SUPERIOR COURT OF  CALIFORNIA 

 COUNTY OF SAN FRANCISCO

THE PEOPLE OF THE STATE OF                ) CALIFORNIA,                                             )

                                                                        )

                                    Plaintiff,                        )                                                                       )

                                    vs.                                )

                                                                        )

ROBERT NAWI                                             )

                                                                        )

                                    Defendant.                    )

___________________________________  )
 

SCN:   176527

MCN:  1812436 

Date:                March 3, 2000

Time:                9:00 a.m. 

Department:      24 

REPLY TO PEOPLE”S OPPOSITION TO MOTION TO EXCLUDE FINGERPRINT IDENTIFICATION EVIDENCE AND REQUEST FOR A HEARING PURSUANT TO PEOPLE V. KELLY (1976) 17 CAL. 3D 24, OR, IN THE ALTERNATIVE, MOTION FOR FUNDS TO RETAIN FINGERPRINT EXPERTS AND TO PERMIT THEIR TESTIMONY BEFORE THE JURY

                        PRELIMINARY STATEMENT  

            The prosecutor repeatedly makes one, and only one, point about Mr. Nawi’s fingerprint motion: “It’s too long.” The Bible is long; so is Tolstoy’s War and Peace. By the prosecutor’s child-like reasoning, these works of art should be rejected out of hand because they are  “too long.” The defendant’s  motion, and the extensive scientific evidence upon which it is based, must be judged on it’s merits, not on the basis of catchy phrases or irrational fears of  information contained on the internet. As former prosecutor and now  Northern California Federal District Court Judge Lowell Jensen put the matter, “The government is correct in their assertion that pre-Daubert/Kumho/ Ninth Circuit precedent supports the admissibility of (expert) testimony; however, the world has changed. The Court believes that... a past history of admissibility does not relieve this Court of the responsibility of now conducting Daubert/Kumho analysis as to this proffered expert testimony.” United States v. Santillan (N.D. Cal. 1999) __F.Supp. __,1999 WL 1201765 at p. 4. See also, People v. Soto, (1999) 21 Cal. 4^th 512, 540-541 n. 31(“(I)n a context of rapidly changing technology, every effort should be made to base (decision) on the very latest scientific opinions...”); People v. Allen (1999) 72 Cal. App. 4^th 1093, 1101(“The issue is not when a new scientific technique is validated, but whether it is or is not valid; that is why the results generated by a scientific test once considered  valid can be challenged by evidence the test has since been invalidated.”);People v. Smith (1989) 215 Cal.App.3d 19, 25(in determining whether a particular technique is generally accepted "defendant is not foreclosed from showing new information which may question the continuing reliability of the test in question or to show a change in the consensus within the scientific community concerning the scientific technique").

             The prosecutor admits that a Kelly hearing is required if the defendant presents “new evidence... reflecting a change in the attitude of the scientific community.” People v. Kelly (1976) 17 Cal. 3d 24.(emphasis added).¹ In a non-sequiter, however, the prosecutor then argues that defendant’s motion for a Kelly hearing should be denied because Mr. Nawi “cites more to web pages than cases.” But by the prosecutor’s own admission the focus must be on the attitude of the scientific community, not on how many cases Mr. Nawi can cite as reflective of the attitudes of the judicial community.

            The government has failed to demonstrate that its latent fingerprint identification evidence possesses the various indicia of scientific and forensic  reliability set forth by the Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579  (1993) and our Supreme Court in People v. Venegas (1998) 18 Cal.4th 47 and the other cases cited in defendant’s motion.   Most significantly, as even a cursory review of the Mitchell transcripts and other Exhibits in support of the motion will reveal, there has been no testing of any of the fundamental premises that underlie latent fingerprint identifications.  As the government witnesses in Mitchell conceded, there has been no controlled studies performed so as to determine the reliability of identifications which are based on small distorted latent fingerprint fragments such as those at issue in the case at bar.  Nor has there been any testing done so as to determine the minimum amount of corresponding detail that a fingerprint examiner should find before making an identification.

            Because no testing of this nature has been performed, there is no established error rate for latent fingerprint examiners.   It is clear, however, from many real life examples, that errors do occur.   Moreover, there is substantial reason to believe that additional errors have gone undetected, given the alarmingly high rate of misidentifications that have occurred during the last several years on latent fingerprint examiner proficiency exams.   The government’s experts in Mitchell provided no explanation for the shockingly poor results of these exams and the prosecutor in this case is also silent on this issue, but does admit the astonishing fact that the AFIS system “has an error rate of 70-75%.” (Opposition to Motion to Suppress) .  Instead, the government in Mitchell and the prosecutor here resort to the argument that human error rate is irrelevant, a contention that not only conflicts with Daubert and Venegas, but which is patently absurd in light of the fact that latent fingerprint identifications are based entirely on subjective human judgment.

            The lack of experimentation that has been done with respect to latent fingerprint identifications also has resulted in the failure to establish an objective identification standard.  As the government’s experts in Mitchell all acknowledged, a fingerprint examiner’s opinion of identification is completely subjective.   Shockingly, the government’s experts conceded that the varying point standards that fingerprint examiners have been relying upon, and testifying on, for the past 90 years are not scientifically based and are instead the product of what can, at best, be described as  educated conjecture.   The extreme subjectiveness of latent fingerprint identifications is especially troubling given that many latent print examiners are poorly trained and minimally qualified.

             In addition to these various Daubert factors, the unreliability of latent print comparisons was amply demonstrated in Mitchell the survey that the government conducted of state law enforcement agencies.   Of the 35 agencies that, initially responded, eight (23%) found that there was an insufficient basis to make an identification with respect to one of the two latent prints at issue in this case and six (17 %) found an insufficient basis as to the other.   That the government subsequently went to the extreme lengths that it did to convince each of  these agencies to change their opinions, only serves to demonstrate just how significant the original results of this test actually are.        

