Fingerprint Dogma Final Exam

[Dogma (formerly Guest)]

Is it 100% error free? We doubt that. What is the error rate -- nobody knows, yet.

First, welcome to the debate, Ms. Steele.

The idividualization of a latent print to a known source can be 100% error free as a methodology or process. In the cases you cite, it was not fingerprint identification that was wrong. It was its application and the conclusions of the examiners. In the Cowans case, it was a matter of fraud by the initial examiner and possibly "confirmation bias" on the part of the verifier.

My earlier questions concerning the establishment of values for our oft repeated concepts of quality, clarity, quantity and sufficiency go to the heart of the dilema. I believe that our problem is one of "threshold." A car traveling down the road at 40 mph is 100% stable. A car speeding down the road at 120 mph (speaking of an average vehicle and roadway surface) is unstable. Some drivers can handle it, most probably can not. If the speeding car crashes, is it the fault of the car or the driver? Were the limits of the car surpassed or was it the ability of the driver to control the car at fast speeds? Or, a bit of both?

I believe that one of the problems with fingerprint examiners today is that when we philosophically changed from point counting to ridgeology, there wasn't the re-education to go with it. The IAI came out with its "no minimum" resolution in 1973. Seven years later, when I received my FBI fingerprint training, we were still counting points and there was no mention of level 3 detail.

AFIS has contributed to the problem, offering prints that are more and more similar than we were used to in the pre-AFIS environment. It also adds to the "confirmation bias."

The focus of the challenge should be on the ability of the examiner. The way the argument is currently framed will make us jump through many hoops just to end up back where we started.

BTW, thank you for participating in the panel discussion in VT. Unfortunately, I came away unsated. I didn't hear anything that would move us past the current quagmire.

[L.J. Steele]

The idividualization of a latent print to a known source can be 100% error free as a methodology or process.

This is one of those tricky areas. If you were at VT, you probably saw Cole and Wertheim debating this very point. I have some philosophical trouble with the idea of an error-free method that fundamentally depends on a human decision about, for example, what is an explainable distortion/discrepancy/difference due to some factor in the surface, matrix, etc. and what is an inexplicable distortion/discrepancy/difference that means there's no match.

There are likely some tests that are 100% reliable. Is that blood A+ or O-? Does that object meet the standards for size, weight, etc., or not? Unless one does the test wrong, the answer is always going to be the same.

Then we get into tests that can have false results no matter how well they are done -- take Pat W's example, does that blood sample test positive for the markers for prostate cancer or not? As he pointed out, the test has a known false positive and false negative rate -- presumably there are some things that aren't prostate cancer that have the same markers (thus a false positive) and/or types of cancer that are under the threshold of detection or don't produce the right markers (false negative). With enough statistics, one figure out the chance the test itself will be misleading.

Then we get to things like fingerprints, and a number of the other trace evidence fields were the examiner's judgment and experience are fundamentally part of the decision making. Where human judgement is part of the process, I'm not sure it is meaningful to seperate practioner error from human error.

To take this to rhetorical excess -- I'm sure an astrologer or alchemist could claim that his or her methods are 100% accure. Any failure to replicate the results is human error -- the skeptic isn't doing the test right. (No, I'm not claiming that fingerprints are the same as astrology, just that there's a problem with trying to focus just on practitioner error.)

In the Cowans case, it was a matter of fraud by the initial examiner and possibly "confirmation bias" on the part of the verifier.

As an aside, I would really like to see a published accounting of the Cowans case, with the underlying trial evidence published. I've seen one of the official reports and talked to a couple of examiners who saw the Cowans evidence (all of whom have slightly different opinions). The Boston AG's office, DA's office, and Boston PD have given me the run-around in trying to track down what was shown to the Cowans jury.

The IAI came out with its "no minimum" resolution in 1973.

This one has always confused me. I assume that no examiner would testify to a match with a single Galton point. I'm not sure I've ever seen a case with less than 8 points (either all from the same latent or on allegedly simultaneous latents). Is there really no minimum? Would someone testify with only 1 or 2 points?

The focus of the challenge should be on the ability of the examiner.

But what of the Stacey report and the Mayfield error? The three FBI guys were well qualified and experienced. Moses likewise. In the abstract, I doubt anyone would have questioned their ability. If four experienced folks make the same mistake, is that really mere human error?

