FILED
United States Court of Appeals
Tenth Circuit
July 20, 2009
PUBLISH Elisabeth A. Shumaker
Clerk of Court
UNITED STATES COURT OF APPEALS
TENTH CIRCUIT
UNITED STATES OF AMERICA,
Plaintiff - Appellee,
v. No. 08-2098
ROBERT ABDUL BAINES,
Defendant - Appellant.
APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE
DISTRICT OF NEW MEXICO
(D.C. No. 2:06-CR-01797-MV-3)
Terri J. Abernathy, Assistant United States Attorney, (Gregory J. Fouratt, United
States Attorney, with her on the brief), Las Cruces, New Mexico, for Plaintiff -
Appellee.
Zachary A. Ives, Freedman Boyd Hollander Goldberg & Ives P.A. (Molly
Schmidt-Nowara, with him on the briefs), Albuquerque, New Mexico, for
Defendant - Appellant.
Before McCONNELL, HOLLOWAY, and BALDOCK, Circuit Judges.
HOLLOWAY, Circuit Judge.
After a jury trial in federal district court, defendant-appellant Robert Abdul
Baines was convicted on five counts: conspiracy to possess marijuana with intent
to distribute; possession of marijuana with intent to distribute; possessing a
firearm in furtherance of a drug trafficking crime; possession of a firearm after
former conviction of a felony; and possession of ammunition after former
conviction of a felony. He was sentenced to a total of 123 months of
imprisonment to be followed by three years of supervised release, and he was
ordered to pay immediately a special fee assessment of five hundred dollars.
We have jurisdiction under 28 U.S.C. § 1291. Concluding that the district
court did not abuse its discretion in allowing the government to present expert
evidence that a thumb print found on some of the contraband recovered by the
authorities was a match to Baines’ print, we affirm the judgment of the district
court.
I
Because the sole issue Baines raises in this appeal is the admissibility at
trial of fingerprint analysis as expert testimony, a brief overview of the facts
underlying the convictions, as established in the trial testimony, will suffice to
provide context for our discussion.
Baines recruited two young women to travel with him and two male friends
from Pennsylvania to Arizona, offering the women $1,000 each for the trip. Both
young women testified at trial that they realized that the purpose of the trip was to
2
transport drugs, although Baines did not tell them any details. At one point
during the trip Baines said he was “in the business,” which they understood to be
a reference to drug trafficking. The travelers used two rented vehicles for the
trip, a “tannish” Ford sedan and a minivan. The two women and Baines were in
the Ford for the trip to Arizona, with the women taking turns driving, while the
other two men were in the minivan.
Upon arriving in Phoenix, the group spent the night in a motel and went
shopping for awhile the next day. Baines contacted someone he referred to as
“Felix,” and had one of the young women take down directions to a place where
they would meet with Felix in Tucson. After the group spent part of the day at a
place the women assumed to be the residence of Felix, the three men left in the
Ford. Upon their return, the young women were told to get ready for the return to
Pennsylvania. They were both seated in the Ford when they saw Baines approach
carrying a blue bag. The woman in the driver’s seat unlatched the trunk, and both
could feel things being put in the trunk and moved around, though neither of them
saw what was happening with the trunk lid open. For their return journey, Baines
chose to ride in the minivan, leaving the two women as the only occupants in the
Ford.
Baines had told the two young women to plan a route for their trip back to
Pennsylvania. One of them decided that they should go through Texas rather than
return the way that they had come. That decision proved fateful because this
3
route took them through Las Cruces, New Mexico, and near there they
unexpectedly entered a border checkpoint.
When the trunk of the Ford was opened at the border checkpoint, the agent
immediately noticed the scent of fresh marijuana. The car was taken to the
secondary inspection area, where a dog trained to detect illegal drugs alerted to
the trunk. One of the agents had asked one of the women about the van that was
behind them at the checkpoint and learned that the two vehicles were traveling
together. One of the occupants of the van confirmed that fact. Accordingly, the
van and its three occupants were also directed to the secondary inspection center.
Officers found packages of marijuana in each of three bags, two laundry bags and
a black duffle bag, and the black duffle bag also contained two pistols and
ammunition. The two women were arrested.
When it appeared that the men were going to be released, one of the young
women decided to tell the officers about defendant’s role in arranging the trip and
his apparent role in acquiring and loading the marijuana. Both women also told
the agents that during the trip defendant had said that he was “in the business.”
With the testimony of the two young women and other evidence, such as
records of calls made from cell phones, the jury was persuaded to convict Mr.