            Unable to satisfy any of the key Daubert and Downing factors discussed above, and reeling from the results of the state examiner survey, the government, at the conclusion of the Daubert hearing in Mitchell, submitted that even if its fingerprint identification evidence is not admissible as “scientific” knowledge, it may still be admitted as “technical” or “specialized” knowledge.    The government, however, cannot so easily evade the criteria of Daubert.   As the Supreme Court recently held in Kumho Tire v. Carmichael, 119 S. Ct. 1167 (1999), the Daubert factors may properly be applied to evaluate the reliability of all expert testimony,  not just scientific expert testimony.  The prosecutor in this case can offer no plausible reason why the Daubert factors should not be used to assess the reliability of such critically important evidence as  latent fingerprint identification testimony.   It is beyond dispute that testing in the fingerprint field can be performed, an error rate established, and an objective identification standard developed.   That none of this has occurred, cannot be glossed over simply by characterizing the evidence as well accepted by a heretofore uncritical.   The essential fact remains; the government has not established the reliability or general scientific acceptance of its latent fingerprint identification evidence.

            Finally, the unreliability of  latent fingerprint identifications has already been judicially recognized.   In the only known fingerprint case in which a federal trial court has performed the type of analysis that is now mandated by Daubert, the district court excluded the government’s fingerprint identification evidence.  United States v. Parks (C.D. Cal. 1991) (No. CR-91-358-JSL) (Ex. 48).  The district court in Parks reached its determination after hearing from three different fingerprint experts produced by the government in an effort to have the evidence admitted.  In excluding the evidence, the district court recognized, among other things, the lack of testing that has been done in the field, the failure of  latent fingerprint examiners to employ uniform objective standards, and the minimal training that latent  fingerprint examiners typically receive.

            Accordingly, for all of the foregoing reasons, Mr. Nawi requests that this Court preclude the government from introducing its fingerprint identification evidence at his upcoming trial. Mr. Nawi does not think he has “saddl(ed) the court with hundreds of pages of testimony from (United States v. Mitchell).” Opp. p 3). The prosecutor cannot have it both ways. He cannot complain that defendent must demonstrate a change in the attitude of the scientific community and then turn around and whine when the defendant attempts to satisfy that burden. Nevertheless, defendant is sensitive to the burden which his showing places upon the Court. He therefore offers the following summary of the testimony elicited in Mitchell as an aid to understanding the scientific basis of the defendant’s motion. 

 THE DAUBERT HEARING IN MITCHELL 

                        A.        The Government’s Witnesses   

            The government offered seven experts to support its position that fingerprint evidence is admissible as scientific or technical evidence.  These experts, by and large, testified to their acceptance of the prosecution’s three premises, set forth in Government Exhibit A:

1.  Human friction ridges are unique and permanent;

            2.  human friction ridge skin arrangements are unique and permanent;

3.  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. 

            The first two of these premises, however, that entire fingerprint patterns are “unique and permanent,” are largely irrelevant to the question concerning the reliability of latent fingerprint identifications which are based on small distorted fingerprint fragments.²  And, with respect to the much more significant third premise, none of the government’s witnesses were able to answer the fundamental question which that premise leaves open:  what is a sufficient quantity and quality of  friction ridge detail from which a latent fingerprint identification can reliably be made.   The reason for this failure is simple; there has been no testing in the fingerprint field so as to determine this.

                        (i)  William Babler

                        The government’s first witness was Dr. William Babler, an expert in the prenatal development of friction ridges.  While Dr. Babler presented an interesting discussion of certain embryology studies concerning how and when fingerprints form during fetal development, he readily conceded that these studies have nothing to do with latent fingerprint identifications. (Tr. 7/7 at 75).   Indeed, while Dr. Babler testified to his belief in premises one and two set forth above, he did not opine as to premise three, that positive identifications can be made from a comparison of a latent impression with a known exemplar.  (Id. at 72-74). 

                        (ii)  David Ashbaugh

                        The next government witness was Staff Sergeant David Ashbaugh of the Royal Canadian Mounted Police.   Ashbaugh, whose education did not extend beyond high school, testified that the scientific basis for latent fingerprint identifications consists of the embryology studies that Dr. Babler referred to.  (Id. at 86, 213-214).  Ashbaugh, however, did not offer any explanation as to how these studies demonstrate the reliability of latent fingerprint identifications or provide any inkling as to what constitutes a sufficient quantity of friction ridge detail on which an identification may reasonably be based.  

            Ashbaugh repeatedly acknowledged that a latent fingerprint examiner’s opinion of identification is subjective.  (Id. at 115, 166).   Indeed, in a 1993 article that Ashbaugh wrote entitled, Premises of Friction Ridge Identifications, he went so far as to state that the opinion of identification is “very subjective.”  (Id. at 151).   While Ashbaugh testified that he was “exaggerating” when he wrote this article (Id. at 153), in his most recent writing on the subject, Ashbaugh again emphasized the subjectiveness of a latent print identification, this time by writing the word “subjective” in all capital letters.  (Id. at 167); (Gov’t Ex. 10 at 99).   

            Ashbaugh also revealed in his testimony that the various “point” standards that latent fingerprint examiners have employed, and have been testifying on, over the past 100 years are not scientifically based.  (Id. at 168).  Rather, as Ashbaugh candidly conceded, these  point standards are based on nothing more than “educated conjecture.”  (Id.).  Accordingly,  Ashbaugh testified that it is unacceptable for point standards to still be utilized.  (Id. at 168, 169).  He admitted that he was unaware that examiners in this country were continuing to use them.³  (Id. at 169).

