United States v. Mathis

Opinions of the United
2001 Decisions States Court of Appeals
for the Third Circuit

8-30-2001

United States v. Mathis
Precedential or Non-Precedential:

Docket 99-5940

Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2001

Recommended Citation
"United States v. Mathis" (2001). 2001 Decisions. Paper 200.
http://digitalcommons.law.villanova.edu/thirdcircuit_2001/200

This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova
University School of Law Digital Repository. It has been accepted for inclusion in 2001 Decisions by an authorized administrator of Villanova
University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu.
Filed August 30, 2001

UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT

No. 99-5940

UNITED STATES OF AMERICA
Appellee

v.

KEITH MATHIS
Appellant

On Appeal from the United States District Court
for the District of New Jersey
NO. 98-cr-656-2
District Judge: Honorable Alfred J. Lechner, Jr.

Argued June 1, 2000

Before: SCIRICA and NYGAARD, Circuit Judges and
POLLAK,* District Judge

(Filed: August 30, 2001)

_________________________________________________________________
* Honorable Louis H. Pollak, United States District Judge for the
Eastern District of Pennsylvania, sitting by designation.
 Robert J. Cleary, Esq.
George S. Leone, Esq.
Phillip H. Kwon, Esq. (argued)
United States Attorney
Assistant United States Attorney
970 Broad Street
Room 700
Newark, NJ 07102

Counsel for Appellee United States
of America

Mark A. Berman, Esq. (argued)
Peter J. Torcicollo, Esq.
Gibbons, Del Deo, Dolan,
Griffinger & Vecchione
One Riverfront Plaza
Newark, NJ 07102

Counsel for Appellant

OPINION OF THE COURT

POLLAK, District Judge:

This appeal concerns three challenges to Keith Mathis's
conviction for bank robbery and conspiracy to commit bank
robbery. First, Mr. Mathis claims that the District Court
wrongly admitted into evidence testimony describing his
involvement in previous, uncharged bank robberies.
Second, Mr. Mathis asserts that the District Court erred in
admitting evidence that his picture was selected from a
photographic array, because the array was
unconstitutionally suggestive. Third, Mr. Mathis argues
that the District Court wrongly excluded expert testimony
that called into question eyewitness testimony identifying
Mr. Mathis as he fled from the robbery. As described
herein, we disagree with Mr. Mathis's first and second
arguments on their merits; with respect to the third, we
hold that the District Court erred in part, but we find it
highly improbable that such error affected the jury's
decision. Thus, we affirm Mr. Mathis's conviction.

2
I. Background

On October 20, 1998, a grand jury indicted Mr. Mathis,
Steven Gantt, and Jeffrey Seaberry on one count of bank
robbery, in violation of 18 U.S.C. SS 2113(a) and 2, and one
count of conspiracy to commit bank robbery, in violation of
18 U.S.C. S 371. The indictment charged the three men
with conspiring, from October 11 to October 14, 1998, to
rob the Sun National Bank in Maple Shade, New Jersey,
and with successfully robbing that bank on October 14,
1998. Mr. Mathis pled not guilty to both counts, and his
trial commenced on January 19, 1999.

At Mr. Mathis's trial, two witnesses testified that they
saw, from an adjoining office building, three masked men
with guns run into, and then out of, Sun National Bank.
These witnesses testified that the three masked men drove
away in a dark-colored Jeep. Also, several Sun National
Bank employees testified that, when the three were inside
the bank, one stood near the door holding two guns, while
the other two jumped over the bank counter and took
money from tellers' drawers. Video footage taken by the
bank's security cameras confirmed the basic details of
these accounts.

The prosecution's primary witnesses were Sergeant Gary
Gubbei and one of Mr. Mathis's alleged co-conspirators, Mr.
Gantt. Sergeant Gubbei testified that, on the morning of
October 14, he responded to a radio dispatch describing the
robbery and the getaway vehicle. Soon thereafter, Sergeant
Gubbei spotted a black Jeep Cherokee with an African-
American driver in the opposing lane of traffic; Sergeant
Gubbei turned and gave chase. After a period of pursuit,
the Jeep left the highway and drove onto a grass median,
where the vehicle apparently stalled and coasted to a halt.
Three men exited the Jeep while it was still moving--the
driver first, then a forward passenger, then a rear
passenger--and escaped over a guardrail on the highway's
far side. The forward passenger, while stumbling over the
guardrail, dropped a black bag containing money. The rear
passenger carried a gun in his right hand as he exited the
Jeep and held the weapon near his head, pointing toward
the sky. The rear passenger momentarily looked back at
Sergeant Gubbei before running away, and Sergeant

3
Gubbei testified at trial that he was able, based on that
brief view, to identify the fleeing man as Mr. Mathis.
Sergeant Gubbei also testified that, at 1:00 p.m. on October
15, he selected Mr. Mathis's picture from an eight-picture
photographic array as depicting one of the Jeep's
occupants.

Mr. Gantt, who pled guilty in this case, testified that he
and Mr. Mathis jointly robbed a total of twelve banks,
including the Sun National Bank, and that seven of these
robberies, including that of the Sun National Bank, also
involved Mr. Seaberry. Mr. Gantt further testified that these
seven robberies shared other characteristics: The robbers
covered their faces with masks; the robbed banks were
located near Camden, New Jersey; and Mr. Mathis often
stood as an armed guard, while Mr. Gantt and Mr. Seaberry
vaulted the counters and stole cash from the drawers.

In describing the Sun National Bank robbery, Mr. Gantt
testified that Mr. Mathis, Mr. Seaberry, and he parked a
Jeep at the rear of the bank's parking lot, donned face
masks, ran alongside the bank, and entered through the
front of the building. Mr. Gantt stated that he and Mr.
Seaberry jumped over the tellers' counters and placed
money from the drawers in a bag they were carrying, while
Mr. Mathis stood guard with two pistols. The three then
returned to the Jeep. Mr. Seaberry drove, Mr. Gantt rode in
the forward passenger's seat, and Mr. Mathis rode in the
rear passenger's seat. Mr. Gantt testified that, as the Jeep
tried to evade a pursuing police cruiser (apparently driven
by Sergeant Gubbei), Mr. Seaberry tried to drive over the
highway's grass divider, causing the Jeep's engine to stall,
and the three men exited while the vehicle was still rolling
forward. Mr. Gantt stated that he and Mr. Mathis jumped
over the highway guardrail and, without Mr. Seaberry,
escaped after stealing a nearby delivery truck. At trial,
Detective Jeff Hoch testified that he arrested Mr. Seaberry
near an apartment complex beside the highway.