Baines of the drug counts. But neither of the women had seen Mr. Baines with
the black duffle bag in which the guns and ammunition were found, and the
officers had received information that another one of the men owned a bag of that
4
description. So the government relied on fingerprint evidence to connect Baines
with the guns and ammunition.
Two fingerprints were discovered on one of the magazines found with the
two pistols in the black duffle bag. Defendant filed a motion before trial to bar
the government from presenting evidence that a fingerprint specialist had
determined that one of the recovered “latent” prints matched the “known”
fingerprint of defendant Baines. (See infra for an explanation of the two terms in
quotation marks.) Defendant’s motion invoked Rules 104(a) and 702 of the
Federal Rules of Evidence and Daubert v. Merrill Dow Pharm., Inc., 509 U.S.
579, 593-94 (1993). Defendant requested a pretrial hearing on the admissibility
of the government’s expert testimony, and the district court granted the request.
II
A
Two witnesses testified at the pretrial hearing on defendant’s motion to
exclude the fingerprint evidence, Mr. Fullerton, the state-employed fingerprint
examiner who later testified at trial, and FBI Agent Meagher, who is a fingerprint
specialist with the bureau. 1 Agent Meagher’s testimony was wide-ranging,
1
Both expert witnesses had years of training and experience. Defendant did
not challenge their qualifications as experts in the district court, nor does he on
appeal.
5
explaining basic concepts underlying fingerprint identification, the procedure
followed by fingerprint examiners, and branching out from there to respond to
inquiries aimed at some of the factors suggested in Daubert as relevant to the
consideration of expert testimony.
Agent Meagher defined “fingerprint” as the “ridged skin which appears on
the palmar side of the hand for each of the fingers.” He used a photograph to
demonstrate that the ridges are visible on the hand. He explained that the ridge
pattern can then be “transferred to an object when it is touched, or intentionally
recorded on a known fingerprint card.”
To explore the issues involved in fingerprint identification, it is first
necessary to understand the difference between what the witnesses called latent
prints and known prints. A known print is the kind that is made intentionally, as
when a person is arrested. Law enforcement agencies and others taking
fingerprints will attempt to get full prints of each finger. Agent Meagher
explained that this seemingly simple task is actually not so simple; practice and
training are needed to develop the skill of recording prints to obtain a clear and
complete image. In previous years prints were normally taken by applying ink to
the fingers and then applying the fingers to a paper card with a rolling motion. In
recent years, some agencies have adopted a digital photo scanning technique in
place of the old method. Even with trained personnel recording the prints, the
quality of known prints varies substantially.
6
Latent fingerprints are partial prints like those found at crime scenes and
often are invisible to the naked eye. One study determined that a latent print is
only, on average, about 22% of a known print. III R. 52. The gist of defendant’s
challenge is that the government in this case did not establish that the method for
matching the latent print at issue with defendant’s known print was reliable.
The field of fingerprint identification ultimately rests on two premises: that
each individual’s fingerprints are unique and that the unique pattern of a person’s
prints does not change over time. These basic principles are essentially
unchallenged in this appeal. Nor does defendant contest that the latent print
found on the magazine in this case was accurately reproduced for analysis.
Defendant’s challenge is to the reliability of the process of comparing the latent
print to known prints.
Agent Meagher described the approach used in fingerprint comparison.
The first step a fingerprint examiner takes is a close observation of the
characteristics of the latent print under study. A latent print may have three
levels of detail. The first level is the “ridge flow” or pattern of the ridges. There
are three basic patterns, known as arch, loop, and whirl. An “individualization” 2
cannot be based on this level alone although a decision to exclude a candidate
2
Agent Meagher explained that in the “fingerprint discipline” the term
“individualization” is used (instead of “identification”) and that the term means
that the examiner has identified the “donor” of the latent print to the exclusion of
all other possible donors. III R. 24-25.
7
may be made on Level 1 detail alone.
The second level of detail is “the ridge path of the individual ridges.” Id.
at 26. The examiner chooses ridges on the print, follows them, and makes
observations. For example, the examiner may note points where a ridge ends or
divides into two ridges. Agent Meagher testified that an examiner can
“individualize” or establish an identity at this point. The third level of detail is
observed by “zooming in” more closely to gather additional information about an
individual ridge, including features like sweat pores and differences in size and
shape of the ridge.
The process used for determining whether a latent print matches a known
print has been given the acronym ACE-V, with the letters standing for the steps in
a four-stage process: (1) analysis, (2) comparison, (3) evaluation, and (4)
verification. In the initial analysis step, the examiner looks at the latent print and
the known print separately. The purpose of this step is to discern characteristics
at all three levels of detail and to evaluate the quality and quantity of information
on each print. The examiner may find a disparity in characteristics that compels
the conclusion that the prints cannot be a match, or may find that one or the other
– usually the latent – is of too poor a quality or simply reveals too little
information for further examination to be fruitful. If, however, the examiner
determines that there is nothing to exclude the possibility of a match and that the
quantity and quality of the information is sufficient, then he moves to the next
8
step.