            In Ashbaugh’s opinion, there should not be any minimum identification standard. (Id. at 168).   Rather, he asserts that the “standard” in the fingerprint field “is in the training of the expert and the knowledge that the person has.” (Id. at 168, 186).  Ashbaugh acknowledged, however, that he is “not familiar with the training” that latent print examiners in this country are provided.  (Id. at 174).  Ashbaugh testified that in his opinion latent fingerprint examiners should be board certified and subjected to “blind” proficiency testing before they are permitted to testify in court.  (Id. at 200, 201).  As he recognized though, there is no board certification requirement in this country.  (Id. at 200).  Ashbaugh professed ignorance as to whether any blind proficiency testing is conducted in this country.  (Id. at 201).   

            Ashbaugh testified that rather than simply looking for points of comprison he and other latent fingerprint examiners currently employ a “quantatative - qualitative” process known as “ridgeology” that consists of four components,  analysis, comparison, evaluation and verification.  (Id. at 110, 135, 146).  Ashbaugh, as well as some of the government’s other witnesses, also referred to this process by the acronym ACE-V.”  Although Ashbaugh conceded that the critical evaluation step of this process is highly subjective, he insisted during his testimony that the analysis and comparison stages are of an objective nature.  (Id. at 115).  His testimony, however, disclosed otherwise.    In demonstrating to the court how he conducts a comparison, Ashbaugh unwittingly revealed the very subjective judgements that are made throughout the entire process:

So when I looked at the first ridge, that was acceptable.  The second ridge, this area is a little different.  And because there’s no lateral pressure, I would have a concern.  This area here, I would accept the top -- the top area I would accept.  The shape of this particular ridge feature, I feel even with less pressure you should be able to see that pointed aspect of that short ridge.

When I move on to this ridge, there’s a dissimilarity here that will -- in my opinion, would not mean a difference.  Dissimilarity to me is something you can’t accept, but when it looks like this, I believe that is now moved into the difference category.

Then, of course, the ridge ends here and that’s reasonably in agreement.

.     .    . 

The next ridge comes down and has a dogleg.  This ridge comes down and the dogleg is not as abrupt.  And that again would concern me.

I feel that is bordering on disagreement.  This particular short ridge in the unknown has a wide area at this end and then moves narrow, and this is pretty well straight all the way through.

I feel that this would be a disagreement because even with less pressure, things would get thinner on this ridge.  You would still have that shape.

The next ridge in the unknown print -- I apologize.  I’m so used to looking at unknown on the left-hand side and this is backwards.

            In the unknown print, this ridge runs straight through between two short ridges.  This ridge makes a major dogleg and that is disagreement.  That isn’t acceptable.

This other short ridge, it just doesn’t seem to be the same length.  And where this ridge ends, a little bit farther down the ridge, where, if you look at -- I’m sorry, this ridge ends a little farther up the ridge of the short ridge, if you draw a line straight across here to their relationship.  And yet this one is down quite a bit farther.

If I moved to the next ridge across and followed it down, it has a dogleg again coming in nice and tight here.  I’m not sure I would accept that, but I likely would.

The next ridge, the angle here is a little bit different than the angle here (indicating).  But again, there’s a great deal of pressure so I may even accept that as being an agreement. 

(Id. at 121-123) (emphasis added).

                                                                                                                 

            As Ashbaugh’s demonstration thus revealed, latent fingerprint examiners make subjective jugement calls throughout the entire comparison process as to whether various ridge characteristics are in agreement or disagreement.  Neither Ashbaugh, nor any of the government’s other witnesses, testified to any standard rules or measurements that exist which would serve to guide latent print examiners in making these determinations.

               Ashbaugh testified that “ridgeologists”, such as himself, do not simply look at the traditional “Galton” characteristics when comparing a latent print to a known exemplar.  (Id. at 150).   The Galton characteristics, which Ashbaugh referred to as “second level detail,” are what latent print examiners have traditionally counted when making identifications on the basis of point standards.  (Id. at 126, 130).  Ashbaugh testified that in addition to Galton characteristics he looks for “third level” detail such as sweat pores and the shapes of the ridge edges.  (Id. at 150).  In so testifying, Ashbaugh expressed his disagreement with the position of James Cowger that because “prints of friction skin are rarely well recorded ... comparison of pores or edges is only rarely practical.”  (Id. at 150); see James F. Cowger, Friction Ridge Skin: Comparison and Identification of Fingerprints at 143 (1983).  Ashbaugh testified that he was unaware that the FBI had previously expressed the same view as Cowger.  (Id. at 213); see, An Analysis of Standards in Fingerprint Identification, FBI Law Enforcement Bulletin (June 1972) (Ex 1 at 3).  (“observations on pores have shown that they are not reliably present and that they can be obliterated or altered by pressure, fingerprint ink, or developing media;” the FBI “knows of no case in the United States in which pores have been used in the identification of fragmentary impressions.”).

            Ashbaugh acknowledged during his testimony that all fingerprints suffer from distortion which can result from 1) the pressure exerted at the time the print is made, 2) the shape of the surface on which the print is found, 3) the material used to lift the print, 4) the medium or subatrate in which the print is found, and 5) anatomical aspects of the finger itself.  (Id. at 159-162).  Ashbaugh conceded that these distortions may cause a ridge characteristic to appear to be something other than what it really is.  (Id. at 160, 161).  As Ashbaugh recognized, a major part of a latent print examiner’s job is to determine whether the features that are observable in a particular print are genuine or a product of distortion.  (Id. at 162).  