II. Grounds for Appeal

A. Evidence of Uncharged Robberies

At trial, the government moved in limine to admit into
evidence testimony from Mr. Gantt concerning eleven

4
robberies that he and Mr. Mathis had jointly undertaken
prior to the Sun National Bank robbery. The government
first addressed the testimony's admissibility under Federal
Rule of Evidence 404(b), which states:

Evidence of other crimes, wrongs, or acts is not
admissible to prove the character of a person in order
to show action in conformity therewith. It may,
however, be admissible for other purposes, such as
proof of motive, opportunity, intent, preparation, plan,
knowledge, identity, or absence of mistake or accident
. . . .

Fed. R. Evid. 404(b). According to the government, the
disputed evidence was admissible, inter alia, to show Mr.
Mathis's familiarity with Mr. Gantt and Mr. Seaberry prior
to the charged offense and to show the group's modus
operandi for robbing banks. In response, Mr. Mathis
conceded that Rule 404(b) was satisfied, but he claimed
that Mr. Gantt's testimony presented a risk of unfair
prejudice that substantially outweighed its probative value
and was therefore barred by Rule 403, which provides:

Although relevant, evidence may be excluded if its
probative value is substantially outweighed by the
danger of unfair prejudice, confusion of the issues, or
misleading the jury, or by considerations of undue
delay, waste of time, or needless presentation of
cumulative evidence.

Fed. R. Evid. 403.

The District Court admitted Mr. Gantt's testimony into
evidence, holding, as a preliminary matter, that Rule 404(b)
was satisfied because the testimony's description of
uncharged acts was linked to charged conduct by"[t]he
same general area, the same general time period, the intent
entering into the conspiracy, the background of this, the
familiarity [of] one with the other, the agreement among
everyone and the knowledge with regard to how to carry out
bank robberies." With respect to Rule 403, the District
Court acknowledged that there was "prejudice in this
evidence," but it also found that the evidence"possess[ed]
great probative value." Mr. Mathis's defense was one of
mistaken identity, and the District Court found that"[t]his

5
evidence from a co-defendant links him directly to the
crime, directly with the evidence of the police officer who
will testify. . . . I find it is not unfairly prejudicial looking at
what the Government must prove, what is available to the
Government and who is going to be offering the testimony."
When Mr. Gantt completed his testimony regarding the
uncharged robberies, the District Court read a jury
instruction (whose substance is not challenged) limiting use
of the testimony to showing "defendant's knowledge of the
crime of bank robbery and . . . his familiarity with the other
alleged co-conspirators" and to showing "that the robbery of
Sun National Bank was part of a scheme or plan in which
the defendant participated." The District Court stressed
that the jury could consider such evidence "only for [those]
limited purposes . . . [and] to use the evidence for any other
purpose would be improper and violative of your oath."

We review a district court's decision to admit or exclude
evidence for abuse of discretion, and such discretion is
construed especially broadly in the context of Rule 403.
E.g., Hurley v. Atl. Police Dep't, 174 F.3d 95, 110 (3d Cir.
1999). On appeal, Mr. Mathis claims that Rule 403 should
have barred the contested portion of Mr. Gantt's testimony
because "evidence relating to so many other uncharged
bank robberies" inevitably prejudiced the jury"by
demonstrating that Mathis had a propensity to rob banks."
Appellant's Br. at 28. He also claims that such testimony
was not significantly probative because the government
"could establish Mathis' identity, or his association to Gantt
. . . by Sergeant Gubbei's eyewitness identification. . . and
[by] Steven Gantt's testimony that he was acquainted with
Mathis and that he . . . robbed the Sun National Bank with
the defendant." Id. at 30.

With respect to the testimony's risk of prejudice, we agree
with the District Court that the disputed testimony's
description of eleven uncharged robberies, in a case seeking
conviction on only one, entailed some risk of unfair
prejudice. That is, the jury could well have interpreted Mr.
Gantt's testimony as proving that Mr. Mathis was a person
of bad character who deserved punishment independent of
the government's particular proof regarding the Sun
National Bank robbery. But in United States v. O'Leary, 739

6
F.2d 135 (3d Cir. 1984), this court held that the number of
uncharged acts alleged, even if disproportionate to the
number of charged acts, does not alone render admission of
such evidence an abuse of discretion. The defendant in
O'Leary was charged with one count of conspiracy to
distribute cocaine and five counts of possession with intent
to distribute cocaine. Mr. O'Leary claimed that the district
court erroneously admitted testimony stating that he had
sold drugs on "[a]bout 20" occasions during the course of
"[a]bout a year." Id. at 136. We affirmed the conviction,
finding that the district court had expressly analyzed the
defendant's Rule 403 challenge and had issued an
instruction "clearly explaining to the jury the narrow way in
which to use the evidence," thereby "lessen[ing] any
possibility of prejudice." Id. Similarly, in the case at bar, the
District Court explicitly balanced the testimony's probative
value against its danger of unfair prejudice, and it
instructed the jury on the permissible uses of Mr. Gantt's
testimony.1

The chief authority on which Mr. Mathis relies is United
States v. Hans, 738 F.2d 88 (3d Cir. 1984), where we
reversed a defendant's conviction for bank robbery and
assault during the commission of a bank robbery. The
testimony in Hans was from an FBI agent, and it described
the surprising speed with which the defendant became a
suspect in the case. Such testimony was deemed extremely
prejudicial because, in context, "the only reasonable
inference that a reasonable juror could draw from the
testimony was that Hans was well-known as a bank robber
_________________________________________________________________

1. Mr. Mathis seeks to distinguish O'Leary because the witness in that
case answered relatively few, relatively cursory questions about the
defendant's previous misconduct, while "the government's examination of
Gantt regarding the details of the eleven uncharged bank robberies went
on for 34 pages, and dominated Gantt's trial testimony." Reply Br. at 26.
We cannot agree that the government's efficiency, or lack thereof, in
conducting Mr. Gantt's examination significantly determines the
resultant testimony's prejudicial effect. Here, as in O'Leary, the
government's questions were sufficient in duration and substance to
raise the defendant's prior misconduct as an issue, and this fact
alone--not any claim of "undue delay, waste of time, or needless
presentation of cumulative evidence"--is the defendant's asserted basis
for prejudice.

7
to the Detroit F.B.I." Id. at 94; id. ("[I]t is difficult to imagine
testimony more prejudicial than [the] implication that Hans
was known to the Detroit police as a professional bank
robber . . . ."). Mr. Mathis has presented no argument that
Mr. Gantt's testimony implicated the unusual risk of
prejudice discussed in Hans, and we see no reason to
believe it did so. Whereas the testimony in Hans generated
untestable inferences about what was "well-known" by the
"Detroit F.B.I.," Mr. Gantt testified directly regarding his
own experience. Thus, Mr. Gantt's testimony was properly
challengeable, and was in fact challenged, by vigorous
cross-examination, which significantly mitigated its risk of
unfair prejudice. Cf. United States v. Gonzalez-Lira, 936
F.2d 184, 191 (5th Cir. 1991) (finding the prejudicial effect
caused by evidence of past acts mitigated by the
opportunity to cross-examine and by a proper limiting
instruction); United States v. Lynn, 856 F.2d 430, 437 (1st
Cir. 1988) (finding the risk of prejudice "especially
distressing" where the defendant "could not . . . cross-
examine the agent on the reliability of the information").