The Cowans error wasn't just the examiner and the verifier -- two unnamed defense experts concurred. Again, four seperate human errors, or a problem with the method?

Cole's paper has a few other cases with senior, experienced examiners making mistakes, which were often not caught by defense counsel. Mostly, the errors seem to involve folks who's training and credentials look good.

I'd tend think of the human factors (confirmation bias, tunnel vision, pressure from high profile cases) as both a procedural error problem and a practitioner error problem. A department's views on dissent could likewise be either or both -- it goes beyond one individual and affects everyone in that lab

BTW, thank you for participating in the panel discussion in VT. Unfortunately, I came away unsated. I didn't hear anything that would move us past the current quagmire.

Quite welcome. I have hopes that the NIJ validation studies will help. I suspect that protocols and guidelines that specifically try to exclude biasing information and insulate examiners from pressure might help. I too am dissatisfied with the situation -- a Daubert hearing is a terrible way to resolve a scientific debate. We've got experts who's testimony is filtered through the direct/cross-exam format and limited by the attorneys' understanding of the terminology and state of the art trying to educate a judge who's equally unfamiliar with the area, about some very technical issues.

I wonder if it would be possible to set up something like the Justice Department's Technical Working Group that set up the Guide on Eyewitness ID. That got the critics and supporters, defense counsel, prosecutors, and police into the same room to hash out the issues and come up with a concensus guide to improve the procedures. Would it make sense to try to get the best-n-brightest on both sides to talk and work something out about the print issues? (My impression is that SWGFAST doesn't include folks like Cole, Starrs, Dror, etc.)

[Heidi Fraser]

I am so glad to see that this debate has returned from the land of childish name-calling into a reasoned, mature discussion. As a trained scientist and a Forensic Scientist in other disciplines who has only recently joined the ranks of the Latent Print Examiners, I am personally very interested in learning the ins and outs of how we practice and defend comparative analyses as a science. I think this sort of debate and discussion is beneficial to all of us, both as individuals who will have to explain this in court, and as a discipline that (more than a hundred years after its inception) is still struggling to define itself. The more we discuss, debate, and research, the stronger we become.

That being said, I always enjoy your posts, Ms. Steele, as they give a great perspective and thoughtful insight into the defence perspective and I believe help us to understand one another, without being pejorative or closed-minded.

I have responses to a couple of your comments below that I hope will present a new perpective.

The IAI came out with its "no minimum" resolution in 1973.

This one has always confused me. I assume that no examiner would testify to a match with a single Galton point. I'm not sure I've ever seen a case with less than 8 points (either all from the same latent or on allegedly simultaneous latents). Is there really no minimum? Would someone testify with only 1 or 2 points?

My understanding of this concept (and as I said, I am fairly new to the world of Latent Prints) is not that there is truly no minimum amount of detail needed, but rather that it is not a constant. There is no scientific basis for specifying a minimum number of points for an individualization, because that minimum amount of necessary detail changes on a case-by-case basis.

Consider for a moment two latent prints. One is nearly complete and has 7 or 8 or even 9 clear points of minutiae. I believe most examiners would be quite comfortable individualizing that print. However, the second print is a very small partial print in which perhaps only 3 or 4 or 5 points of minutiae are visible. But let's say that this small partial print also has very clear pores, unique and clear edgeoscopy, several incipient ridges, and some funky little twists to it that are very distinctive. The sort of detail that (while not a level 2 "point") you would recognize again anywhere. I think most examiners would also be comfortable individualizing this print. True, there were fewer "points", but there was a wealth of other information that was very distinct and very clearly unique. Is there a scientific basis for saying that only the first print is sufficient for individualization? Of course not. However, if we had a very distorted print in which 4 or 5 points could be seen, would we be as comfortable in definitely indiviudalizing that print? Probably not. You see, you have to take the entirety of the impression into account. It's very contextual and situational. Five points that might be very distinct and well within the analyst's comfort level in one case might be entirely unacceptable in another case. This is exactly why the training to become a latent print examiner takes 2 to 3 YEARS whereas the training to become, say, a controlled substances analyst (my last discipline) takes 2 to 3 MONTHS. One is very cut-and-dry whereas the other takes "developing an eye". There is a lot to be said for the experience and training gained from seeing hundreds of thousands of prints and learning what is within your comfort zone. So... I guess this is a really long-winded way of saying, "Yes, there must be a minimum amount of detail present, but there is no way of determining the SAME minimum that would be appropriate for every case".