            Ashbaugh took issue with the view expressed by Dr. James Thorton that latent fingerprint examiners will routinely make up explanations regarding distortions so as to explain away differences in prints once the  examiners have become convinced that the prints were made by the same finger.  (Id. at 140).; see John Thornton, The One Dissimilarity Doctrine in Fingerprint Identifications, 306 Int’l Crim. Police Rev. 89 (March 1977) (Ex. 38).  Ashbaugh, conceded however, that he is unaware of any paper that has ever been written disagreeing with Dr. Thorton’s article.  (Id. at 166).

            Finally, Ashbaugh acknowledged that there have been no studies performed from which fingerprint examiners can determine the probability of different people having a certain number of fingerprint ridge characteristics in common.  (Id. at 187, 190, 193)   Lacking any such probability studies, examiners testify in terms of absolute certainty; i.e. that a latent print was made by a particular finger to the exclusion of all other prints in the world.  (Id. at 190). 

                        (iii)  Edward German   

                        The next prosecution expert was Edward German, another law enforcement witness.  Mr. German is a latent print examiner, employed by the U.S. Army Criminal Investigation Laboratory.  (Tr. 7/8 at 2).  Mr. German, who also does  not possess a college diploma, testified that he is a member of the Scientific Working Group on Friction Ridge Analysis Study and Technology (“SWGFAST”)⁴.  (Id. at 12, 14).  This group has recently promulgated recommended guidelines concerning qualifications and training of latent fingerprint examiners.  (Id. at 12).  These recommended guidelines were introduced as government exhibit 4-4.  As this exhibit states on its first page, these “draft”guidelines have been presented for “consideration and comment.”  Neither Mr. German, nor any of the government’s witnesses, presented any testimony or evidence that these draft guidelines have been adopted by any law enforcement agency at either the state or federal level, including the FBI.  Accordingly, in this country, there continues to be no minimum training or qualifications requirements for latent fingerprint examiners. 

            Mr. German also testified about certain twin studies which he and others have conducted.  (Id. at 20–28).  In Mr. German’s opinion, these studies have established that fingerprints are not genetically determined.⁵  (Id. at 46).   Neither Mr. German, however, nor any of the government’s other witnesses claimed that these studies provide even a clue as to what constitutes a sufficient quantity of friction ridge detail on which a latent print identification can reliably be based.  

                        (iv)  Steven Meagher   

                        The government’s next witness was Stephen Meagher, employed by the F.B.I. as a fingerprint specialist unit chief.  Mr. Meagher, who also does not have a college degree, testified, like Sergeant Ashbaugh, that the embryology studies referred to by Dr. Babler, provide the scientific basis for latent fingerprint identifications.  (Tr. 7/8 at 61; 7/9 at 13).  Like Ashbaugh, however, Meagher was unable to explain how these studies demonstrate the reliability of latent fingerprint identifications, or provide any inkling as to what constitutes a sufficient quantity of friction ridge detail on which a latent print  identification can reasonably be based.     Accordingly, Meagher did not disagree with the opinion of forensic science commentator, Michael J. Saks, that a “vote to admit fingerprints is a rejection of conventional science” and that “a vote for science is a vote to exclude fingerprint expert opinions.” (Tr. 7/9 at 17, 18).  

            Like Ashbaugh, Meagher conceded that a latent fingerprint examiners’s opinion of identification is subjective.  (Tr. 7/9 at 15).  Meagher testified that the FBI stopped using an objective identification standard in the late 1940's and that the FBI currently uses the ACE-V methodology testified to by Ashbaugh.  (Tr. 7/8 at 105, 106).  When asked to explain why the FBI fingerprint examiner at Mr. Mitchell’s first trial testified in terms of points of similarity, Meagher asserted that this was just the examiner’s simplistic way of explaining the identification to the jury.  (Id. at 98, 99).  Meagher stated that he had no opinon as to whether latent fingerprint examiners outside of the FBI continue to employ point standards.  (Tr. 7/8 at 227).⁶  

            Meagher also described the various surveys he had constructed for the purpose of the Daubert  hearing.  Most relevant to the proceeding was the survey in which photographs of the latent prints at issue in this case and Mr. Mitchell’s inked prints were sent to 53 law enforcement agencies for comparison.  Of the 35 agencies that initially responded to the government’s request, eight (23%) did not make an identification with respect to one of the two latents and six (17%) did not make an identification as to the other.  (Tr. 7/9 at 207).   Mr. Meagher testified that he subsequently sent the photographs back out to these agencies, along with marked up enlargements displaying the common characteristics that he (Meagher) had found, with the request that the agencies perform a new examination and complete a new survey form.  (Id. at 210, 211).  Meagher acknowledged that it is “not common” for “examiners to get blowups as part of their examinations.”  (Id. at 118).  Nevertheless,  Meagher claimed that he did not send the marked up enlargements to the state agencies because of any concern on his part that the agencies’ initial responses might be detrimental to the government’s interests.  (Id. at 154).  Indeed, Meagher testified that in his view “practictioner error” is irrelevant to a Daubert hearing.⁷  (Id. at 152-154).  Meagher testified that he sent out the enlargemnents to the agencies as a training tool so as to demonstrate the mistakes which, in his view, they had made.  (Id. at 124).  When questioned, however, as to why, if his motivation was simply to educate the examiners, he urged them to quickly complete the new response form and return it to him in advance of the Daubert hearing, Meagher testified that it “was just decided that it would be in our best interests to do that.” (Tr. 7/9 at 8, 9).