With respect to the other part of Rule 403's analysis, the
District Court found that Mr. Gantt's testimony possessed
"great probative value." On appeal, the government argues
that the disputed testimony "was admissible to show the
familiarity and interrelationship of Mathis, Gantt, and
Seaberry," thereby tending to prove Mr. Mathis's conspiracy
count. Appellee's Br. at 53-55. In evaluating this position,
we take guidance from this court's holding in United States
v. Scarfo, 850 F.2d 1015 (3d Cir. 1988). The defendant in
Scarfo was convicted of conspiring to extort, and of
extorting, money from a real estate developer. The district
court admitted testimony from two of the defendant's co-
conspirators stating that, prior to the charged extortion,
they had committed two murders pursuant to the
defendant's orders. These co-conspirators also testified
about numerous details of the defendant's "crime family,"
including uncharged murders, briberies, and extortions. On
appeal, we affirmed the defendant's conviction, holding that
"[b]ecause the government's case against the defendants
rested on the testimony of unindicted co-conspirators,
evidence describing their relationship to the defendants and
providing background information to illustrate the

8
witnesses' roles in the scheme was properly admitted." Id.
at 1020. Also, in order for the jury "to realize that [the co-
conspirators] had been granted immunity for the very
murders that they asserted Scarfo had ordered" and
thereby to assess the witnesses' motive for testifying, we
concluded that such evidence was "essential in the
government's effort to establish the credibility of its
disreputable, yet indispensable witness." Id.

The charged conduct here, of course, lacks much of the
gravity and extensiveness associated with the activities
described in Scarfo, but the applicable legal principles are
not so dissimilar. Mr. Mathis was indicted for a conspiracy,
whose details the government was required to prove, and
Mr. Gantt was the government's only witness who could
describe the conspiracy directly. Mr. Gantt's testimony
regarding the eleven uncharged robberies, if believed by the
jury, explained both Mr. Gantt's role in the conspiracy and
his motive for testifying, i.e., to fulfill his plea agreement
and reduce his sentence. This testimony also tended to
show that, in addition to their previous associations, Mr.
Gantt and Mr. Mathis conspired in robbing Sun National
Bank, on the theory that co-conspirators more often trust
those with whom they are already intimate. Contrary to Mr.
Mathis's argument that the disputed testimony was
unnecessary given other evidence in the case, Mr. Gantt's
testimony was the only direct evidence offered for the
purposes listed above. Accordingly, for reasons parallel to
those in Scarfo, we deem Mr. Gantt's testimony significantly
probative of Mr. Mathis's involvement in the charged
conspiracy.2
_________________________________________________________________

2. The government also argues that Mr. Gantt's testimony demonstrates
a common modus operandi. The government's position begins with the
solid premise that "[a] jury can rationally infer from evidence that the
defendant committed a prior crime in an unusual and distinctive manner
and evidence that a second similar crime was committed in the same
unusual and distinctive manner that the defendant committed the
second crime." Gov't of Virgin Islands v. Pinney, 967 F.2d 912, 916 (3d
Cir. 1992). In order to distinguish this type of inference from
impermissible conclusions based on propensity or bad character,
however, the admissibility of such evidence critically depends on the
degree to which the "manner" employed is "unusual and distinctive." 1

9
After reviewing the District Court's explanation for its
ruling, we find no abuse of discretion in the District Court's
conclusion that the disputed testimony's danger of unfair
prejudice did not substantially outweigh its probative value.
Mr. Gantt's testimony about the uncharged robberies
provided significant background information bolstering the
charged conspiracy's plausibility, and the risk of prejudice
attending such evidence was mitigated by the opportunity
for cross-examination and by the District Court's limiting
instruction. Thus, we hold that the District Court did not
abuse its discretion by declining to exclude Mr. Gantt's
testimony.

B. Evidence of Photographic Identification

Before trial, the District Court conducted an evidentiary
hearing to determine the admissibility of testimony from
Sergeant Gubbei stating that he saw Mr. Mathis exit the
Jeep on October 14, 1998, and that he selected Mr.
Mathis's picture from a photographic array on October 15,
1998. At that hearing, Sergeant Gubbei testified that he
chose Mr. Mathis's photograph from an array of eight
photographs at the Maple Shade police department, and
Mr. Mathis moved to exclude such testimony because
Sergeant Gubbei had previously seen an identical
_________________________________________________________________

McCormick on Evidence 662-63 (John W. Strong ed., 5th ed. 1999)
("Much more is demanded than the mere repeated commission of crimes
of the same class, such as repeated murders, robberies, or rapes. The
pattern and characteristics must be so unusual and distinctive as to be
like a signature.") (footnotes omitted). On this point, Mr. Gantt's
testimony indicated only that all of the uncharged robberies at issue
occurred near Camden, New Jersey, within approximately four months of
one another, and that the robbers wore gloves and masks. Mr. Gantt
further testified that sometimes his girlfriend waited outside the target
bank in a car, other times she did not; sometimes one robber acted as
an armed guard near the bank's door, sometimes not; sometimes the
robbers would use a stolen getaway car, other times not; sometimes they
used a .45 caliber handgun, other times a .22 caliber handgun, and
sometimes they used toy guns or screwdrivers taped to look like guns.
None of the authorities cited by the government has approved modus
operandi admissibility based on such inconsistent and generally non-
distinctive characteristics, and we find the evidence in this case
insufficient to support such a theory.

10
photograph of Mr. Mathis. Specifically, a Maple Shade
detective had, roughly one month before the robbery, given
Sergeant Gubbei materials describing various bank robbery
suspects. According to Sergeant Gubbei, these materials
included between thirteen and fifteen pages with
approximately seven photographs of bank robbery suspects,
and one of these photographs--virtually identical to that
later placed in the array--depicted Mr. Mathis. 3

Mr. Mathis argued that admitting this testimony from
Sergeant Gubbei would violate the Due Process Clause's
protection against certain suggestive identification
procedures. In particular, Neil v. Biggers, 409 U.S. 188
(1972), held that, for out-of-court identifications,
"convictions based on eye-witness identification . . . will be
set aside only if the photographic identification was so
impermissibly suggestive as to give rise to a very
substantial likelihood of . . . misidentification." Id. at 196-
97 (internal quotation omitted).4 Biggers further declared
that likelihood of misidentification should be measured by
a totality of circumstances including: the witness's initial
opportunity to view the suspect at the crime scene and
degree of attention at that time, the witness's level of
certainty in the disputed identification, the length of time
between initial viewing and disputed identification, and the
accuracy of any intervening description of the suspect
occurring between those two events. Id. at 199-200.

