CLPEX Chat Board

The IAI put out a position paper on the 5th that has been making its way around the various forums, fb groups and mailing lists. Ultimately I would argue it says very little and puts the examiner in a bind. I wrote this on another forum, so I'll just paste it here. But I'm interested in what other people think.

While it is nice to see the IAI recognize that conclusions are actually inferences and need qualification, it does little remove the examiner from the rhetorical corner they have been painted into. The 1979/1980 resolutions regarding not being able to testify to probable or possible literally only allowed an examiner to state that their conclusion was not certain nor was it an impossibility that their conclusion was wrong. Anything else is merely an appeal to probability or possibility no matter how implicit.

The position of the IAI then shifts to a seemingly impossible task. Avoid overstatement of decisions by qualifying them without giving standards for qualification. It merely reintroduces the possible/probable conundrum without actually giving advice how to navigate it while adding an admonition that overstating is wrong. In my mind, that really amounts to zero guidance.

I would add to this that this problem is not new to our discipline and is categorized under words of estimative probability. I would add that we need better training around the concepts behind the shifts in thinking that are occurring here, not just new words and to that end I've created a document that attempts to do that. If anyone wants to rip it, add to it, or do their own, feel free.

Hello Boyd,

Firstly thank you for taking over the running of this website, it is a very important forum and has proven its worth many times and is still a credit to Kasey for having created its existence. As you said this subject has been done before and until the cows come home it will keep repeating like an unwanted hiccup. I nutted out a few things to add to your thread from an Australian perspective and when I saw the list I thought just for fun I would tackle the order in reverse rather than the logical sequence. I haven't read your paper but I assume it is very logical and well thought through as all of your previous work has been to a very high standard. Conclusions deal with the principle of uniqueness in fingerprint patterns of detail, firstly we could disagree that the principle is sound and scientifically true. We accept that it is true, proven by lots of research and experience. To reach a conclusion means we have conjured up lots of confidence in our result, enough to state our reputation before a court of law. To prove we have this confidence we need to be able to show records of our examination methodology and to articulate the correct hypothesis that led to the final result. The oral presentation must deliver both objective and subjective data and details using a clear and easily understood language. The underpinning knowledge and experience of the examiner should be qualified by a body of study and examinations that are at academic standard. Sciences of biology, human anatomy and physiology, mathematics, computing and chemistry are some of the pillars that support fingerprint science. To supplement that knowledge there must be continual improvement on the part of the individual expert as a highly motivated researcher challenging all aspects of fingerprint science and its principles. Finally and this was my first listed point there is way too much dogma still embedded in fingerprint science, very hard to move. Cheers from oz Les

Hello Boyd, I will stay on this thread but I am aware of the later threads dealing with much the same issue of credibility in both fingerprint science and the fingerprint expert. Although retired in 2012 I left behind a body of 10 years research study that covers some of the issues under these discussions. The work was generally titled Golden Hands and Silver Feet, quite a few CD's were mailed to the US. If anyone Googles Les Bush and Fingerprints one of my articles is still there for reviewing. I agree fully with your comment Boyd that more words will not solve the problem of any language using a lot of dogma and little science. If you Google how fingerprints are formed the California university site speaks volumes as to how far removed the general community is from good science. So how do we catch the thief that is trying to steal the credibility of fingerprint science as not being a Gold Standard? I will keep this short since more words are not the answer. Firstly experts need to be able to explain how fingerprints are formed and what combination of factors make each fingerprint unique. Then experts need to be able to explain how the geometry of a pattern of detail and the combination of detail types are collectively analysed and progressively incorporated into an evaluation of a hypothesis. Culminating in the matching of a patterns spatial arrangement and using a reliable predictability test to confirm the experts confidence. For those with time please look into the work of Christophe Champod and his many researchers who have laboured hard and long to come up with a stats model that can account for the limitless combinations of pattern details in their spatial configurations. I am not a critic of Christophe's work I applaud the many advances that were made, I would wish that at sometime he actually completed expert training in the identification of latent fingerprints by the tried and true method we all know works. Remember the lengths that Simon Cole went to during your countries many years of Daubert! The science is fundamentally okay, it just needs more advancement in science and having a general understanding of how to explain its workings and principles. It badly needs good peer review both scientific and pragmatic. Cheers Les

Hi Les,

Thanks for the thoughtful replies. I guess the general concern that is present for me anyway is attempting to hold up probability and more specifically Bayesianism as a sort of scientific rationality while not addressing some of the pre-existing conditions, even if addressing them just means talking about them not necessarily solving the problems. If I had to sum up my current vantage point there seems to be some tension that exists within the PCAST findings, the NCSF Draft for Comment and the IAI's position paper, that we're uncertain how to deal with uncertainty while at the same time advancing specific models in casework. All with no discussion on how to deal with errors that will inevitably occur. There will be a statistical equivalent of Mayfield, and prior to that there will be degrees of optimization that will occur that will change outcomes. Why wait for these events to occur before discussing how examiners might be affected and potential courses of action. After all, if these are policy makers, shouldn't they be thinking about these things prior to them happening?

Boyd

Boyd some quick thoughts for those policy makers. Fingerprints has been functioning for a long time both in Ten set and Latent examinations. To date no two persons have been shown to have the same prints. The argument from others is that not all persons have been printed. Their fear is that in Probability Theory the Law of Large Numbers already means fingerprint science has been performing extremely well and continues to do so even under the watchful eye of AFIS technology. The onus for introducing the probability of uncertainty in fingerprint examinations is not on the side of fingerprint science. Fingerprints by its inherent biology are both permanent and unique. There is no uncertainty about Ten print examinations and neither is there uncertainty about a clear and well defined latent fingerprint. Those levels of detail available for evaluation make it clear that any error in a conclusion of identity would be as a result of human failure to follow correct methodology. So the uncertainty argument comes down to the weak latent fingerprint which requires more from the expert to complete the examination. Remembering that the origin of the weak latent is, for this example, one of ten digits on the hand of the human. The skin of this one finger tip is an area around 1cm by 1cm and contains a constellation of large and small details which may include pattern shape, ridge flow, central features and ridge details. Every constellation has values in both its geometry and morphology. Our science eliminates the result on the basis of exclusion or insufficient detail, or confirms it by capture of the details present. In a controlled repeat of the whole process, latent deposit, development and examination would see the result confirmed. So three things exist for the science to remain 'uncertainty' free, the biology of friction skin, the complexity of arrangement of details and the standard applied to identify. As Paul Kirk once said only human error in failing to find the evidence stands in the way of the success of forensic science. The policy makers need to look at what is before them before deciding the science is failing. Cheers from oz Les

This was funny to me so I thought it relevant to this thread, thank God for fingers it saved the day for maths!
A high school teacher was arrested today at Toronto Pearson Airport as he attempted to board a flight while in possession of a ruler, a protractor, a compass, a slide-rule, and a calculator.
At a press conference, Premier Kathleen Wynne said she believes the man is a member of the notorious extremist Al-Gebra movement. She did not identify the man, who has been charged by the RCMP with carrying weapons of math instruction.
'Al-Gebra is a problem for us', the Premier said. 'They derive solutions by means and extremes, and sometimes go off on tangents in search of absolute values. They use secret code-names like "X" and "Y" and refer to themselves as "unknowns," but we have determined that they belong to a common denominator of the axis of medieval with coordinates in every country. As the Greek Philosopher Isosceles used to say, "There are three sides to every triangle."'
When asked to comment on the arrest, Prime Minister Trudeau said, "If God had wanted us to have better weapons of math instruction, He would have given us more fingers and toes."