            Meagher also offered explanations as to why the state latent print examiners had not made identifications.  Either they had “just screwed up”, or they had lacked the appropriate experience, or it was late in the day,  or they didn’t realize the result was being used in a pre-trial hearing.  (Tr. 7/8 at 134-150).  (This latter explanation implies that knowledge of the purpose of the survey would have influenced the determination of identification.)⁸

            Finally, Meagher testified, on his direct examination, with respect to two experiments that were conducted on the FBI’s Automated Fingerprint Identification System (“AFIS”), by the government’s AFIS provider, Locckheed Martin.   On cross examination, however, Meagher revealed that he did even possess the necessary expertise so as to be able to explain the basic terminology that Lockheed used in the conclusion section of the test report that it generated in connection with these two experiments.  (Tr. 7/8 at 196, 198).⁹

                        (v)   Don Zeisig            

                        The government next called Don Zeisig, an electrical engineer at Lockheed Martin who conducted the AFIS experiments to which Meagher referred.   Mr. Zeisig testified that the two experiments utilized a database of 50, 000 fingerprints extracted from the FBI’s Criminal Master File.  (Tr. 7/9 at 81).  The first experiment compared each fingerprint with itself, as well as with all of the other fingerprints in the database.  (Id.)  The scores generated by AFIS from these comparisons were converted into “Z” scores and then probability measures.  (Id. at 82, 83).  Not surprisingly, whenever AFIS compared a fingerprint image with itself, an extremely high score was generated.¹⁰  (Id. at 82).  Much lower scores were obtained when different fingerprint images were compared.  (Id. at 84).  The conclusion that Lockheed derived from this experiment was that the “probability of a non-mate rolled fingerprint being identical to any particular fingerprint is less than 1/10 (1 followed by 97 zeros).  (Id. at 84).  As Mr. Zeisig explained, this probability was taken from the lowest score that AFIS generated when a particular fingerprint image was compared with itself.  (Id. at 85).

            Significantly, Mr. Zeisig conceded on cross examination that even rolled fingerprints which are taken of the exact same finger will not be identical because of the various distortions that occur in the rolling process.  (Id. at 86, 87).¹¹  Indeed, this fact was unintentionally demonstrated by the experiment.  Included in the 50,000 print database were multiple fingerprints that had been taken of the same fingers. (Apparently some individuals had been fingerprinted twice by the FBI).  (Id. at 87).  Three different examples of this were found to exist in the database, though Mr. Zeisig conceded that there could be others that went undetected.  (Id. at 87-94; Ex. 54 at 4).  Significantly, in each of these instances, the scores that were generated by  AFIS, when comparing two different fingerprints of the exact same finger, were significantly lower than the scores that were obtained when each fingerprint in the database was compared with itself.  (Id.)   In fact, in some instances, the scores generated was so low as to fall well within the range of scores that were generated when fingerprints of different fingers were compared.  (Id. at 91, 92).  Thus, there were some fingerprints of different fingers that AFIS found to have greater similarity than fingerprints of the same finger.  Accordingly, Mr. Zeisig acknowledged that if two people actually had fingers with identical fingerprint patterns, and rolled fingerprints of those fingers were compared by AFIS, the score generated would in all likelihood be far less than the scores that Lockheed obtained when each fingerprint image was compared with itself.  (Id. at 94-95).   In other words, fingerprints of identical fingers would not meet the definition of identical that Lockheed established through the methodology of comparing fingerprints with themselves.  (Id. at 92, 94-95).

            The second AFIS experiment which Mr. Zeisig testified to was very similar to the first.  Again each fingerprint was compared with itself and with all of the other fingerprints in the database.  (Id. at 95).  However, this time each fingerprint, prior to comparison, was converted into a simulated “latent” fingerprint by extracting the central 21.7% of the print.  (Id. at 96).  This central portion was than compared with the entire print from which it had been extracted.  (Id.)  Accordingly, with respect to the middle 21.7%, identical images were again being compared, and as such the scores generated from these comparisons were extremely high, much higher than the scores that were obtained when the simulated latents were compared with different fingerprints.  (Id. at 96, 97).  The conclusion that Lockheed derived from this experiment was that “the probability of a non-mate fingerprint being identical to a minutia subset of any particular fingerprint is less than 1/10 (1 chance in 1 followed by 27 zeros) for small numbers of minutiae (in this case small means four), decreasing to less than 1/10 (1 chance in 1 followed by 97 zeros) for larger numbers of minutia (in this case larger means greater than eighteen.).”  (Id. at 97, Gov’t Ex. 6-8).  Again, these astronomical probabilities were simply derived from the scores that were generated when fingerprint images were compared with themselves.  (Id. at 98).   And, as with the first experiment, Mr. Zeisig answered yes to the hypothetical posed by defense counsel: If  two people possessed fingers bearing identical  fingerprint patterns, and a simulated latent print was created from a rolled print of one of these fingers and than compared on AFIS with a rolled print of the other identical finger, the score generated would in all likelihood be far less than the standard for identical that Lockheed created by comparing images with themselves.  (Id. at 99, 100).

                        (vi)  Bruce Budowle

                        The government’s final witness on its direct case was Dr. Bruce Budowle, a geneticist employed by the FBI.  (Id. at 105).   Dr. Budowle’s lengthy resume does not reveal any background or experience with fingerprints, and he did not claim any such experience or expertise during his testimony.  (Gov’t Ex. 8; Tr. 7/9 at 105-110).   Nevertheless, Dr. Budowle offered his opinion that latent fingerprint identifications are scientific.  (Id. at 161).  Unlike the government witnesses that testified before him, Dr. Budowle did not premise his opinion on the embryology studies that were the subject of Dr. Babler’s testimony.  Rather, Dr. Budowle referred to certain theoretical probability models, including one developed by defense expert Dr. David Stoney.  (Id. at 121, 161).  As Dr. Budowle recognized at another point in his testimony, however, a theoretical model must be tested before it is relied upon.  (Id. at 168).  It was therefore curious that  Dr. Budowle did not comment upon the fact that the probability models that he was relying  upon have never been tested.