In the present case, Mr. Mathis contended that "the very
fact of [a] previous viewing" was itself sufficient to establish
the identification procedure's suggestiveness. He further
argued that a substantial likelihood of misidentification
arose both from Sergeant Gubbei's limited opportunity to
view Mr. Mathis as he fled and from Sergeant Gubbei's
_________________________________________________________________

3. We base much of our analysis on Sergeant Gubbei's testimony
because the materials themselves, aside from the page containing Mr.
Mathis's picture, apparently were not introduced into evidence.

4. See also United States v. Emanuele, 51 F.3d 1123, 1128 (3d Cir. 1995)
(interpreting later cases as requiring only a "substantial risk of
misidentification").

11
degree of attention, which was arguably impaired by other
people and events he witnessed at that time.5

The government replied that Sergeant Gubbei's testimony
showed that he had only reviewed the materials containing
Mr. Mathis's photograph once or twice, immediately after
receiving them approximately a month before the Sun
National Bank robbery. Furthermore, Sergeant Gubbei
stated he had attended most closely to the suspects' modi
operandorum, not to their photographs. The government
also noted that Sergeant Gubbei described his view of Mr.
Mathis's face on October 14 as unobstructed and clear, and
that he exhibited both a high degree of attention in
watching the Jeep's occupants disembark and a high level
of certainty in making his subsequent photographic
identification.

In denying Mr. Mathis's motion to exclude the testimony,
the District Court followed this court's two-step process for
applying Biggers:

The first question is whether the initial identification
procedure was `unnecessarily' . . . suggestive. This
inquiry . . . contains two component parts: that
concerning the suggestiveness of the identification, and
that concerning whether there was some good reason
for the failure to resort to less suggestive procedures. If
a procedure is found . . . unnecessarily suggestive, the
next question is whether the procedure . . . gave rise to
such a `substantial likelihood of . . . misidentification'
that admitting the identification would be a denial of
due process.

United States v. Stevens, 935 F.2d 1380, 1389 (3d Cir.
1991) (internal quotation omitted). The District Court ruled
that Mr. Mathis did not "initially demonstrate[that] the
identification was impermissibly suggestive"; the District
Court then proceeded to find that Sergeant Gubbei's
identification was reliably based in fact and most likely did
_________________________________________________________________

5. Mr. Mathis conceded that Biggers's"length of time" factor weighed in
favor of admitting the evidence, since the contested identification
occurred just one day after Sergeant Gubbei viewed the Jeep's
passenger.

12
not result from impermissible suggestion. The District
Court also found that the "background of the photo of the
defendant is no darker or lighter than the other photos,"
that the cropped pictures did not reveal any difference in
the subjects' build, and that the eight men "have the same
skin tone, have the same facial features, eyes, nose, lip
shape, [and] facial hair."

We review the District Court's decision for abuse of
discretion, applying clear error review to its underlying
factual findings and plenary review to its conclusions
drawn from such facts. United States v. Emanuele , 51 F.3d
1123, 1127 (3d Cir. 1995); see also Sumner v. Mata, 455
U.S. 591, 597 (1982) (discussing the relationship between
fact and law in the context of pretrial identifications). On
appeal, both parties have primarily disputed the factual
reliability of Sergeant Gubbei's identification and whether
there existed a substantial likelihood of misidentification.
We affirm, however, based solely on the first part of the
Stevens test, i.e., the District Court's determination that
Mr. Mathis did not show the photographic identification to
be unnecessarily suggestive. See United States v. Hill, 967
F.2d 226, 230 (6th Cir. 1992) (noting the defendant's
burden of proof on this issue).

Mr. Mathis contends, as he did before the District Court,
that "a photographic array created for purposes of a bank
robbery investigation--which consists of seven photos the
witness has never seen before, and one photo the witness
has seen before in the very context of a bank robbery
investigation--is by its very nature unduly suggestive."
Appellant's Br. at 14. We disagree, finding that, without
more, this allegation speaks only to the weight of the
evidence and, thus, may properly be argued to the jury--as
Mr. Mathis did in this case. Cf. Manson v. Braithwaite, 432
U.S. 98, 116 (1977) ("[E]vidence with some element of
untrustworthiness is customary grist for the jury mill.
Juries are not so susceptible that they cannot measure
intelligently the weight of identification testimony that has
some questionable feature."). The unadorned fact that
Sergeant Gubbei briefly viewed Mr. Mathis's picture, as part
of a collection of materials distributed one month before the
robbery, does not demonstrate sufficient suggestiveness to

13
require constitutional inquiry into whether a
misidentification might have occurred in this case.

Since we have not discovered cases squarely addressing
facts like these, we are primarily guided by principles
articulated in Stevens, 935 F.2d at 1380, where the
defendant's picture was selected from a photographic array
by two witnesses who had previously identified the
defendant from a police station's wanted board. That board,
to which the witnesses were directed by a police officer as
they filed their complaint, contained eight posters."[F]ive
were composite sketches, one contained eight small
photographs, and the other two were both photographs of
[the defendant]. . . . [O]ne of his photographs was in color,
whereas the other nine photographs were in black and
white." Id. at 1388. In evaluating the suggestiveness of the
board's collection of photographs, we acknowledged that
the wanted board had "several suggestive attributes":
"Whereas the posters of Stevens each contained
photographs, most of the other posters had only composite
sketches. Stevens's picture, moreover, was the only one
that appeared twice . . . . Even more bothersome . .. are
that Stevens's photographs were significantly larger than
the others, and that Stevens was the only suspect
portrayed in color." Id. at 1390. Each of these features
"quite possibly could have drawn the victims' attention to
Stevens" independent of the witnesses' recollection. Id. But
we also noted certain considerations that counseled against
deeming such suggestiveness "unnecessary," including the
fact that the wanted board had not been "arranged with
this particular crime in mind. It was, rather, a collection of
random sketches and photographs that had been
assembled in order to facilitate chance identifications . . . ."
Id. Measured against this general law enforcement purpose,
we also noted that wanted boards often cannot be made
with photographs of uniform size. Thus, although aspects
of the particular wanted board in Stevens were
unnecessarily suggestive, we made significant efforts to
stress that such displays generally are not so. Compare id.
with id. at 1390 n.12.