The second step in the process is comparison, a side-by-side examination of
the latent print and the known print. The examiner looks for reasons to exclude
the known print and for similarities between the two. As Mr. Fuller later
explained in his trial testimony, the examiner at this stage is determining if there
could be a match between the known print and the latent print. Id. at 389. If it
appears possible that they might match, the examiner goes to the third step,
evaluation, where the examiner actually tries to reach a conclusion as to whether
there is a match or not.
Verification, the fourth stage, involves having a second examiner look at
the prints being compared. In this appeal, defendant stresses that the verification
process is not truly independent. Not only is the second examiner usually with
the same law enforcement agency, but in this case at least the second examiner
did not conduct a “blind” comparison, but rather was given all the work notes and
other work product of the first examiner.
Agent Meagher was asked about the error rate for friction ridge
identification. His answer (as to many questions during the hearing, all without
objection) was a rambling narrative covering almost six pages of transcript. He
began by positing that there are two types of errors, practitioner error and
methodological error. He then pronounced by his ipse dixit that the subject of the
hearing was methodological error, not practitioner error, and that the error rate for
9
the method was “either no error, or it’s a zero error.” III R. 87. He went on to
acknowledge that practitioners do make mistakes, but then asserted that the
“practitioner error rate goes to the individual, not to the whole of the practitioners
applying the methodology.” It would be “inappropriate,” he testified, to “take the
accumulation of those who have made errors and assign it to those who have not
made errors,” thus at least implying that most practitioners have achieved a level
of perfection that is rather rare, to say the least, in other complex human
endeavors. Agent Meagher did go on to cite one published report in which 92
participants performed a total of 5,861 individualizations, out of which there were
two errors, both of which were noticed and corrected by verifiers.
On cross-examination Agent Meagher testified that the FBI has no statistics
from which error rates of its analysts could be calculated. He said that each
analyst would know his or her error rate from the proficiency examination taken
at the end of training and annually thereafter. With respect to errors in actual
cases, Mr. Meagher first explained that there were three possible types of errors
to consider: false or mistaken identifications in which the analyst incorrectly
identifies a person as the source of a latent print; missed identifications, where
the analyst fails to make an identification when she should have; and clerical
errors. Of these, the first is the proper focus for the court, “the only error of
consequence,” the agent testified. As to these “false positives,” Meagher testified
that the FBI had “made, on average, about one erroneous identification every 11
10
years.” The total number of identifications made has been about one million per
year, he continued, so that the known actual error rate was about one per eleven
million identifications. Id. at 99. He further testified that he knew of no
erroneous identifications in proficiency testing of the FBI’s examiners in the last
ten years that he has been in a managerial position to have access to that
information. There were one or two missed identifications during the ten-year
period.
The second witness at the Daubert hearing was Mr. Fullerton of the New
Mexico forensic lab, who conducted the actual process in this case. His
testimony covered some of the same ground as Agent Meagher’s in describing the
ACE-V procedure in general, for example. Fullerton also testified that he was
able to conclude that the latent print at issue matched the known left thumb print
of defendant Baines.
Mr. Fullerton testified that he could not even determine the basic pattern of
the latent print at level 1 because the left side of the print was not available. The
latent print was an impression of such a small portion of the print that Fullerton
could not say whether it was a part of a left slant loop or a whorl. Id. at 123-24.
Moreover, Fullerton testified, in this case the known prints were also of poor
quality. However, on re-direct examination at trial he further explained that,
although the prints of several fingers on Baines’ card were very poor images, the
11
left thumb print was of good quality and could be matched to the latent print. 3 Id.
at 416.
Notwithstanding these challenges, Fullerton testified, he was able to
conclude that the latent print was from Mr. Baines’s left thumb, based on eleven
points of comparison. Verification was accomplished by giving the data – and
Fullerton’s marks on the copies of the prints and other work notes – to another
examiner in the same lab. Fullerton admitted that this was not an “independent
identification” and that his sharing of his work product with the second examiner
had suggested findings.
The defense presented no witnesses at the hearing.
B
The district court’s ruling.