            In opining that latent fingerprint identifications are scientific,  Dr. Budowle also expressed the view that the “100 years of fingerprint employment has been empirical studies,” which have demonstrated that “no two unrelated individuals or related individuals have the same print.”  (Id. at 115).    As Dr. Budowle well knows, however, there is a substantial difference between the somewhat academic question of whether two people might possess the same entire fingerprint pattern and the  significantly more important real-life issue concerning the frequency with which small distorted latent fingerprint fragments are mistakenly identified with the rolled impressions that they are compared with.  Dr. Budowle did not claim that the 100  years of fingerprint employment constitute  “empirical studies” as to the frequency of such errors.

            In the absence of any empirical studies concerning the subject of errors, Dr. Budowle expressed the view, contrary to that of the Supreme Court in Daubert,  that “calculating an error rate is meaningless and misrepresents the state of the art.”  (Id. at 163).  When asked specifically about the 1995 latent fingerprint examiner proficiency exam, on which 22% of the participating examiners made false identifications, Dr. Budowle testified, somewhat confusingly, that although he has “no experience in what was done in that particular situation,” he nevertheless knows that “in 1995 people weren’t using it in the proper fashion and design for proficiency testing.”  (Id. at 170).¹²

            Dr. Budowle also attempted to analogize the opinion of a latent fingerprint examiner to a diagnosis rendered by a medical doctor.  (Tr. 7/13 at 86-87).  In making this comparison, however, Dr. Budowle failed to address two significant distinctions between medical doctors and latent fingerprint examiners; the first being the extreme level of training and testing that medical doctors are put through before their opinions are deemed sufficently reliable to be trusted; the second distinction being that doctors do not make the inherently unscientific claim of absolute certaintity when rendering their opinions.  In addition, Dr. Budowle failed to support his position that a medical doctor’s opinion would be considered “scientific” if that opinion was not based on any controlled studies or experimentation that had previously been done in the particular field.

            Dr. Budowle also analogised the comparison of fingerprints to the comparison of DNA strands.  (Id. at 86).  But, in drawing this analogy, Dr. Budowle simply ignored his earlier testimony that with respect to DNA there has been a plethora of testing of different population groups so as to determine the statistical probability of different people having the same DNA. (Tr. 7/9 at 150, 151).  As Dr. Budowle well knows, there has been no comprable testing done with respect to fingerprints.

            Finally, Dr. Budowle also vouched for the AFIS experiments conducted by Lockheed Martin.  However, after hearing the testimony of defense expert David Stoney,  Dr. Budowle, in his rebuttal testimony, acknowledged that “no one” would say that these tests have “prove[d] uniqueness” of fingerprints.  (Tr. 7/13 at 82).  Moreover, in opining that these tests have any value at all, Dr. Budowle revealed his fundamental misunderstanding of the tests, which is surprising given that he helped to design them.  (Tr. 7/9 at 116, 177).  Dr. Budowle testified that Lockheed should not be criticised for comparing prints with themselves, because this was simply done as a “quality control measure.”  (Tr. 7/13 at 80).   To the contrary, Mr. Zeisig’s testimony made clear that the astronomical probabilities provided for by these tests, concerning the likelihood of two fingerprints, or two fingerprint subsets, being “identical,” are directly derived from the scores that AFIS generated when each fingerprint image was compared with itself.   Thus, far from being a simple quality control measure, the comparison of each fingerprint image with itself effectively provided the standard for “identical” by which all other fingerprint comparisons were measured.   Dr. Budowle did not explain during his testimony how these experiments tell us anything about uniqueness given that Mr. Zeisig conceded that rolled fingerprints of the exact same finger would not meet the standard of identical which Lockheed had adopted.¹³

                        (vii)  Pat Wertheim

                        In addition to Dr. Budowle, the government also called Pat Wertheim as a rebuttal witness.   Wertheim, who has more than 20 years of experience as fingerprint examiner with various state law enforcement agencies, testified that he has given numerous training courses to fingerprint examiners around the country.  (Tr. 7/13 at 53-56).  He testified that the trainees at these courses would commonly be “reluctant to embrace the philosophy and the methodology of ridgeology,” but that they would ultimately conclude that the ridgeology methodology was consistent with what they had “been doing all along.”  (Id. at 58).   Wertheim did not testify, however, as to whether his trainees had advised him as to whether they were adhering to point standards when making identifications.¹⁴

            Wertheim acknowledged that with respect to fingerprint identifications, verification is usually done by an examiner who not only works with the examiner who initially compared the prints, but who also knows the first examiner’s conclusion.  (Id. at 62).  Wertheim agreed that such verifications are essentially a “cleanup or checkup confirmation process.”  (Id. at 62).  In recounting an experience where he served as a defense expert in England, Wertheim revealed that when he wanted a true verification of his opinion, he provided the prints in question to a colleague, without revealing the opinion that he had formed.  (Id. at 59, 61).

                        B.         The Defense Witnesses 

            The defense called four witnesses at the Daubert hearing, three experts and an investigator.   The three experts all agreed that latent fingerprint identifications are not scientific.   Among other things, each of the defense experts pointed to the lack of testing that has been done in the fingerprint field and the failure of the fingerprint community to develop an objective identification standard.   Each agreed that the case work that has been performed by fingerprint examiners over the past 80 years is no substitute for scientific testing.

