CLPEX Chat Board

In the McKie thread Son of Adam posted:

“I do not understand why Arie Zeelenberg would not produce an ident in court if the mark had one discrepancy. If a mark contains 30 or 40 characteristics in sequence and agreement and one in disagreement and Mr Zeelenberg deems that mark unsafe then I'm afraid he is lacking in skill, confidence and expertise.”

Prior to this post I’ve been discussing this issue with a few friends (not in relationship to the McKie case, just as a general topic – that’s the reason for this being a new thread).

SWGFAST’s definition of a discrepancy is “A difference in two friction ridge impressions due to different sources of the impressions (exclusion).” With this definition, doesn’t it seem like it’s impossible to know if something was a discrepancy or distortion until after a conclusion is arrived at? If we use this definition to exclude someone as the donor of a print, it seems to me that circular reasoning is being used (i.e. using the conclusion that’s trying to be proved as part of the proof itself).

Contrary to what’s considered to be a tenant of our profession, I agree with Son of Adam that one disagreement doesn’t discount an identification. Personally, I think the statement that “One discrepancy or unexplained difference excludes the person as the donor” is a dogmatic statement that’s untrue.

I think our profession needs to move forward and start using the term ‘acceptable difference’ or ‘within tolerance’. There are times where I’ve seen differences in prints. Sometimes there’s evidence of why this may have occurred and I also have examples where I can’t explain why the difference exists, but it doesn’t matter if I can explain it or not because there’s plenty of information to make an identification. This may not be an explainable difference but it is within tolerance due to the other information I have (acceptable). Some people may determine there’s enough information to make an identification and then come up with a reason for the difference (it must be smearing, pressure, double tap, or scar), even when there’s no data to support their reasoning. Maybe people feel compelled to explain differences instead of just saying “I can’t explain that extra ridge, but it doesn’t matter because there’s an overwhelming amount of data to support an identification by even the most conservative fingerprint expert”.

I’d be interested in how many people agree with the idea that one discrepancy or unexplained difference rules out an identification, and if there’s other people who disagree with it.

If you don't want to post your answer here, you can email me at michele.triplett@metrokc.gov

Hi Michelle,

Good choice for thread discussion and timely. I predict the main difficulty will be the interpretation of the latent/inked and what is considered acceptable as quality and quantity. The anatomy of the skin pattern is permanent so its fixed in its geometry. The accurate application of the scientific principles and rules of examination for latent prints means that human capability in the arriving at the correct outcome is a distinct variable. Using fair reasoning and with reasonable latents/inked the rule of one discrepancy is in my view valid. I just cant imagine a latent of 30 plus features in agreement that would have a distinct dissimilarity still to be detected. I'd like to see it. Cheers from oz. Les

I suspect that opinions on this subject may revolve around defining terms such as discrepancy, dis-similarity, acceptable variance, tolerance, or difference.

I have always used the term 'discrepancy' to define an irresolvable difference between two prints that would prevent them from originating with a common source, as opposed to 'dis-similarity' to describe events that may look different in two prints, yet not necessarily negate an identification. Others may use the terms differently. By my definitions, one true discrepancy should virtually always prevent an identification. An identification might still be made in the presence of one or more apparent dis-similarities.

When doing comparisons, we all routinely eliminate from consideration those known prints that obviously and plainly have discrepancies (by my definition) when compared with the questioned print. (arch/whorl, right slope/left slope loop, tented arch/large count loop, etc., as well as Galton details and derivitive configurations). Would you consider that practice dogmatic or based on scientific principle?

Also for consideration, could use of terms such as 'acceptable difference' or 'within tolerance' shorten the distance to 'probable' or 'possible' identifications?

Like Mr. Tavernaro I was trained that 'discrepancy' meant that there was a difference that could not be resolved, while 'dis-similarity' could be explained and was generally explained as some aspect of distortion (although other explanations were sometimes apparent).

One discrepancy would mandate an exclusion while two prints that had been individulized could have several areas that were not similiar in apparance but the observed unit relationship and relative postions between the two prints were the same.

At one time SWGFAST had the two (discrepancy and dissimilarity) definied differently, but in the past 4 years they have changed them to mean the same thing. Why they did that is conjecture for another thread.

I truely believe that in most of the cases involving errorenous identifications that discrepancies are there and observed by the examiner or examiners but by attributing them to some aspect of distortion they have relegated them to being only dis-similiarties and therefore explain them away.

Alice M. will be giving a very good presentation in Boston ( I hope) on the subject of distortion. A topic I feel that a lot more discussion and study should be made of.