The district judge issued her ruling from the bench, followed later by a
Memorandum Opinion and Order. In her written order the judge concisely
summarized the parties’ arguments and the testimony of Agent Meagher of the
FBI and Mr. Fullerton of the state crime lab. Although the judge summarized this
testimony and appears to have accepted it, she did not make formal findings. As
3
A copy of the fingerprint card was admitted in evidence at both the
Daubert hearing and at trial, and a copy of the exhibit is appended to the briefs.
By looking at this exhibit, one can see that the image of the left thumb print is
clearer than that of most of the other fingers, an observation that the jurors had
the opportunity to make for themselves.
12
he does in this appeal, the defendant argued that the methods of fingerprint
analysis have not been tested; that there are no established error rates; that
fingerprint examiners do not adhere to uniform, objective standards; and that
there is an absence of professional literature to support admission of testimony of
fingerprint examiners.
The district judge did not directly address these arguments by defendant’s
counsel. The judge did say that her decision was based on the testimony of Agent
Meagher, thus at least implicitly making findings consistent with that testimony.
The judge concluded that government had met its burden and “shown by a
preponderance of the evidence that the reasoning and methodology underlying
latent fingerprint evidence is scientifically valid and was properly applied by Mr.
Fullerton to the facts at issue in this case.” The judge also noted her agreement
with the cases cited by the government in support of its position. Accordingly,
she held that the evidence was shown to be relevant and reliable, meeting the
requirements of Fed. R. Evid. 702. In closing, the judge addressed the core of
defendant’s argument, that fingerprint analysis rests substantially on the
subjective interpretations of the examiner. The judge said that this argument went
to the weight of the evidence, not its admissibility, and she quoted Daubert’s
observation that “[v]igorous cross-examination, presentation of contrary evidence
and careful instruction on the burden of proof are the traditional and appropriate
means of attacking shaky but admissible evidence.” Daubert, 509 U.S. at 596.
13
At trial, Mr. Fullerton testified to his opinion that the latent print from the
magazine matched the known left thumb print of defendant Baines.
III
As noted, the only issue in this appeal is whether the trial court should have
excluded the fingerprint evidence. As with other evidentiary rulings, we review
the district court’s decisions to admit expert testimony only for abuse of
discretion. See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999). The
issue in this case is whether the district court properly fulfilled its duty, as
established in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 597
(1993), to ensure that expert testimony “rests on a reliable foundation and is
relevant to the task at hand.” 4
General principles.
Expert testimony is admissible only if it is potentially helpful to the jury
and “(1) the testimony is based upon sufficient facts or data, (2) the testimony is
the product of reliable principles and methods, and (3) the witness has applied the
principles and methods reliably to the facts of the case.” Fed. R. Evid. 702. The
burden of proof is on the proponent of the evidence, here the government.
The Court has suggested some factors, which are not necessarily
4
The district judge has discretion also in determining how to perform the
“gatekeeper” function mandated by Daubert, see Kumho Tire, 526 U.S. at 152,
but no contention is made in this appeal that the judge abused her discretion in
that regard. The only contention is that the judge reached the wrong conclusion.
14
exhaustive, that will be helpful to the trial courts in determining whether
proposed expert testimony is based on reliable methods and principles: (1)
whether the particular theory can be and has been tested; (2) whether the theory
has been subjected to peer review and publication; (3) the known or potential rate
of error; (4) the existence and maintenance of standards controlling the
technique’s operation; and (5) whether the technique has achieved general
acceptance in the relevant scientific or expert community. Daubert v. Merrill
Dow Pharm., 509 U.S. 579, 593-94 (1993). These factors do not constitute a
“definitive checklist or test.” Id. at 593. The gatekeeping inquiry must be “tied
to the facts of a particular case.” Id. at 591. The factors “may or may not be
pertinent in assessing reliability, depending on the nature of the issue, the
expert’s particular expertise, and the subject of his testimony.” Kumho Tire, 526
U.S. at 150.
Daubert was limited to scientific evidence. In Kumho Tire Co., the Court
held that the district courts’ “gatekeeping” obligation as described in Daubert
applies to all expert testimony and that in performing this function in particular
cases, the district courts may consider the specific Daubert factors to the extent
relevant. The Court specifically noted that “no clear line” divides “scientific”
and “technical or other specialized” knowledge, all of which are treated together
under Fed. R. Evid. 702. 526 U.S. at 148.
Defendant’s argument.
15
Defendant discusses each of the Daubert factors. First, defendant asserts
that fingerprint identification lacks objective standards and so must rely largely
on the subjective impressions of the individual examiner. Agent Meagher
admitted that the FBI does not use any objective standard for the number of
similarities between a latent print and a known print necessary to make a match,
and also that the subjective views of the individual examiner play a significant
role in the process.