                        (i)  David Stoney

                        The first expert witness called by the defense was Dr. David Stoney.  Dr. Stoney is currently the Director of the McCrone Research Institute, a not-for-profit teaching and research institution located in Chicago, Illinois.  (Tr. 7/12 at 36, 37).  Dr. Stoney was awarded his Ph.D. in forensic science in 1985 from the University of California at Berkeley.  (Id. at 36).  His dissertation concerned a quantitative assessment of fingerprint individuality.  (Id. at 36).  It included a statistical model for fingerprint individuality that Dr. Stoney personally created.  (Id. at 66).  This work was subsequently recast into a peer reviewed journal article, entitled, A Critical Analysis of Quantitative Fingerprint Individuality Models, 31 Journal of Forensic Sciences, 1187 (1986), which Dr. Stoney coauthored with Dr. John Thornton.   Dr. Stoney has authored approximately 20-25 peer reviewed publications in the forensic science area, more than 10 of which have concerned the area of fingerprints.  (Id. at 40-42).  One of these publications is a book chapter concerning the scientific status, or lack thereof, of fingerprint identifications which is included as part of a West publication entitled Modern Scientific Evidence: The Law and Science of Expert Testimony (David L. Faigman et al. eds., West 1997) (Ex. 15).

            Dr. Stoney is himself a trained fingerprint analyst.  (Id. at 56-58).  He received his training as part of his forensic science education at Berkeley.  (Id.).   After graduating from Berkeley, Dr. Stoney worked for several years at a private criminalistics laboratory in California.  (Id. at 39, 40).  Approximately one third of his work there involved fingerprint comparisons.  (Id. at 40).  Dr. Stoney has previously been qualified to testify as an expert witness concerning fingerprint identifications.  (Id. at 45).

              At the Daubert hearing in the case at bar, Dr. Stoney testified that latent fingerprint identifications are not scientific.  Rather, a fingerprint examiner’s opinion of identification is a “subjective determination, without objective standards.”  (Id. at 87).¹⁵   Consistent with the Supreme Court’s decision in Daubert, Dr. Stoney testified that for a technique, such as latent fingerprint identification, to be considered scientific, the technique must be tested.  (Id. at 87-89).   Dr. Stoney explained two ways such testing could be done with respect to fingerprints.  (Id.).  First, an objective identification standard could be proposed and that standard then tested so as to determine the degree to which identifications meeting the criteria of the standard are correct or incorrect.  (Id.).  Second, a process could be proffered, such as the ACE-V process testified to by the government’s witnesses, and that process then tested so as to determine the degree to which examiners utilizing the process are producing correct answers.  (Id. at 87, 89, 102).

            Dr. Stoney provided a specific example of a test that might be done in this regard.  Different fingerprint examiners could be asked to compare fingerprints which, though deposited by different fingers, nevertheless contain several common ridge characteristics.  (Id. at 105).  An example of such prints was provided in defense exhibit 6, a journal article discussing a latent fingerprint that had seven characteristics in common with a rolled fingerprint impression that had been taken from a different person.  There was some disagreement among the government experts as to the likelihood of a fingerprint examiner making a misidentification when confronted with prints of this nature.  (Tr. 7/7 at 195, 196; Tr. 7/8 at 224-226).  As Dr. Stoney recognized, it would be a very simple and interesting study to provide these prints to different examiners so as to see how  many would actually make misidentifications.  (Tr. 7/12 at 105) (“That would be applying science to the issue; testing.”).

            Dr. Stoney testified that the casework that fingerprint examiners have been doing over the past eighty years cannot be considered a substitute for scientific testing.  (Id.  at 120, 121).  There is no simply no way of knowing, Dr. Stoney testified, how often fingerprint examiners have erred in their casework.  (Id. at 121).  In order to accurately assess whether fingerprint examiners are producing correct opinions, it is necessary to test them by providing them with prints that are known ahead of time to have either arisen or not arisen from the same source.  (Id.).

            Dr. Stoney disagreed with Dr. Budowle’s position that  practitioner error rate is irrelevant and that the error rate for the ACE-V methodology is 0. (Id. at 102-104).  As Dr. Stoney recognized, and as the government experts conceded, the ACE-V process is entirely dependant upon a practitioner’s “subjective determination.”  (Id. at 103).    Accordingly, since “the individual is an inherent part of getting to the opinion in this process [,]... errors that individuals make are a very important part of evaluating whether or not it works.”  (Id. at 104).    Dr. Stoney testified that the claim of a zero error rate differentiates fingerprints from all of the other forensic sciences.  (Id. at 103).   

            Dr. Stoney also disagreed with Dr. Budowle as to whether the statistical probability models that he (Stoney) and others have created provide a scientific basis for fingerprint identifications.  (Id. at 119, 120).  Dr. Stoney testified that none of these models, including his own, have ever been tested.  (Id. at 120).   Accordingly, he testified that the models, as they now stand, are simply “reasonable guesses or speculations ... as to what might be a model that would apply to the individuality of fingerprints.”  (Id.).  Moreover, none of these models concern the reliability of identifications that are based on small distorted latent fingerprint fragments.

            As to the embryology studies that some of the government witnesses referred to, Dr. Stoney testified that these studies have nothing to do with the fundamental question of what constitutes a sufficient basis to make a reliable latent fingerprint identification. (Id. at 106,107).   What these studies address, Dr. Stoney testified, is  the issue of how “friction ridges come to be on the fingers.”  (Id. at 92).    Dr. Stoney testified that while these studies provide important “background information,” they in no way make latent fingerprint identifications scientific, since they neither tell us when identifications should be made or when they are correctly made.  (Id. at 107).