The focus of the challenge should be on the ability of the examiner.

But what of the Stacey report and the Mayfield error? The three FBI guys were well qualified and experienced. Moses likewise. In the abstract, I doubt anyone would have questioned their ability. If four experienced folks make the same mistake, is that really mere human error?

I agree with you that the fact that most of these high-profile errors seem to have been made by highly respected and qualified individuals is very disturbing. However, I find your use of the word "ability" to be very liberating and insightful. The fact is, "ability" is made up of a number of variables. Yes, training and experience are important parts of "ability". But "ability" is also affected by a number of other factors, which are highly variable. These can include stress, pressure, and even fatigue. My ability to perform a task one day may be very different from my ability to perform the same task another day.

Let's say that my job was performing open heart surgery, not comparing fingerprints. Would I be as confident in my "ability" to perform that job to the same level on a day when I was overfatigued, hung over, being supervised closely by a rigorous boss whom I feared, or expected to perform the entire surgery in under ten minutes? Of course not! All of these situations will be detrimental to my cool nerve, steady hands, etc. I believe the same can be said of comparisons. These analysts had a great deal of experience and training, but who can say what their ability level was at the moment during which they performed those comparisons? I know (from hearing him speak at a conference) that Ken Moses compared the Mayfield print under great time constraint. Would he have made the same mistake if he had been given the time necessary to do a complete and thorough analysis and comparison? Possibly not. We'll never know.

My point is, we don't want to jump to conclusions and state that if a person is truly qualified they can never make a mistake and therefore the system is flawed. Ability is variable and even the most experienced analyst can make a bad comparison on a bad day under the wrong circumstances.

Thanks again for your great comments,
Heidi

[Michele]

Lisa,
People have made identifications without any Galton points. In the JFI Vol. 53, Robert Reneau published an identification with zero Galton points. The ident was based on incipient information. There was so much corresponding information that anyone would be hard pressed to disagree with the ident.

The erroneous identifications stated above have one common thread, the weight of the conclusion was based on the training, credentials, and notoriety of the examiners involved and not on the supporting information that lead the examiner to their conclusion. In science, the position of the practitioner shouldn’t carry the weight, only the available data. Is this a problem with the method or a problem of people to understand the method?

Michele
michele.triplett@metrokc.gov

[Guest]

Van julle is besig om die dam onder die eend uit te ruk !

[Les Bush]

When I first responded I thought the cheese was a little off. The one truth coming out of this thread is that we have a general consensus that this forum is about discussion and not research. Once again I invite the author to read our texts, study the experiments we perform and others, then get back here with any specific problems for discussion. Its a bit like the feather in the wind, gravity is suspended for a while until the planets are once again aligned. I wish you well on your journey of discovery since it appears your desire to learn is highly motivated.

[Guest]

In science, the position of the practitioner shouldn’t carry the weight, only the available data. Is this a problem with the method or a problem of people to understand the method?

I hope you do not really believe this. Your description of science as a one way street illuminates your lack of understanding.

Science is in fact a two way street that relies upon both deductive and inductive logic. It is not just about the available data but how we use that data to come to a conclusion.

If the credentials of the examiner did not matter we could get Pee Wee Herman to make all the ID's in the world, and folks like you could spend your time writing great, albeit misguided, theses about scientific philosophy.

[Shaheen]

Dear Guest & Michele

Guest:

Yes, we are always busy...

I would put to you that, to coin a phrase,

Your description... illuminates your lack of understanding

As an aside, I don't appreciate the denigrating nature of some of your comments. May I make a friendly suggestion that others would be more willing to help you with your research / participate in your debate if you curbed that temptation in the future. I personally am happy to stick my neck out and post a response as part of a debate even if I am later proved wrong. But that changes quite quickly if someone equates my error to my ability as a person. Understand?

Allow me to respond to the topic at hand ...