Here, the materials that Sergeant Gubbei received were
apparently assembled for a law enforcement purpose whose

14
validity is not challenged. Such materials obviously were
not prepared or circulated with the Sun National Bank
robbery (which occurred thereafter) in mind, nor, it seems,
with awareness of influencing any photographic
identification procedures. Cf. Emanuele, 51 F.3d at 1129
("[W]e expressly do not require defendant to establish the
government's state of mind. On the other hand, evidence
that the government intended and arranged such [a
potentially suggestive] encounter would be a substantial
factor in the court's analysis."). We hold that, with respect
to the present facts, due process does not protect
defendants like Mr. Mathis from photographic materials
briefly viewed one month before the disputed identification,
when such materials are distributed only for purposes of
general law enforcement. The brief time that elapsed
between Sergeant Gubbei's receiving the materials and his
selecting Mr. Mathis's picture from the photographic array
significantly reduces the level of cognizable "suggestiveness"
at issue. And though it might seem preferable, or even more
effective, for police departments to use photographs in
identification arrays that are different from those used in
previously distributed suspect data, we do not find an
identification "unnecessarily suggestive" simply because the
same photograph, or a very similar photograph, is used in
both. Law enforcement agencies commonly lack an
abundance of particular suspects' photographs for use in
assembling photographic arrays. And even where various
pictures are on hand, lists are not often kept recording
which pictures have been circulated to officers who later
happen to be eyewitnesses. Given such operational
circumstances--similar to those concerning police station
wanted boards--we cannot agree that Sergeant Gubbei's
prior exposure to Mr. Mathis's photograph, in and of itself,
renders his subsequent identification unnecessarily
suggestive.

Mr. Mathis further claims that certain "additional
circumstances" aggravated the suggestiveness of Sergeant
Gubbei's photographic identification. First, Mr. Mathis
notes that Sergeant Gubbei chose Mr. Mathis's photograph
only after Mr. Gantt had already identified one of his co-
conspirators by name as "Mr. Mathis." We agree with the
government, however, that there is no evidence to connect

15
Mr. Gantt's identification of Mr. Mathis by name with
Sergeant Gubbei's identification of Mr. Mathis by image.
Since the photographic array did not list the names of its
subjects, Mr. Mathis has failed to show how Mr. Gantt's
prior identification by name increased the photographic
array's suggestiveness. Second, Mr. Mathis claims that, of
the eight photographs in the array, Mr. Mathis's picture
"quite obviously" was the only one of a "slender and slight"
build. We review for clear error the District Court's factual
finding to the contrary, and we find none. The array at
issue portrays only the subjects' heads and necks, making
their general physiques very difficult to estimate. Even from
this limited view, however, it seems that at least two other
photographs depict men of a build similar to Mr. Mathis's.
Third, Mr. Mathis claims that the background of Mr.
Mathis's photograph "was noticeably darker than the
remaining seven." Upon reviewing the array, we agree and
find the District Court's finding to the contrary clearly
erroneous; thus, we must assess the impact of this
circumstance on the alleged "unnecessary suggestiveness"
of Sergeant Gubbei's identification. Although Mr. Mathis's
picture is darker than the others in the array, there are at
least three other pictures whose backgrounds are
somewhat dark. Given the current technological state of
photographic reproduction, and the variety of backgrounds
evident in the eight pictures at issue, we hold that the
slightly darker background of Mr. Mathis's picture did not
significantly contribute to the array's unnecessary
suggestiveness. See United States v. Burdeau, 168 F.3d
352, 357 (9th Cir. 1999).

In sum, we find that Sergeant Gubbei's photographic
identification was not "unnecessarily suggestive" to the
degree required by Biggers. Due process does not mandate
the removal of testimony like Sergeant Gubbei's from the
jury's consideration, even if such evidence presents (as
much admissible evidence does) a risk of misidentification.
On the contrary, such modest levels of "suggestiveness,"
which are not deemed constitutionally "unnecessary,"
speak only to the evidence's ultimate reliability and should
therefore be argued for authoritative appraisal to the
factfinder. On the current record, we find no abuse of

16
discretion in the District Court's decision to admit Sergeant
Gubbei's testimony into evidence.

C. Expert Testimony Regarding Eyewitnesses

After the government had presented its witnesses, Mr.
Mathis sought to present testimony from Dr. Geoffrey
Loftus regarding the reliability of eyewitness identifications
such as Sergeant Gubbei's. The District Court conducted
an evidentiary hearing to determine whether such
testimony was admissible. During that hearing, Dr. Loftus
testified that he had received a doctoral degree from
Samford University in clinical psychology and had served as
a professor at the University of Washington for over twenty-
five years. His doctoral dissertation concerned humans'
ability to recall and recognize pictures from memory, and
his subsequent studies in the field of human perception
and scientific methodology were published as books and
professional journal articles. Based on these and other
academic credentials, the District Court qualified Dr. Loftus
as an expert in the field of human perception and memory.

Dr. Loftus then proffered testimony on four topics
concerning the operation of human memory: (i) the
preconditions for forming accurate memories, (ii) the
confounding impact of "double identification" or "post-event
information," (iii) the relationship between individuals'
confidence in describing memories and the accuracy of
such memories, and (iv) potential disruptions caused by
"weapons focus." With respect to memory formation, Dr.
Loftus observed that, among other circumstances, the
brevity of one's visual exposure to an object tends to cause
an incomplete memory, which might later be especially
susceptible to biasing influences. Dr. Loftus described
"double identification" as a problem in determining whether
one's memory derives from one of two or more possible
visual exposures to an object. According to Dr. Loftus, a
witness's recognizing someone, under circumstances
similar to those described by Sergeant Gubbei, could derive
either from the witness's actually having seen the
recognized person or from a previous exposure to that
person's photograph. Dr. Loftus indicated, citing scientific
studies, that when one encounters a remembered image
along with "post-event information" suggesting a particular

17
context from which the image might be remembered (such
as the Sun National Bank robbery's photographic array),
such post-event information can become incorporated with
the original memory, creating an inaccurately
"remembered" association between the image and its
source. Regarding witnesses' confidence in their
identifications, Dr. Loftus stated that "the correspondence
between confidence and accuracy is, at best, about 25
percent." He further explained that when conditions
attending the recalled memories are poor, "when, for
example, there is only a brief duration . . . [or] incorrect
post-event information introduced to the [observer], it's
under those circumstances that the relation between
confidence and [accurate] memory is zero." And on the topic
of "weapons focus," Dr. Loftus discussed studies finding
that the presence of a dangerous weapon can weaken one's
ability to recall other aspects of a remembered scene,
including individuals present therein.

In applying these principles to hypothetical questions
intended to simulate the facts of this case, Dr. Loftus stated
that "it's two to three times as likely that the identification
in the photo montage was made based on seeing the
photograph four weeks earlier than it was based on seeing
the individual" who fled on October 14, 1998. Dr. Loftus
also concluded that, if the photo array did not actually
depict the person who exited the Jeep, that array itself
would constitute incorrect post-event information. And
such misinformation might have caused Sergeant Gubbei's
identification of Mr. Mathis to reflect the man seen in the
previously viewed photograph, who was observed under
conditions favorable to memory formation, rather than the
man seen running away, who was observed under
conditions unfavorable to memory formation. Dr. Loftus
further opined that the officer's confidence in this
identification and the factual accuracy thereof would have
a probabilistic relationship of "essentially zero" under such
circumstances.