Meagher identified only two standards, both of which involve subjective
determinations, defendant contends. First, Meagher testified that the examiner
must not find any discrepancy between the two prints. But Meagher explained
that “no discrepancy” really means no discrepancy without a “viable or plausible
or valid explanation,” and whether an explanation meets that amorphous standard
is a subjective judgment by the examiner. III R. 103-04.
Second, Meagher testified that there must be “agreement of sufficient
friction ridge details in sequence,” but again it is up to the examiner to determine
what is “sufficient.” Defendant cites one scholar who opines that the fingerprint
community has been unable to answer the “crucial question” of “where the
boundary lies between insufficient and sufficient correspondences.” Simon A.
Cole, More Than Zero: Accounting For Error in Latent Fingerprint
Identification, 95 J. Crim. L. & Criminology 985, 993-94 (Spring 2005) [hereafter
Cole, More Than Zero]. Indeed, one appellate court that held the evidence
16
admissible found that this factor weighed against admissibility. See United States
v. Mitchell, 365 F.3d 215, 241 (3d Cir. 2004).
Turning to the other Daubert factors, defendant contends that the
government failed to show that the process for latent fingerprint identification has
been tested. As one judge said, “there have not been any studies to establish how
likely it is that partial prints taken from a crime scene will be a match for only
one set of fingerprints in the world.” United States v. Crisp, 324 F.3d 261, 273
(4th Cir. 2003) (Michael, J., dissenting). 5
Defendant then turns to attacking the one survey and one study that the
government, through Agent Meagher, proffered as evidence of reliability. In the
survey, the FBI polled law enforcement agencies in all 50 states, the District of
Columbia, Canada and the United Kingdom and learned that none of these
agencies had ever found two different people with the same fingerprints and that
none of the agencies had ever found that a latent fingerprint had been identified
5
To clarify, both the dissenting judge in the case cited and counsel for
Baines seem to be questioning the likelihood of a mistaken match of a latent print
to a known print, rather than challenging the fundamental principle that each
individual’s prints are unique. Indeed, the dissent included this quote from a
treatise: “The issue is not the finding of two fingerprints that are alike, but rather
the finding of prints from two different fingers that can be mistakenly adjudged to
be alike by a fingerprint examination.” Mitchell, 324 F.3d 261, 274 (Michael, J.,
dissenting) (quoting David A. Stoney, Fingerprint Identification: The Scientific
Basis of Expert Testimony on Fingerprint Identification, in 3 Modern Scientific
Evidence: The Law and Science of Expert Testimony § 27-2.3.2 (2002)).
17
with two different people. 6 But, defendant says, this is not the same as saying
that latent prints had never been misidentified.
The government also relied on a statistical study commissioned by the FBI
and conducted by Lockheed Martin. Studying 50,000 prints and comparing each
by computer against every other one, this study confirmed to an extremely high
degree of probability that no two persons’ fingerprints are identical. 7 Again,
defendant responds that is not at issue here.
In the second part of this study, an attempt was made to simulate latent
prints by extracting about 20% of the data from each print and then comparing
these partial prints to every other print in the database. 8 The study concluded,
with a very high degree of certainty, that there is almost no chance of ever finding
two persons to have the same print, even when based on such partial prints. But
defendant points out that a leading case found these “pseudo-latent” prints are
“poor approximations of real latent prints.” Mitchell, 365 F.3d at 237. Because
the study did not adequately model real-world conditions, it does not provide
significant support for the government’s position, Mitchell held. Id. at 238.
6
The FBI conducted this survey in preparation for the exhaustive Daubert
hearing described in United States v. Mitchell, 365 F.3d 215, 223-225 (3d Cir.
2004).
7
In Mitchell, the court noted that this study was also done specifically in
preparation for the Daubert hearing in that case.
8
The researchers had determined that the average latent print has just over
20% of the image of a known print.
18
Nor has fingerprint identification been subject to peer review, defendant
continues. The government claimed that the verification step in the ACE-V
process is peer review, but defendant insists this is not accurate. The Court in
Daubert referred to a process that serves to assess the scientific validity of the
methodology, which is not accomplished merely by having two persons apply the
same technique, defendant argues. Consistency of results does not prove that the
results are valid.
Moreover, the verification at issue here is not truly independent. In fact, it
fails to show independence in two ways. Often, as here, the reviewer is
associated with the first examiner and both are employed by the same agency.
Second, unlike true peer review in the scientific process, the reviewer in this
system is not independent in that he receives all of the examiner’s work product,
rather than perform the analysis himself. Indeed, Mr. Fullerton admitted in his
testimony that this was not an “independent identification” and that giving his
work product to the verifier was suggestive. In a truly independent verification
process, the reviewer should not even know the conclusion of the first examiner,
much less all the steps taken on the path to that conclusion, defendant asserts.