            Finally, Dr. Stoney harshly criticized the government’s AFIS experiments.  (Id. at 109-119).  He testified that the methodology of these experiments was “fundamentally flawed” in that the standard for “identical” was derived by comparing a fingerprint image with itself.  (Id. at 110, 111).   It is a basic element of forensic science, Dr. Stoney testified, that no two representations of anything, be it a person’s signature or his fingerprints, will be exactly alike.  (Id. at 111).    Dr. Stoney testified that the experiments, therefore, do not “mean anything” with respect to the issue of whether fingerprints are unique or the question of what constitutes a sufficient basis to make an identification.  (Id. at 113).  This was made clear, Dr. Stoney testified, by Mr. Zeisig’s admission that fingerprints taken from identical fingers would not meet the definition of identical employed in the experiments.  (Id. at 110).  Dr. Stoney testified that the conclusions that have been drawn from these experiments, as expressed in the Lockheed Martin test report, are very misleading:

Q.  [The first conclusion] states that:  The probability of a non-mate rolled fingerprint being identical to any particular fingerprint is less than 1/10 to the 97th.

                                    Do you have any opinions as to the way that this conclusion is stated?

 

A.  Well, if this is meant as a comment or to have anything to do with the notion of how common is it that people either have identical skin patterns on their finger or if it is meant to comment on the process of will different prints from the same person how common will they appear to be the same by either judgment that we are talking about here in terms of an examiner’s opinion, this is completely irrelevant to those determinations.  By completely irrelevant to it, I mean that it doesn’t make it to this point in the exhibit where I was describing earlier as foundation for my opinion.  If I were to use this as a foundation for my opinion.  I would be grossly in error in making that statement.

It’s so fundamentally inaccurate to do that type, so as it sits here in this report, it causes me grave concern.  I consider it very misleading. . ..   .   .  

Q.  [The second conclusion] states that:  The probability of a minutia subset of a non-mate fingerprint being identical to a minutia subset of any particular fingerprint is less than 1/10 to the 27th for small numbers of minutiae, in this case, small means four, decreasing to less than one in 10 to the 27th for larger numbers of minutiae.

Do you have any comments on the way that this Experiment 2 Conclusion is stated?

A.  Well, I have already given my opinion that I feel these experiments have -- are well removed from any kind of forensic science, from describing any process in forensic science.  I mean, so to the degree that a person would read that and think that it did or read into this, this is the probabilities of a comparison coming out wrong or this is a random probability of encountering a particular minutia configuration, that would -- it does not present that to us. 

Q.  So one cannot properly read from this that the chance of two different people having four minutia points in common is in 1 in 10 to the 27,000? 

A.  Absolutely not. 

(Id. at 113-114; 118-119). 

                        (ii)  James Starrs

                        The defendant next called Professor James E. Starrs, a Professor of Forensic Sciences at The George Washington University, the Columbian School of Arts and Sciences, and a Professor of Law at The George Washington University Law School.  (Ex. 50).  Professor Starrs has held these positions for the past 30 years and is one of the founding fathers of the Forensic Sciences department at George Washington University.  (Tr. 7/12 at 123, 124).  Professor Starrs has published more than 80 articles in the forensic science area and he is a co-author of the leading text in the field: Scientific Evidence in Civil and Criminal Cases (4th ed. 1995).  (Id. at 127, 128).  Among his many notable honors and achievements, Professor Starrs is a Distinguished Fellow of the American Academy of Forensic Sciences, a distinction that only 24 other persons have received.  (Id. at 133,134).  Professor Starrs testified that fingerprints has been one of his major areas of interest and study throughout his lengthy career.  (Id. at 125, 126).

            In the opinion of Professor Starrs, there is not a scientific basis for latent fingerprint identifications.  (Id. at 151).  Like Dr. Stoney, Professor Starrs pointed to the lack of testing that has been done in the field.  (Id. at 151, 154).   He pointed out, for example, that there has been no  experimentation done to determine how many different people might have the same fragmentary print in common.  (Id. at 156).  Nor has testing been done to determine the different assessments that different examiners might make when presented with the same latent print to analyze.  (Id. at 154).¹⁶

            Closely associated with the lack of testing, Professor Starrs also pointed to the failure of the fingerprint community to establish an error rate.  (Id. at 157).   He testified that it is “scientific balderdash” for the FBI to claim that fingerprints are infallible and that there is a zero error rate.  (Id. at 161).  (“The infallibility of fingerprinting is only as fallible or infallible as the one conducting the examination.”).   Moreover, he stated that if the error rate is actually as high as is indicated by the 1995 latent print examiner proficiency exam, on which there was a 22% false identification rate, than “we have got more than a fly  in the ointment, we have a bee hive in the ointment of fingerprint analysis.”  (Id. at 157).            

            In opining that latent fingerprint identifications are not scientific, Professor Starrs also pointed to the lack of standards in the fingerprint field.  (Id. at 161-165).  He testified that the various fingerprint examiners that he spoke with in advance of the hearing advised him that they are continuing to use point standards rather than the ridgeology approach espoused by the government’s witnesses.  (Id.).   Moreover, Professor Starrs testified that there is no standard classification or terminology with respect to the basic ridge characteristics.  (Id. at 161, 162).  Accordingly, as seen in this case (see supra at 5), what some examiners might count as one characteristic, another examiner might consider to be two.  (Id.).

              Finally, Professor Starrs pointed to the fundamental lack of skepticism that pervades the fingerprint community.  (Id. at 167-169).   The essence of science, Professor Starrs testified, is the willingness to consider that your hypothesis may be wrong and to attempt through testing to falsify it.  (Id. at 157, 169).  Such a willingness is completely missing in the fingerprint field.  (Id.).  Rather, Professor Starrs testified, you have fingerprint examiners providing their subjective opinions of identification in the inherently unscientific terms of absolute certainty.  (Id. at 152, 153). 

                        (iii)  Simon Cole