Mr. Mathis moved to admit the foregoing testimony into
evidence pursuant to Federal Rule of Evidence 702, which
provides:

18
 If scientific, technical, or other specialized knowledge
will assist the trier of fact to understand the evidence
or to determine a fact in issue, a witness qualified as
an expert . . . may testify thereto in the form of an
opinion or otherwise, if (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. In particular, Mr. Mathis argued that Dr.
Loftus's testimony was based on generally accepted
scientific methods and data, that there was a direct relation
between the proffered testimony and the factual issues in
dispute, and that the proffered testimony would"refute
common assumptions about the reliability of eye witness
identification testimony that will assist the jury to weigh
the credibility of Sergeant Gubbei's eye witness
identification."

The District Court excluded Dr. Loftus's testimony, but it
ratified Dr. Loftus's qualifications as an expert and
dismissed all methodological challenges to the studies he
cited. The following constitutes the District Court's
explanation of its ruling:

I'm particularly concerned with [Rule] 403 and I'm
inclined not to allow this testimony. I believe this
testimony has the probability of confusing and
misleading the jury. I find no problem with his
qualifications. Obviously he is qualified in this area.
The debate over the differences in the studies is an
interesting academic debate. As [defense counsel]
points out, however, there is only one [w]ay to run the
studies. You can't run them in a live situation, they
have to be in a controlled situation. That does not give
me pause.

What does give me pause is the fact that I believe that
in point of fact, there is a probability that there will be
unfair prejudice here. The aura of reliability that's
attached to an expert witness, I believe, is significant.
Listening to this expert, it seems to me, that the
testimony itself has the potential, if not controlling

19
 probability of confusing the jury. There is no issue in
this case about cross-racial identification. There is no
issue about stress. They were the two more compelling
issues about which I was concerned.

The issues concerning the focus at the time of the
chase and when the defendant exited the vehicle are
subject to the jury. The defense in this case had an
untrammeled opportunity to cross and did, in fact,
cross the police officer vigorously on that area. The
testimony is, I think clear. Whether the jury accepts it
or not, how the jury accepts it is clearly within their
province. I do not see how this evidence will do
anything, other than to confuse and mislead the jury.
I decline the invitation to admit.

1. Admissibility of Dr. Loftuss Testimony

We review the District Court's decision to exclude Dr.
Loftus's testimony for abuse of discretion. General Electric
Co. v. Joiner, 522 U.S. 136, 143 (1997); Oddi v. Ford Motor
Corp., 234 F.3d 136, 146 (3d Cir. 2000). Although the
District Court explicitly rested its decision on Rule 403's
standards for probative value and prejudice, the
government on appeal invokes only Rule 702's prescriptions
regarding expert testimony. We will address both rules,
beginning with Rule 702.

This court has construed Rule 702 as embodying "three
distinct substantive restrictions on the admission of expert
testimony: qualifications, reliability, and fit." Elcock v.
Kmart Corp., 233 F.3d 734, 741 (3d Cir. 2000). Only "fit" is
contested here; the government acknowledges that Dr.
Loftus is a properly qualified expert and that his methods,
principles, and data are of a sufficiently reliable scientific
character. See generally Daubert v. Merrell Dow Pharm.,
509 U.S. 579, 592-94 (1993) (discussing factors that may
be considered in assessing scientific reliability); In re Paoli
R.R. Yard PCB Litig., 35 F.3d 717, 742 n.8 (3d Cir. 1994)
(same).

Rule 702's fit requirement derives from the textual
provision that "scientific, technical, or other specialized
knowledge will assist the trier of fact to understand the
evidence or to determine a fact in issue." Fed. R. Evid. 702.

20
The requirement is "not intended to be a high one,"
however, Oddi, 234 F.3d at 145, and its principle is not
dissimilar to the Federal Rules' general provision that,
unless otherwise specified, "[a]ll relevant evidence is
admissible" and "[e]vidence which is not relevant is not
admissible." Fed. R. Evid. 402; Kumho Tire Co. v.
Carmichael, 526 U.S. 137 (1999) (interpreting Rule 702 as
requiring all expert testimony to be relevant and reliable).
Compare United States v. Downing, 753 F.2d 1224, 1230
(3d Cir. 1985) (noting "the liberal standard of admissibility
mandated by Rule 702"), with Daubert, 509 U.S. at 587
(noting that Rule 402's "basic standard of relevance is a
liberal one"). See generally Daubert, 509 U.S. at 592
(describing inquiry under Rule 702 as a determination "of
whether the reasoning or methodology underlying the
testimony is scientifically valid and of whether that
reasoning or methodology properly can be applied to the
facts in issue").

The clearest guide for interpreting Rule 702 in the
context of this case is United States v. Downing , 753 F.2d
at 1226, wherein this court, in 1985, first considered the
admissibility of expert testimony concerning human
memory and eyewitness testimony. In Downing, we
discussed certain of the rationales advanced by other
courts of appeals in prior years for excluding such
testimony, including notions that relevant issues could
adequately be raised through cross-examination and
common sense, that such testimony usurps the jury's
function, and that such evidence would lead to an unduly
confusing "battle" of experts. Id. at 1229-30 & n.4. In
explaining why we found those rationales unpersuasive, we
disavowed skepticism of such testimony as a matter of
principle, and we remanded for the district court to apply
Rule 702's "helpfulness test." We described this test, in
pertinent part, as follows: "[A]dmission depends upon the
`fit,' i.e., upon a specific proffer showing that scientific
research has established that particular features of the
eyewitness identifications involved may have impaired the
accuracy of those identifications." Id. at 1226.

In the case at bar, the government challenges the fit of
Dr. Loftus's testimony in three respects. First, on the

21
subject of "double identification" and "post-event
information," the government claims that Dr. Loftus's
testimony did not pertain to the present facts because,
according to Dr. Loftus's own theory, Sergeant Gubbei
viewed Mr. Mathis's face only once, as part of the suspect
materials, before selecting from the photographic array.
Thus, the government claims, this case did not involve any
"double" identification. Appellee Br. at 36. The government
also argues that there was no post-event information
"between the visual identification of Mathis on the day of
the robbery and the photo array identification." Id. at 38.
Each of these arguments misconstrues the substance of Dr.
Loftus's testimony. As explained supra, double
identification theory indicates that a person confronting an
image may have a firm memory of having previously viewed
the image, but may not have a similarly firm memory of
when that viewing occurred. Given the suboptimal
conditions for memory formation during the Jeep incident,
and given the relative calm under which Sergeant Gubbei
had previously seen a photograph of Mr. Mathis, Dr. Loftus
testified that Sergeant Gubbei might have recalled Mr.
Mathis's face from that previously viewed photograph
but--because he was searching the photographic array in
order to identify the Jeep's passenger--he might have
wrongly associated the recalled photographic image with
the fleeing robbery suspect. Contrary to the government's
misunderstanding, the two pertinent "identifications" were
the view of the suspect materials and the view of the man
who fled.