Michele in response to your question, I feel that the term dis-similarity or not similiar in appearance would be the same as "an acceptable difference" or "within tolerance". Dis-similarity could certainly be explained that way. With discrepancy, if you cannot explain it with logic and reason, then you need to pass on it.

Discrepancy equals exclusion.
Dis-similarities equals explainable differences.

I posted this with really bad timing. I'm off work until Friday.

I have two 10-prints cards, very clear and fully rolled. I haven't counted the level two detail but I'd guess there's well over 100 characteristics (a valid ID). In one portion of the print there's an extra ridge. There's no signs of pressure, smearing, double taps, or scarring, but there is an extra ridge. If I were at work, I'd ask Kasey to add them to the 'Distortion or Dissimilarity' section (maybe I'll do this on Friday). In this type of situation, is it a dissimilarity or a discrepancy. It can't be resolved though any observation information, so I'd say it's a discrepancy....but it's a discrepancy within tolerance. Some people may say, "it must be from scarring" which is probably true but there's no evidence of this. Labeling it as a dissimilarity and attributing it to scarring after an identification is established doesn't seem like appropriate use of logic. In this case, if you cropped the image down to just this area (say 12 characteristics with 1 difference), then someone may conclude that it's not an identification.

If we have 100 characteristics then would we call this a dissimilarity? And if it were only the 12 characteristic would we call it a discrepancy?

My point is that our industry seems to promote labeling things as either distortion or a discrepancy, when in reality it may not be that simple. Sometimes we might not be able to explain something but that shouldn't discount the other information we have. I may not be able to explain every portion of a latent but that doesn't discount the other information I have.

This may be an excercise in futility,without seeing the prints in question, or having answers to questions such as: Are one or both ten-print cards live-scan? If so, who was the livescan vendor and what was the capture resolution? How were the hardcopies produced? What printer/software was used? Is there any history of artifacting in the process? How do the original high resolution screen images compare to the hardcopy? Is any additional visual information available on-screen at high magnification?

Sometimes we might not be able to explain something but that shouldn't discount the other information we have. I may not be able to explain every portion of a latent but that doesn't discount the other information I have.

Au contraire! The potential negative impact of the unexplained dis-similarity carries a weight that must be counter-balanced and overcome by the weight of your positive data. Qualitative/Quanatative factors must be considered not only for the positive data, but also the (potentially) negative. 100 clear and unambiguous level two details in sequence and agreement would likely carry a considerable amount of counter-balancing weight against a single apparent dis-similarity.

More to the point, you have at least partially provided an answer to your own concern. I think a key is your comment:

Some people may say, "it must be from scarring" which is probably true but there's no evidence of this.

An 'unexplained difference' is not the same as an 'unexplainable difference'. The fact that there is a logical explanation that could work adds weight to your positive data, even if you do not actively use it due to lack of conclusive proof. Conversely, if there is no logical explanation that 'could' work, considerable weight would of necessity be added to the (potentially) negative data.

Does any ot this make sense?

...

If there was concern with a ‘discrepancy’ in ridge formations in prints from the same individual, the acceptance of fingerprint identification may have ended over 100 years ago from the same source that launched the profession. An acknowledged ‘discrepancy’ was illustrated in ‘Fingerprints’ by Francis Galton in 1892 in the chapter on ‘Persistence’. You can obtain more information in “Galton’s Dilemma-The Split Ridge Syndrome”, John Berry, Fingerprint Whorld, 19:74 (October 1993).

Another informative article on this subject is “The one-dissimilarity doctrine in fingerprint identification”, John I. Thronton, International Criminal Police Review, No. 306 (March 1977). The Galton split-ridge is discussed as well as other examples. The following is part of a summation by Dr. Thronton, “If the one-dissimilarity doctrine is allowed to become enshrined as an immutable law of fingerprints, fingerprints such as those illustrated by Galton and in the present work could not be identified as having been made by the same finger. Such an absurd circumstance cannot be allowed to occur. The science of fingerprints is developed, both empirically and theoretically, to the point where we can afford some flexibility on this point; we have no need of pseudo-scientific dogma. It is wrong to establish a standard that is impossible to fulfill. Fingerprints are perfectly capable of being established as a certain means of identification. Let us acknowledge that the one-dissimilarity doctrine has never been demonstrated to have originated from a firm scientific basis. Once we recognize this, we will not be forced to guess the manner of occurrence of unexplained differences. In view of a preponderance of matching characteristics, one dissimilarity isn’t important.”

Ernie,

Excellent citations! It had been far too long since I last opened my copy of "Finger Prints".

