Dave, thanks and you have made some good points that I would like to address. First let’s you and I keep this discussion friendly and if I step over the line (or appear to) then by all means kick me back into play. Second, on these type of discussions I believe most of the differences are actual differences in beliefs.
It should be noted that my research with Dr Dror is about the study of psychology and the human mind. As such, it is the academic question that needs to be answered. If we accept that we are all human, then the frailties of the mind apply to us all, expert or not.
(1) Yes we are all human but your statement seems to me to imply an equality of cognition, and therefore ability. Would the following statement be out of line with psychology principles to say “We are all human, and the frailties of the mind apply to all human beings but not equally”. If all of us are human, and the frailties of the mind are equally shared then where do Mozarts and Shakespeares come from. Where does that inner action come from that separates a hero from a coward. I do not believe that the same frailties of the mind come about equally among all human beings.
(2) If the question is an academic one, then the study using non-fingerprint people to show bias in fingerprints from my POV should have stayed within the academic world where its content could be truly appreciated. By submitting it to the fingerprint discipline it is my belief that it did more harm than good.
This was done, twice, with more subtle variables in the 2nd study using experts, but nevertheless the effects were noticed and replicated.
Correct me if I am wrong but Type II errors was replicated not Type I. Other than the first study Type I errors have never been made from intentional biasing study?
I tend to argue that the fingerprint community needs to embrace challenge and engage academia.
Sounds good, but which academics? Are they selected by themselves, or does the government select, and most importantly does the fingerprint community have a say in which academics are chosen? I think either way you go there will be no consensus.
It is my assertion that bias is a factor in all our lives...because we are human beings....we cannot help it. However, I also hypothesize that training and understanding of such issues can greatly mitigate the more obvious effects. I contest the notion that applied methodology and good training can always be a major issue in offsetting contextual error...just ask the FBI in Mayfield…
(1) Here we have a belief=assertion hurdle. Do we as human being have biases? Yes! Do the biases pervade our lives as much as some say they do? My current belief is no. There are over 200 catalogued biases in the cognitive sciences and they even have one for cognitive studies that are biased. If we have as many biases as some say we do, how can one even function in today’s world?
(2) In discussions such as this, Mayfield is often brought up and as well it should. It does represent itself towards such discussions whether bias was an actual issue or not. But on a different note, it is one out of 79 years of history. Now someone is going to tell me that it is the tip of the iceberg that if their is one their is another, just buried is all. It is a very nice theory, but like all conspiracy theories it does not require a shred of proof just an allegation. Yet very learned and esteemed men and women will say just that. What true scientist or academician would offer a statement without a shred of proof? My main point here is how often has it occurred, to what extent and in what context?
Using ACE-V do we look automatically for features that match in a print and latent? Or do we look for obvious anomalies? If the former is true, at what point do I reach that tipping point, that escalation of commitment beyond which there is no going back? Even if I subsequently find anomalous features. By the way, this could be the case in negative calls too...so maybe we miss matches too!! Ths of course is only highlighted in the verification phase.....unless there is total blind verification?
I look for all information for and against. It is a total assessment of all information shared and not shared between the two images. In regards to Type II errors I think that subject is broached further down. I do not believe that type I and type II errors should have the same weight. This is a human endeavor and not a mechanical one. In reference to blind verification I will submit the last issue of the Chesapeake Bay Division has an article on the topic and which I happen to agree with (mostly). I will refer you to that article on blind verification.
Bias and other psychological effects are a human trait that are generic across many disciplines. Apples and bananas re both fruit are they not ;.))
Are you saying that bias is also present in other disciplines? Yes they are. Does eyewitness identification have a propensity for biases? I would have to say yes. And they also both occur with human beings and happen on the planet Earth. That does not make them the same. Yes they are fruit and I will go you a couple better. They grow on trees, consumed by human beings, and evolved on planet earth. But they are a different shape, color, skin size, texture, taste, and you cannot cross pollinate them and get an Appanana. In conducting an experiment and you are looking at 50 issues and three are in agreement and 47 are not would you say your experiment is a success because you had 3 in agreement. I think not.
But I think it is true to say that we as a science have sat on our collective back sides for many generations on the assumption that our work was beyond reproach. We would be collectively be better placed now to offset Daubert etc etc had we engaged in long term research over many many years to look at error rate and other issues.