Furthermore, according to Dr. Loftus, the photographic
array constituted potentially inaccurate post-event
information, which might have influenced Sergeant Gubbei
wrongly to associate Mr. Mathis's image with the Sun
National Robbery.6 Dr. Loftus also was of the view that
_________________________________________________________________

6. The government argues that the photographic array was not distortive
post-event information if, as we have now held, the array was not
unconstitutionally suggestive. As our earlier discussion makes clear,
however, the standards of "unnecessary suggestiveness" for admissibility
under Biggers simply have no bearing on Dr. Loftus's testimony. A
photographic identification may be highly suggestive as a psychological
matter and still be deemed legally admissible as a constitutional one. A
determination that such an identification is admissible poses no greater
barrier to attack by expert testimony than it does to cross-examination.

22
Sergeant Gubbei's act of choosing Mr. Mathis's photograph
from the several photographs constituting the photo array
was itself post-event information that could have influenced
Sergeant Gubbei's face-to-face identification of Mr. Mathis
at trial. Dr. Loftus argued that these pieces of post-event
information may have distorted each of Sergeant Gubbei's
purported identifications of Mr. Mathis. Given the close
correspondence between these aspects of Dr. Loftus's
testimony and the factual issues disputed at trial, we
conclude that the proffered evidence did not lack
evidentiary fit for purposes of Rule 702.

Second, the government contends that Dr. Loftus's
observations with respect to the weak relationship between
an eyewitness's testimonial confidence and the accuracy of
such testimony did not fit because, in the government's
view, Dr. Loftus presented "nothing more than a general
thesis on the correlation between confidence and accuracy."
Id. at 44. As an initial matter, this characterization of the
evidence appears mistaken. Dr. Loftus testified about the
relationship between confidence and accuracy under
specific circumstances similar to those in this case: an
identification based on a brief view, in the presence of a
deadly weapon, and several potentially distracting events
on-going. Also, Sergeant Gubbei's testimony on direct
examination indirectly implicated his confidence as an
element supporting the credibility of his identification. In
describing how he selected Mr. Mathis's photograph,
Sergeant Gubbei was asked, "Q: Did you have any
hesitation in determining those individuals [Mr. Mathis and
Mr. Gantt]? A: No, I didn't." Later, after already having
described his choice of Mr. Mathis's photograph from the
array, Sergeant Gubbei was again asked, "Q: Sitting here
today Sergeant, was your identification of Keith Mathis
from the photo array based upon your viewing him the day
before, following the bank robbery or on the photograph
that you had been shown in the packet? A: I identified him
from the incident that took place from the Sun National on
the day before." The level of confidence underlying Sergeant
Gubbei's identification came into more explicit focus during
cross-examination:

Q: [A]s we sit here today, is it even possible that the
identification you made of Mr. Mathis in that photo

23
 array was based not on seeing him exiting that Jeep
but on the previous opportunity to observe that
photograph.

A: No, I'm positive by him getting out of the vehicle
. . . .

Q: Your answer is it's not even a possibility?

A: I guess there is a remote possibility, but I'm positive
of the identification when he exited the vehicle.

Where Sergeant Gubbei testified so directly regarding his
high confidence in the disputed identification, it seems
clear that Dr. Loftus's testimony did address, for purposes
of Rule 702, pertinent factual disputes in the case. See
Stevens, 935 F.2d at 1400 (reaching an identical conclusion
where the defendant sought "[t]o rebut the natural
assumption that [the witnesses'] strong expression of
confidence indicates an unusually reliable identification").

Third, the government argues that Dr. Loftus's opinions
about "weapons focus" were not relevant to the present
case because Sergeant Gubbei was a police officer who was
quite accustomed to firearms, and because the relevant gun
was pointed into the air, not at Sergeant Gubbei. Such
arguments may be useful topics for cross-examination, but
we do not find them persuasive grounds for exclusion
under Rule 702. Dr. Loftus expressed the view that the
phenomenon of weapons focus derives both from the
infrequency with which most individuals view deadly
weapons and from a survival instinct that draws one's
attention to potentially threatening objects. He pointed out
that various studies, performed by him and by other
scientists, showed that a weapon's presence in a scene
substantially affects subjects' ability to recall individuals in
that scene. He further pointed out that close physical
proximity of a weapon to another object (e.g. , the face of the
Jeep's passenger) would not appreciably reduce the
weapon's distortive influence on a witness's memory.

With respect to Sergeant Gubbei's being a police officer,
Dr. Loftus acknowledged that one's degree of familiarity or
unfamiliarity with the weapon at issue generally would
affect the practical impact of weapons focus. But Dr. Loftus

24
also noted that the primary study on which he relied
involved a knife, which, in his view, was an adequately
familiar object for most people. Moreover, Dr. Loftus
specifically rejected any notion that police officers would be
immune to the psychological principles he described. With
respect to the particular location of the gun, Dr. Loftus
indicated that psychological studies inevitably take place in
secure environments; thus, all of the cited studies assumed
observers who were actually safe from the weapon at issue.
He observed that the "survival" impetus for weapons focus
would, in real life situations, likely be even higher than the
cited studies indicated. Thus, we cannot agree with the
government's suggestion that Dr. Loftus's expert opinions
regarding weapons focus were not relevant for purposes of
Rule 702.

Having concluded that the government's arguments for
exclusion under Rule 702 cannot succeed, we turn now to
Rule 403. As an initial matter, "[w]e are mindful that a trial
court is in a far better position than an appellate court to
strike the sensitive balance dictated by Rule 403. When a
trial court engages in such a balancing process and
articulates on the record the rationale for its conclusion, its
conclusion should rarely be disturbed." Pinney, 967 F.2d at
918; see also Downing, 753 F.2d at 1242-43 (noting Rule
403's important role in assessing whether expert
testimony's probative value is "substantially outweighed by
other dangers, e.g., confusion of the issue or waste of
time"). But when a district court does not explain its ruling,
and no adequate explanation can be gleaned from the
record, appellate courts often have little choice but to
undertake such evaluation themselves. See e.g. , United
States v. Murray, 103 F.3d 310, 319 (3d Cir. 1997) (noting
that when district courts fail to explain their rulings under
Rule 403, "we are able to perform this balancing here,
[though] other cases may require remand . . . or even . . .
a new trial"); United States v. Himelwright , 42 F.3d 777,
785 (3d Cir. 1994).