Michele wrote:

In science, the position of the practitioner shouldn’t carry the weight, only the available data. Is this a problem with the method or a problem of people to understand the method?

"weight" is a legal term that signifies that the proof for one hypothesis or "side" is greater than for another. Michele is right that the position of the practitioner should not be used as proof that the evidence supports one hypothesis. The original Daubert case (See Daubert v Merril Dow Pharmaceuticals) identified this as a problem, and the guidelines provided during the case were developed to ensure that a practitioner with impressive credentials could back up what he was saying, reducing the chance of a miscarriage of justice. So this means a court no longer has to accept the word of an expert - the expert now has to be able to explain how he came to that conclusion. This also means that the data becomes the focal point of the conclusion, not the credentials of the practitioner.

In effect, you are right that Pee Wee Hermanns have to understand how the practitioner came to the conclusion they arrived at based on the explanations provided by the practitioner.

Michele:

I also agree with Lisa that a Daubert hearing is a poor way to deal with a scientific debate.

So Michele, in my humble opinion, I believe it is not a problem with the method or the ability of the people to understand the method. I think there are a few more things to consider (please realise I am talking in generalities as I am not personally familiar with the case you mention):

1) Perhaps it is in part to do with the ability of the practitioner to adequately explain the method to the people (since I believe that understanding is in part a function of the teaching ability). I think this is something that can be improved on.

2) I also fear that the ability to teach the jury may then be lost in the court process since it is so long and drawn out. This unfortunately is out of our hands, but I guess a good teacher is more memorable than someone that talks over your head. So we get back to point 1.

3) Having said that, how can you shift the focus back to the data and away from the personal attacks to practitioners in court?

Lisa's idea to involve representatives from the court in a working group would be a great idea.

I would like to hear more from others on how (if!!) we can work on these things? Have they been tried before?

Kind regards,
Shaheen

[CHERIM22]

Dear Guest,
I have been reading this message board for some time now and for the first time feel compelled to reply to the personal attack you made on Michele. Michele's comment was obviously a generalized response to the idea that an expert's credentials are not as important as the data and their conclusions. She made no comments in this post about a true definition of "science". Yet you felt it appropriate to use this forum in an anonymous capacity to malign her. I find your comments offensive and in the future if you have something negative to say have the courage to log in and stand behind them!
I appreciate everyone's contributions to this forum. It is not always easy to place your opinions out there for comment. But this space is about sharing ideas, trying to learn, and to broaden our own perspectives. I have read with great interest the many engaging exchanges that have taken place.
Being new to the field, I haven't been real confident in sharing my ideas and opinions, but the one thing I can tell you is that it is comments like yours that make this space not feel comfortable for those such as myself who feel they may have something to contribute, but fear ignorant backlash from people such as yourself.
Thank you for listening and please, the rest of you, continue to provide your valuable insight to those of us who are new and trying to learn all that we can!

Cheri

[Dogma (formerly Guest)]

Please don't confuse me with the current user of the moniker "Guest."
Thank you.

[insert clever name here]

anything else...mystery guest(s)? the suspense is killing me.....

[Guest]

OK I just read my previous post concerning science being a two way street and I have to say the words came out all wrong. It was a long day and I was in a hurry to catch a ride. I will do more proof reading in the future. I apologize if I offended anyone.

Please allow me to further explain... I have seen the dark side of the "it's all about the data" argument. As a profession we used to say that the fingerprint science is 100% objective, no subjectivity involved. Old dogma. This philosophy helped us breeze through courtroom testimony but crippled our profession in the long run. In effect we were saying that fingerprints required no interpretation. It’s all about the data

In reality science relies on both data and interpretation of data. Objectivity and subjectivity. Inductive and deductive logic. The two go together why do we try to pretend the subjectivity should not exist? I believe that is misguided.

In our discipline the practitioner has data available to him or her. Then the practitioner interprets that data using their training, knowledge, and experience. The data is part of the equation and experience is also a big part of the equation. Without the experience, a.k.a credentials, correct interpretation fails to exist. So I argue the credentials we earn comparing all those images are an intrinsic aspect of our science.

Each time I see the old argument that it is all about the data and not also about the experts I feel like I am seeing a flashback to an old misguided philosophy. Now should experts hide their credentials at some point by using blind verification?