I don't believe the given examples depart from my position regarding a true, unexplainable discrepancy (as I would define the term). However they do demonstrate and amplify on the single dis-similarity theme.

For those without access to Galton's book, he presents a case (among several others) of two sets of fingerprints taken from a child of 2 1/2 years, and then the same individual as a boy of 15 years. Included are illustrations in the form of a photographic plate depicting two pairs of fingerprints, each pair representing a separate finger. In this particular case, Galton notes a "single point of disagreement" in one of the pairs of fingerprints. He combines his successful points of agreement in the two separate comparisons as totalling 68 (26 + 42), with a single failure in agreement (in the pair with 42 points of agreement), stating: "This one exception, small though it may be, is in my experience unique." Galton concluds his discussion of this case with: "The student will find it well worth his while to study these and the following prints step by step, to satisfy himself of the extraordinary exact coincidences between the two members of either of the pairs...."

One must see the images to decide for themselves, however there is arguably a fitting explanation for the "single point of disagreement". The analysis is hampered (in my copy of the book) by blue text over a critical area of ridge structure.

You can go to www.galton.org and download PDF's of most of Galton's books and papers. I don't have a downloaded copy in front of me, but it may give some more perspective on this. There are both high and low resolution versions of the PDF, being 40mb and 7mb in size.

Since an examiner is unaware of the exact condition of a latent print deposit most distortion analysis must rely on experience. Hence, one discrepancy or unexplained difference within a print is quite likely. Distortion analysis is a learned trait and we would expect more unexplained differences with novice examiners.
What does help is the constant comparisons of matching impressions made over many years (and conditions). Some agencies keep 10-print cards on every arrest and make these available to the latent print examiners. Imagine the ability to look for the best exemplar to compare to a crime scene print out of stack of 15. Frequent comparisons of this type, highlight imperfections with the exemplar impressions that could often go unexplained with any accuracy. Yet, these differences do not preclude the fact of an identical source with the impressions. Thus, a rule of exclusion would have to be made on the data available at the time.

Craig,

Well said!!! The “One Discrepancy Rule” seems to be more of a quality assurance guideline for new examiners rather than a tenant of our industry. It may be a good QA policy but as an industry tenant I think it promotes false information.

It’s nice to have quick and easy explanations for what we do and how we come to our conclusions but sometimes over simplifying something just makes it wrong. “Wrong” may be a little strong. Actually it seems like the accuracy of it comes from how someone is interpreting the statement. When statements are so brief, they leave a lot more open for interpretation.

This is a case of “You Can’t Have Your Kate and Edith, Too.” Or, maybe it’s “cake” and “eat it,” anyway, you get the idea. During a very excellent thread, titled “Fingerprint Dogma Final Exam,” many of you offered clear and well thought out explanations about the underlying principles of fingerprint identification. We observed in that thread that the two leading traits of friction ridge structure that allow us to individualize to a single source are uniqueness and permanence. Uniqueness meaning that no print is identical to any other and permanence meaning that that details or arrangement of the ridge structure never changes absent accident, mutilation or disease. That being the case, how can you move forward with a declaration of a fingerprint individualization if there exists an unexplainable difference? We hear all the horror stories about how AFIS is presenting us with prints that are closer and closer in appearance of the ridge structure. In one of his first public appearances after the Mayfield debacle, Ken Moses suggested to a rather large audience that we may have to acknowledge the existence of prints from different sources with as many as 8 details in agreement. I know that you Ridgeologists out there will call a print on less than 8 Galton Points.

I suppose what I am getting at is that if we are going to move in the direction that the “scientists” wish us to head, that is away from the absolute declaration of a match to the exclusion of all other sources to the more acceptable “Within an high degree of scientific certainty,” we can live with that unexplained difference. However, if we are going to stick with the absolutist posture, then you better be able to explain why that difference exists. It can’t be left unaccounted for.

Re: C. Coppock’s post, I would hope that those of you who are matching prints to specific sources are not doing so without having the opportunity to print the suspected source (as many times as needed to record the area of interest). Or, you are wording your reports to refer to the fingerprint image which may appear on a certain record or card and not to an individual.

Hi Michele,

I had a look over the images you posted and your observations. It was interesting to see the width of the sulci in the area before the extra ridge appears. It is also interesting to see the closeness of the sulci on one side of the extra ridge. Where I am going too is the inconsistency of the spacing of the sulci which asks the question about the reliability of the data. To resolve this matter I would suggest an experiment where the donor was asked to supply several repeats of the digit. When the digits is being rolled/scanned then care should be taken to avoid pressure distortion. Regards. les