T H E
W E E K L Y D E
T A I L
Welcome to the second "Detail," a weekly e-mail
newsletter that greets you every Monday morning. The purpose of the Detail
is to help keep you informed of the current state of affairs in the latent print
community, to provide an avenue to circulate original fingerprint-related
articles, and to announce important events as they happen in our field.
I must say, I am truly humbled and amazed by the response to both the website and the newsletter. The site has had over 300 hits in the last week, and there are now nearly 50 subscribers to the newsletter, with new sign-ups almost daily! Thank you for your confidence, and I look forward to the exchange of information for the advancement of the latent print discipline.
We are continuing a short series on Daubert issues, with this week's topic being Testing and Validation. In U.S. v. Mitchell, the defense claimed that no testing had been carried out providing scientific proof that fingerprints are permanent and unique; therefore, fingerprint evidence could not meet the legal standard of "reliable evidence" as set forth in the Federal Rule 702 and could not to be admitted into evidence by the trial judge. The prosecutions argument involved several elements. First, testimony as to the biological uniqueness of skin was given, emphasizing that so many factors affect the stress and strain across the surface of skin during formation, that no two friction ridge arrangements, or even friction ridges themselves, could ever develop in exactly the same way. The statistical models proposed by researchers from Galton up to a recent FBI study involving AFIS were then offered as supporting the biological uniqueness of skin and validating fingerprint methodology. Further, the prosecution argued that even though no specific controlled testing had been done, 100 years worth of thousands of law enforcement agencies conducting millions, probably even billions of comparisons across the globe and never finding two areas of skin exactly the same was considered empirical validation of the use of fingerprints to individualize.
The FBI AFIS study briefly mentioned above, involved comparing 50 thousand prints against themselves, one by one, resulting in 2.5 Billion comparisons. A statistical analysis of the data was conducted to determine the probabilities involved in a certain number of details corresponding in two different prints. To make the test more difficult (or more realistic, depending on how you look at it), only left slope loop patterns were used. By analyzing the score differences between the candidates, including the score upon "hitting" itself, a "Z" score was obtained. The results show that there is a very significant difference between the score of a matching print and that of the next closest candidate to it. In short, even in a database of 50K prints, the statistical probability that two different prints would have 4 minutia in the same position is 1X10 to the 27th power. Apply this staggering statistic to the 5,900,000,000 population of the earth, and the chances of any two areas of skin of anyone on earth having the level of agreement tested in the FBI study is approximately 1 in 10,000,000,000,000,000. And that's just with an AFIS system using level 2 detail alone. Include level three detail, and the chances are even smaller. A lot smaller. (As if it was within comprehension to begin with)
For those of you who are thinking to yourself, "I have seen 4 points of agreement in two different prints," that would depend on how you define "agreement." The FBI study used the same print to compare against the database, including itself, so naturally their agreement was fairly strict. But even if distortion were included in a test of this nature, it probably wouldn't take many more details to reach those same astronomical chances. In other words, instead of 4, it might take a few more level 2 details in agreement to reach odds of 10^27. Regardless, if biological uniqueness is fully understood, an examiner on the witness stand can explain that there is no chance of two areas of skin ever developing in exactly the same way. Nature never repeats itself. Friction ridge skin develops in an environment that is subject to random stresses and strains due to unique growth of the individual. The detail which is formed during this critical 6-7 week period of formation is affected by those and other environmental factors to such an extent that is entirely and completely unique in every aspect. Every detail of this unique template is faithfully represented on the volar surface due to elements of skin structure which prevent single cells from shifting during their migration to the surface. Basically, biology fixes in place, very early in development, a unique root system that consistently feeds new growth, in concert, to the surface.