[Gerald Clough]

On Objectivity

It would seem that there have always been and always will be subjective, as we commonly speak, componants to latent print comparison, since we interpret material that varies in clarity and detail. But the terms "objective" and "subjective" can be difficult to apply, and our common usage doesn't stand up to close examination.

Although I think we worry too much about the definitions of such things as "science", "applied science" and "art", it useful to look at the applicable portions of the definitions of the two:

OBJECTIVE: "...expressing or dealing with facts or conditions as perceived without distortion by personal feelings, prejudices, or interpretations <objective art> <an objective history of the war> <an objective judgment>"

Somewhat awkward, as it speaks of "as perceived" but "without distortion by...interpretations." It would seem to imply that a fact cannot be said to be judged objectively if the perception may be altered by the nature of observer.

SUBJECTIVE: "...arising from conditions within the brain or sense organs and not directly caused by external stimuli <subjective sensations>"

Not very useful for our purposes. It could be seen to leave a gap. If directly caused by external stimuli, it's not subjective. If it's perceived but subject to interpretation, it's not objective. And in that apparent gap would lie much of what we do. But to be useful, we can consider OBJECTIVE to refer to judgments that can be reduced to a sufficiently small number of alternatives and when the judgment is made through one or more demonstrable and rational reflections on what is observed.

"Objective" does not require what we often mistakenly take to be the absolute "1 or 0" condition of a binary bit of information. Even the digital componant's judgment of "1" or "0" is based upon arbitrary definitions of the two states. A "1" is preceived when the electrical potential is greater than a. A "0" is perceived when the potential is less than b. Depending on design requirements, there may or may not be a value between a and b that cannot be defined as either "1" or "0". Such a value is deemed unusable, and no valid result can be obtained using that value. And the measurement of the potential is always subject to the accuracy of the sensor.

Our subject material presents us with a continum of clarity and detail from as sharp and detailed as could be hoped for down to hopelessly muddy. The devil is, of course, in the details. Everything that could be a legitimate feature is judged according to criteria that are, if different in nature, not really different in objectivity from the electronic interpretation of a memory bit. Having made statement and having no doubt drawn frowns, let me say that there is a difference. We are subject to influences to which the computer is, at least at the componant level, more or less immune.

But the point is (finally got to the point - aren't you glad?) that in general cases of observation and measurement where high clarity can not be assumed, there isn't any absolute objectivity. Any objectivity claimed, must be conditioned to admit some perceptive and interpretive slack. Just as in the electronic bit example, we cannot say exactly how much slack is too much or not enough to serve our practical ends. In our trade, we set the tolerance for doubt very low.

The question then becomes - How do we judge the individual and cummulative interpretations that lead to the final conclusion? In particular, how do we critically evaluate those interpretations, when they are made by our most adept? How do we recognize when there has been too much tolerance?

And a very useful question, I think, is - When several interpretations combine to lead to a conclusion, do interpretations that lie close to the tolerance limit add to the overall certainty according by the sheer number of calls consistent with the conclusion, or do do they degrade overall certainty below acceptable tolerance by accumulation of small degrees of doubt?

[Guest]

In a nutshell I think objectivity and subjectivity can be generalized by the following:

I place a clear, rolled, 12 count loop alongside a clear rolled 16 count loop in front of my sister in law (I chose her even though she is in law school--I hope that does not invalidate the study) and give her 15 seconds to compare the two images to determine a match. After it is done I ask, "Do the prints match." Bottom line, she cannot answer with any confidence.

I do the same thing in front of a latent print examiner and they laugh and make fun of me (unless they are from Scotland).

I chose my sister in law because she is intelligent, at the top of her class, with good eyesight and not corrupted by me or any other evil LPE. If it was only about the available data sister in law could do the job as well as LPE. But she can't, not even at the easiest level.

The stimuli that reached her brain was the same as for me yet I am able to process that stimuli because of my knowledge training and experience where a layman cannot. If you are into semantics, black and white objectivity and subjectivity may pose a problem. I do not have an answer for that. In general terms I think they work fine.

[Kasey Wertheim]

Dogma (Formerly Guest)

If you could contact me off-line, I would appreciate it.

Thanks,

-Kasey Wertheim
Board Administrator
(kaseywertheim@aol.com)