Next, defendant contends that the government failed to show a meaningful
rate of error for latent fingerprint identification. Mr. Meagher testified that the
rate of error for latent fingerprint identifications is zero, yet he admitted that
innocent people have been convicted based on misidentification of their
19
fingerprints. Defendant cites one study that describes 22 cases of latent
fingerprint misidentification. Cole, More Than Zero, 95 J. Crim. L. &
Criminology at 985-987, 1001-16. Those include the much-publicized recent case
where the FBI identified a Portland lawyer as a suspect in the terrorist bomb
attack on the Madrid train station that killed 191 people, in spite of the fact that
the Spanish authorities insisted, correctly, that the fingerprints did not match. 9
Defendant criticizes Agent Meagher’s attempt to distinguish between
methodological error and practitioner error. Defendant argues that this is a false
and meaningless distinction. One scholar, expressing the same view, said that
because the method “depends so heavily on subjective human judgment . . . the
method literally is the people who employ it.” Jonathan J. Koehler, Fingerprint
Error Rates and Proficiency Tests: What They Are And Why They Matter, 59
Hastings L.J. 1077, 1090 (May 2008). In any event, defendant goes on, the
purported distinction is irrelevant under Daubert.
The government produced no evidence, defendant says, about error rates in
real-world cases. Mr. Meagher admitted that the FBI’s only actual error rates are
based on proficiency tests the examiner candidates take under controlled
conditions. The FBI does not compile error rates for examinations in real cases.
9
The incident is described in Cole, More Than Zero, 95 J. Crim. L. &
Criminology at 985-87, where the author notes that the mistaken identification
was made by a total of four examiners, each of whom had considerable experience
in the field.
20
As Judge Michael observed in his Crisp dissent, “where tests have attempted to
imitate actual conditions, the error rates have been alarmingly high.” 324 F.3d at
275. 10
In sum, defendant contends that the error rate is not zero, and the
government failed to establish an actual error rate. Moreover, Baines argues that
the effort by fingerprint examiners to create an aura of infallibility has the
potential to seriously mislead jurors.
Finally, Baines argues that the government failed to show that fingerprint
identification has been generally accepted in any unbiased scientific or technical
community. It is not enough, according to him, that courts have accepted the
technique.
The government’s argument.
The Daubert inquiry is a flexible one, the government notes, and the factors
that case set out “do not all necessarily apply” in every case. Kumho Tire, 526
U.S. at 150-51. Every published decision to address this issue has found the
evidence admissible. Fingerprint evidence has been admissible in this country for
almost 100 years. The government urges this court to adopt the reasoning of
United States v. Mitchell, 365 F.3d 214 (3d Cir. 2004), which it says was based
10
Defendant does not directly challenge Meagher’s claim that the FBI
analysts have cumulatively made only one mistaken identification every eleven
years or one per each eleven million cases.
21
on substantially the same expert testimony, chiefly from Agent Meagher, that was
presented in this case.
First, the Mitchell court found that the theories underlying fingerprint
identification – that fingerprints are unique and permanent, and that identification
matches can be made from fingerprints containing sufficient detail – are testable
and have actually been tested by experience. On the second Daubert factor, the
Mitchell court found that the ACE-V protocol constituted peer review and
weighed in favor of admission.
The Mitchell court then considered the factor of established error rate for
the procedure. Although a precise error rate has not been established, the court
found that various estimates of the error rate all suggested that it was very low.
The court cited evidence which the government characterizes as indistinguishable
from the evidence in this record: the absence of significant numbers of false
identifications in practice, the absence of “false positives” in an FBI survey of
state agencies, and the Lockheed study discussed supra. Agent Meagher testified
that the FBI’s own monitoring has revealed approximately one false identification
every eleven years. About one million comparisons per year were made, he said,
so that the error rate was approximately one for every 11 million comparisons.
On the fourth Daubert factor, the government points to testimony from
Agent Meagher that the ACE-V procedure is a widely accepted standard
governing operation of the methodology. Meagher also testified that additional
22
standards for conclusions are set by the Scientific Working Group on Friction
Ridge Analysis, Study and Technology (SWGFRAST), a professional group.
Meagher testified that the standards for positive identifications of the latter group
included agreement of sufficient friction ridge details in sequence, as determined
by a competent examiner, and applied to “common area and both impressions” (a
phrase that was not explained), and absent any discrepancies. III R. 94.
