for citations in briefs, articles, or using for any other purpose than training and learning, would need to buy a copy from her. She agreed. Interested parties can email me directly at glenn.langenburg@state.mn.us, or contact the court reporter for your own original set. (PS-I have taken the liberty of correcting a few words and statements in my version of the transcripts, knowing what I said versus what came out on paper. They are minor alterations and don't impact the testimony.)

Cheers all,

g.

Glenn Langenburg