As I said error rate is a new concept. If I could use a time machine and go back and warn them I would. If I could use a crystal ball to determine what new concepts will be presented in the future I would so we could prepare now. But I cannot. I can no more predict what the future issues may be 75 years from now than you or anyone else can. Should those in the future condemn us in their past for issues that are present in their world but which are unknown to us now? As more of a historian than even being close to a psychologist, it is a common complaint from historians how some in our current society talk about the past under today’s knowledge. Kind of like talking bad about the Aztecs because they did not use the wheel.
Absolutely!! There lies the problem. 35 people turn up to a meeting and group dynamics, personalities and politics conspire to a trend toward the lowest common denominator.
Is that not the same problem with Academia? Do we throw out one bunch with group dynamics, personalities and politics for a different bunch with group dynamics, personalities and politics? I do not care where you go or who you align yourself with there will always be group dynamics, personalities and politics. You would be hard put to convince me that academia/science in general have no politics or group dynamics. History would not agree with you.
What we need is for SWGFAST to stop working toward common consensus, and move toward envelope pushing ideas and research.
Here we have a belief hurdle. Their mandate is standards. To achieve those standards they need a common consensus. Ideas are like anything else human. Good ones and bad ones. If you implement a bad idea it is even harder to get rid of. Second it confuses a lot of people. I personally think we have enough bad ideas now. To set good solid standards one needs to proceed cautiously and with deep consideration.
Scientists of the world many centuries ago held the view that the earth was flat. That was the consensus. It took those with vision and bravery to counter this argument. This is what fingerprints needs now...not just in my research, but that of Glenn's, John Vanderkolk's, Christophe Champod's etc etc etc
The change of view that the earth was not flat did not happen over night it occurred over a period of time with more and more converging evidence. As stated before I have no problem with good research. I welcome and read with flavor the works of Glenn, John and Christophe.
Well, maybe in some cases. But maybe such people need bigger guns to break down the thick walls of dynamic inertia ;.))
Perhaps some people built thicker walls because they honestly believed that the rocks being thrown were false, allegations without merit, conspiracies without proof or just being thrown to fulfill a personal agenda. Would not most people or disciplines build walls if they honestly felt they were being attacked falsely? Most of the people I know are pretty reasonable. If you want them to understand something you need to show them in the best way possible not just tell them to take your word for it. I have yet to see beliefs change by throwing rocks at them. It does nothing but create a defensive posture (what some have called the fingerprint community being a closed group)
I would simply ask if we assume that ACE-V is applied correctly then any technical error is valid to be put in the same category? If I miss many hits that should have been called then there is surely a chance that my lack of technical skill will result in me wrongly calling a match that sends a person to jail for life?
(1) This is a big belief hurdle and I doubt I have the communication skills to do it justice but I am going to try. There are several ifs in your statement “if we assume ACE-V….any technical error is valid…same category?” Why must they be put in the same category? I believe they should not and my reasons are that they are not equal, they have never been equal in fingerprint analysis, being considered separately has been a standard for long time, they do not occur under the same level of consideration, they are separate because of different experience levels, different ability levels, it puts an enormous stress on young examiners to make calls they are not comfortable with. What would happen if a senior examiner says yes and a junior examiner says there is not enough for them? Do you terminate or send back for retraining. Putting type I and II errors under the same umbrella I believe would be one of the most harmful concepts. I believe in fact it would do just the very thing you are seeking to prevent—bad calls. I have often wondered why the critics would want to consider type I and type II errors together. Perhaps you could answer that---why?
(2) “If I miss many hits……lack technical skill….calling a wrong match..” (A) that is what peer review is for. If that situation is determined then you take that person off bench work for re-training or if it continues a decision to terminate. But if you have BV then there is no peer. (B) You have a big assumption there (a leap of faith is what it is called), but that is a admin and training issue. Again PR and the correct SOP and standards will correct that, not BV.
Much more research is required and should be conducted!! But the community must be open to such research and act whatever the findings may be.
I agree much more research. I agree the community must be open. However I am stuck on the last one “…act whatever the findings may be.” I would think a better discipline would want to look at the research, how it was formed, what preconceptions are in place, what bias may be present, is the research reasonable and does it follow scientific principals? You are not asking me to believe that all research that comes from academia, science, and from esteemed and learned individuals is absolutely true and correct 100% of the time. You are asking for quite a leap of faith on that one.
Dave I want to thank you. You have enlightened me with you beliefs and convictions and I hope I have not stepped over the line in my response.
For what it is worth I have always enjoyed your comments on this board and while you were editor of Fingerprint Whorld. I think you did a great job. I hope to meet you some time.
Respectfully