Finally, the Mitchell court found that general acceptance in the fingerprint
community weighed in favor of admissibility. The court rejected the argument
Baines makes here – that the community is not an impartial, scientific community.
The Mitchell court held that, after Kumho Tire, this distinction is irrelevant. The
general acceptance test of Frye v. United States, 293 F. 1013, 1014 (D.C. Cir.
1923), merely asked whether the proposed method has “gained general acceptance
in the particular field in which it belongs.” In Mitchell, the court said that,
although Daubert held that Rule 702 had legislatively overruled Frye, at the same
time the Court had acknowledged that “general acceptance” could still be a factor
in the inquiry. 365 F.3d at 241.
Analysis.
Our task is not to determine the admissibility or inadmissibility of
fingerprint analysis for all cases but merely to decide whether, on this record, the
district judge in this case made a permissible choice in exercising her discretion
23
to admit the expert testimony. 11 Although this record raises multiple questions
regarding whether fingerprint analysis can be considered truly scientific in an
intellectual, abstract sense, nothing in the controlling legal authority we are bound
to apply demands such an extremely high degree of intellectual purity. Instead,
courts applying Fed. R. Evid. 702, Daubert, and Kumho Tire, are charged only
with determining that the expert witness “employs in the courtroom the same
level of intellectual rigor that characterizes the practice of an expert in the
relevant field.” Kumho Tire, 526 U.S. at 152.
To begin our analysis, we will consider the parties’ arguments and the
record on the Daubert factors, with the understanding that they are not exclusive
and that expert testimony does not have to meet all of them to be deemed
sufficiently reliable. The inquiry is a “flexible one,” as Daubert itself teaches.
509 U.S. at 594. We also remain mindful that Daubert addressed evidence that
was claimed to be scientific. Kumho Tire held that the trial court’s gatekeeping
function applies to all expert testimony and noted that there is no clear line
separating “scientific” knowledge from technical knowledge or knowledge based
on experience. Nonetheless, the Court there said that “some of Daubert’s
questions can help to evaluate the reliability even of experience-based testimony,”
11
We note, however that appellant’s counsel focuses his argument almost
exclusively on fingerprint identification evidence in general, rather than on the
specific identification in this case by Mr. Fullerton.
24
526 U.S. at 151 (emphasis added), strongly suggesting that the Court realized that
some of the Daubert factors may be less helpful when the evidence under
consideration is not scientific in the strict sense. Although the importance of the
distinction is thus uncertain, we agree with the Third Circuit that fingerprint
analysis is best described as an area of technical rather than scientific knowledge.
Mitchell, 365 F.3d at 234.
The first Daubert question is whether the technique can be and has been
tested. We have seriously considered defendant’s argument that the testing of
fingerprint analysis that has been reported mostly falls short of the rigors
demanded by the ideals of science. On the other hand, the core proposition – that
reliable identifications may be made from comparison of latent prints with known
prints – is testable. And unquestionably the technique has been subject to testing,
albeit less rigorous than a scientific ideal, in the world of criminal investigation,
court proceedings, and other practical applications, such as identification of
victims of disasters.
Thus, while we must agree with defendant that this record does not show
that the technique has been subject to testing that would meet all of the standards
of science, it would be unrealistic in the extreme for us to ignore the
countervailing evidence. Fingerprint identification has been used extensively by
law enforcement agencies all over the world for almost a century. Fingerprint
analysts such as Mr. Fullerton, who have been certified by the FBI, have
25
undergone demanding training culminating in proficiency examinations, followed
by further proficiency examinations at regular intervals during their careers.
Although these proficiency examinations have been criticized on several grounds,
most notably that they do not accurately represent conditions encountered in the
field, we see no basis in this record for totally disregarding these proficiency
tests.
In conclusion, on this record we believe that the first Daubert factor weighs
somewhat in favor of admissibility, although not powerfully.
The second Daubert factor is whether the theory or process has been
subject to peer review and publication. We find little in the record to guide us in
consideration of this factor. Defendant argues persuasively that the verification
stage of the ACE-V process is not the independent peer review of true science.
Agent Meagher’s testimony included some references to professional
publications, but these were too vague and sketchy to enable us to assess the
nature of the professional dialogue offered. In short, the government did not
show in this case that this factor favors admissibility.
The third Daubert factor is the known or potential error rate of the
procedure. As recited supra, testing has been done in training programs and other
environments that are not shown to be accurate facsimiles of the tasks undertaken
by fingerprint analysts in actual cases. Nevertheless, the accumulated data is
impressive. Very few mistakes are reported in testing that trainees must complete
26
before progressing to actual casework. Mr. Fullerton, who made the actual
identification in this case, testified that he has always attained a perfect score in
his proficiency tests.