I plan on a Weekly Detail devoted to the issue of the fingerprints, twins, and clones which will address the issue of pattern formation versus the formation of level 2 and level 3 detail. So for now, let's leave biological uniqueness to what has already been said, and address it further at a later time. The point I want to make here is that the reason fingerprints are unique is because of biological uniqueness, not because statistics say so. Statistical analysis of detail merely approaches the model of uniqueness and validates use of the detail measured, but it cannot account for what it cannot measure. No statistical model can ever measure uniqueness, therefore statistics can support but not prove uniqueness.
If you are pushed to comment on this issue, I believe you should utilize the statistical studies to support biological uniqueness. The issue could be addressed with a statement to the effect of, "The bottom line is, computer models don't even begin to take into account all the detail used by a competent examiner during the course of an analysis, comparison, and evaluation of a latent and an inked print; But even in spite of this, we still see, based on controlled statistical testing involving the limited detail that AFIS systems do use, that the chances of having a significant correspondence of ridge formations in two different prints are so incredibly small as to be considered impossible." Basically, Bruce Budowle's testimony at the Daubert hearing emphasized that the available statistical data pertaining to fingerprint methodology validates fingerprint identification. When coupled with a competent examiner following the Analysis, Comparison, Evaluation process and having their work verified, fingerprint identification is a science, the error rate of the science is zero, and, to borrow from the Government's Motion in Limine, "the science of fingerprints is alive and well and so it should remain."
The prosecutions approach worked quite well. Of course, we know the judge denied the motion to exclude fingerprint evidence, and he even provided judicial notice in his decision that “human friction ridges are unique and permanent throughout the area of friction ridge skin, including small friction ridge areas...” and “...friction ridge skin arrangements are unique and permanent.” The major theme of the objection by the defense was that the use of partial latent prints containing distortion has never been tested. The judge also recognized that "Individualization, that is, a positive identification, can result from comparisons of friction ridge skin or impressions containing a sufficient quality (clarity) and quantity of unique friction ridge detail."
There was another issue brought up later that dealt with the March, 2000 NIJ solicitation for "Forensic Friction Ridge (Fingerprint) Examination Validation Studies." The defense argued that the NIJ issued the solicitation for validation studies, so they must have realized there was no validation to fingerprint identification. Of course, nothing could be further from the truth. Perhaps the NIJ response to this attack sums it up best: "What underlies this solicitation is a desire for more research to confirm the already existing basis that permits fingerprints to be used as a means to individualize." They go on to say, as anyone should say about any scientific endeavor, that "...procedures should be periodically examined and research conducted to enhance their empirical foundation." That was the thrust of the NIJ study.
The 2000 solicitation and the NIJ letter of response can be read in their entirety here.
The NIJ subsequently withdrew it's original solicitation, and issued another more "General Forensic Research" solicitation in June, 2001 which can be viewed here.
Next week we will take a look at the issue of General Acceptance. Rule 702 still provides for a judge to weigh the reliability of scientific evidence by the standard of Frye. It seems silly to even argue the acceptance of fingerprint evidence in the 21st century, but some interesting points can be raised by defense attorneys involving the surveys and tests used by the prosecution to respond to this issue. I think we would all agree that it is much better to be prepared for the question that isn't asked than to not be prepared for the one that is.
If you have questions or comments on this weeks issue, please
visit the Discussion
forum. (I apologize in advance for the advertisements... that is part of
using "free" message boards)
If you have not yet signed up to receive the Weekly Detail in YOUR e-mail inbox, go ahead and join the list now so you don't miss out! Past Details are now available in the Detail Archive room, if you have missed any and want to read up.
Feel free to pass the link to The Detail along to other
examiners. This is a free service
BY latent print examiners, FOR latent print examiners.
There are no copyrights on The Detail, and the website is open for all to
Mailings of The Weekly Detail will be limited to only those who sign up to join the Topica e-mail newsgroup list. Allow me to assure you that privacy is strictly adhered to, and Topica doesn't send unsolicited e-mails. If, instead of signing up, you just want to check back often, remember www.clpex.com and click on The Detail.
ON THE SITE THIS WEEK:
Until next Monday morning, don't work too hard or too little.