More significantly, Agent Meagher testified to an error rate of one per
every 11 million cases, and the defense did not – either in the evidentiary hearing
or in the briefs on appeal – challenge that testimony. There may have been
erroneous identifications that never came to light. Defense attorneys rarely have
the resources to hire independent experts for trial, and in the interests of finality
our system has created obstacles to post-conviction review. But even allowing
for the likelihood that the actual error rate for FBI examiners may be higher than
reflected in Mr. Meagher’s testimony, the known error rate remains impressively
low. We are not aware of any attempt to quantify the maximum error rate that
could meet Daubert standards, but surely a rate considerably higher than one per
11 million could still pass the test. We conclude that the evidence of the error
rate on this record strongly supported the judge’s decision to admit the expert
testimony.
The fourth Daubert factor is the existence and maintenance of standards
controlling the technique’s operation. On this point, we are persuaded by the
analysis of the Third Circuit in United States v. Mitchell, 365 F.3d at 241. The
ACE-V system is a procedural standard but not a substantive one. Critical steps
in the process depend on the subjective judgment of the analyst. We hasten to
27
add that subjectivity does not, in itself, preclude a finding of reliability. But in
searching this record for evidence of standards that guide and limit the analyst in
exercise of these subjective judgments, we find very little. Because in the end
determination of this factor is not critical to our decision, we will assume
arguendo that this factor does not support admissibility.
The fifth Daubert factor is whether the technique has attained general
acceptance in the relevant scientific or expert community. Conceding the general
acceptance of fingerprint analysis by law enforcement officials nationwide and
internationally, defendant contends that fingerprint analysis has not been accepted
in “any unbiased scientific or technical community” and cites to the Daubert
formulation of the standard, which was limited to the “relevant scientific
community.” 509 U.S. at 594. This distinction is significant in this case because
the field of fingerprint analysis is dominated by agents of law enforcement, with
apparently little presence of disinterested experts such as academics.
But in Kumho Tire, the Court – dealing with proffered expert testimony that
was characterized as technical rather than scientific – referred with apparent
approval to a lower court’s inquiry into general acceptance in the “relevant expert
community,” 526 U.S. at 156, and then the Court discussed its own search in the
record for evidence of acceptance of the controverted test by “other experts in the
industry.” Id. at 157. Consequently, while we acknowledge that acceptance by a
community of unbiased experts would carry greater weight, we believe that
28
acceptance by other experts in the field should also be considered. And when we
consider that factor with respect to fingerprint analysis, what we observe is
overwhelming acceptance.
Defendant argues that many of the post-Daubert cases holding fingerprint
analysis admissible placed so much weight on the general acceptance of the
practice that they in effect applied the outdated standard of Frye v. United States,
293 F. 1013 (D.C. Cir. 1923), the standard that Daubert held had been displaced
by Fed. R. Evid. 702. We need not either accept or reject this contention, as we
have examined this issue on the record in this case and have found guidance from
other courts primarily in their discussions of other factors. We have remained
mindful of Frye’s displacement, but also mindful that the Court specifically said
in Daubert that general acceptance “can yet have a bearing on the inquiry.” 509
U.S. at 594.
In reaching a conclusion after this process of focusing on each of the
Daubert factors in turn, we must return to two overriding principles. The first is
that our review here is deferential, limited to the question of whether the district
judge abused her considerable discretion. The second is that the Rule 702
analysis is a flexible one, as both Daubert and Kumho Tire teach. The Daubert
factors are “meant to be helpful, not definitive,” and not all of the factors will be
pertinent in every case. Kumho Tire, 526 U.S. at 150-51. On the whole, it seems
to us that the record supports the district judge’s finding that fingerprint analysis
29
is sufficiently reliable to be admissible. Thus, we find no abuse of discretion.
We apply the Third Circuit’s Mitchell standard: “[T]he usual precepts of abuse-
of-discretion review over the District Court’s decision to admit the government’s
expert testimony.” 365 F.3d at 234.
In closing, we echo the thoughts of Judge Pollak, who said regarding the
desirability of research to provide the scrutiny and independent verification of the
scientific method to aid in assessing the reliability of fingerprint evidence, that
such efforts would be “all to the good. But to postpone present in-court
utilization of this ‘bedrock forensic identifier’ pending such research would be to
make the best the enemy of the good.” United States v. Llera Plaza, 188
F.Supp.2d 549, 572 (E.D. Pa. 2002).
Conclusion
Having found no abuse of discretion in admission of the disputed evidence,
the only issue raised in this appeal, we AFFIRM the judgment of the district
court.
30