We find it difficult to accord the customary degree of
deference to the District Court's discretion in this case
because the District Court explained its ruling with little
more than a series of conclusions:

25
 [T]his testimony has the probability of confusing and
misleading the jury. . . . [T]here is a probability that
there will be unfair prejudice here. The aura of
reliability that's attached to an expert witness, I
believe, is significant. Listening to this expert, it seems
to me, that the testimony itself has the potential, if not
controlling probability of confusing the jury. . . .
[Sergeant Gubbei's] testimony is, I think, clear.
Whether the jury accepts it or not, how the jury
accepts it is clearly within their province. I do not see
how this evidence will do anything, other than to
confuse and mislead the jury.

The District Court clearly expressed concern with"[t]he
aura of reliability that's attached to an expert
witness"--which is one reason for district courts'
"gatekeeping function" in assessing expert testimony under
Rule 702. See Daubert, 509 U.S. at 592, 594. There is no
suggestion, however, that such an aura of reliability was
unwarranted in this case or, to be more precise, that it was
unfairly prejudicial. From the record, it seems that Dr.
Loftus was an extremely qualified, experienced academic
presenting opinions on topics near the heart of his
expertise. Moreover, Dr. Loftus's conclusions seem closely
tied to empirical studies whose reliability is not impeached,
and he explained the bases for these studies at apparently
appropriate length. In short, we see no reason to believe
that Dr. Loftus's aura of reliability reflected anything other
than his actual reliability as an expert witness. With
respect to the District Court's concern with "confusing and
misleading the jury" and "unfair prejudice," we are unable
to discern from these references, any more than from our
own review of the record, how such problems might arise.

The government's oral argument before the District Court
also fails to reveal any valid basis for exclusion under Rule
403. The government focused on two contentions. First, it
claimed that the studies cited by Dr. Loftus did not
sufficiently match the circumstances of this case. The
District Court rejected this position, holding that such
disputes "did not give [him] pause," and the government
has conceded this issue on appeal.

Second, the government stated that:

26
 [T]he problem that bothers me most, your Honor, is the
danger of overwhelming the jury. When we gave this
witness a hypothetical that was slightly--you know,
everyone in this courtroom knows the facts better than
I do, but [it] sounded to me slightly tinged toward the
defense version of the offense, the witness stated as a
fact that the sergeant would not remember accurately.
When I gave him the version that was a little bit closer
to what I believe the facts to be and the Government
version, he stated as a fact that his conclusion would
be very different under those facts. So what he's doing
is telling the jury the answer instead of allowing the
jury to determine the facts for themselves. Any
hypothetical that sets forth those facts is likely to have
a serious effect on the jury's memory of the testimony
rather than allowing the jury to decide what the jury's
going to decide.

The full significance of this language is not clear. In part,
the objection apparently derives from the fact that Dr.
Loftus's answers regarding Sergeant Gubbei's memory
changed when the hypotheses under consideration were
altered regarding, for example, the number of seconds
Sergeant Gubbei saw the fleeing man's face and the
attention Sergeant Gubbei paid to the distributed suspect
materials. Some might view such a change in an expert's
analysis as the reward of a successful cross-examination; it
seems hard to view such variation as a basis for excluding
the testimony under Rule 403. The government did not
demonstrate, and has not done so now, how this dimension
of Dr. Loftus's testimony tended to cause any articulable
degree of unfair prejudice, confusion, or delay.

Another aspect of the government's objection stems from
a concern, which the District Court seemingly shared, that
Dr. Loftus's testimony might usurp the jury's function.
Insofar as this argument constitutes a general attack on
expert testimony concerning eyewitness identifications,
such issues are largely resolved by Downing. See id. at
1229-30 & n.4. Similar to other types of expert witnesses,
who might testify about the flaws of a computerized filing
system or the proper interpretation of satellite photographs,
experts who apply reliable scientific expertise to juridically

27
pertinent aspects of the human mind and body should
generally, absent explicable reasons to the contrary, be
welcomed by federal courts, not turned away. Cf. United
States v. Smithers, 212 F.3d 306, 311-12 & n.1 (6th Cir.
2000) (describing a judicial trend toward, and an
empirically supported need for, accepting expert testimony
about eyewitness identification). In this case, Dr. Loftus
made quite clear that he did not intend to tell the jury
whether Sergeant Gubbei was lying or telling the truth. Like
more typical sorts of expert witnesses, Dr. Loftus attempted
to provide information that, if itself deemed credible, might
cause the jury to evaluate Sergeant Gubbei's testimony in
a different light.

In analyzing for ourselves whether the four components
of Dr. Loftus's testimony7 were admissible, we follow
standards set forth in United States v. Stevens , which is
also discussed supra in Part II.B. In Stevens, the defendant
proffered expert testimony on six topics: "(1) the accuracy of
cross-racial identifications; (2) the effect of weapon focus on
identifications; (3) the effect of stress on identifications; (4)
the suggestiveness of the wanted board [from which the
defendant was initially identified]; (5) the relation back of
subsequent identifications to the initial identification; and
(6) the lack of correlation between confidence and accuracy
in eyewitness identifications." 935 F.2d at 1397. The
district court permitted expert testimony on the first three,
but excluded testimony on the remainder.

On appeal, the Stevens court affirmed the exclusion of
proffered testimony regarding the wanted board's
"suggestiveness," in deference to the district court's
determination that " `[t]he fact that there are two pictures
. . . of the same person and one [is] in color, I think[,] . .
. is certainly apparent [to] the jury . . . for the jury to make
its own determination.' " Id. at 1399; cf. United States v.
Gibbs, 190 F.3d 188, 212 (3d Cir. 1999) ("We have upheld
the exclusion of expert testimony when that testimony
ventures into areas in which the jury needs no aid or
illumination."). Similarly, in this case, Dr. Loftus's
_________________________________________________________________

7. Memory formation, double identification and post-event information,
the relationship between confidence and accuracy, and weapons focus.

28
testimony--claiming that more accurate memories are
produced by objects that are viewed for longer periods of
time--is not the type of information that, in and of itself,
would warrant exposition by an expert witness; thus, we
concur in the District Court's decision to exclude such
testimony under Rule 403.8

Stevens's expert testimony regarding "relation back"
apparently explained that "once a witness makes an
identification, he or she will tend to stick with that initial
choice at subsequent photographic arrays or lineups, even
if it was erroneous." Stevens, 935 F.2d at 1399. Such
evidence closely resembles the part of Dr. Loftus's
testimony on "double identification" and "post-event
information" that explained how, after Sergeant Gubbei had
already selected Mr. Mathis's picture from the array, his in-
court identification was likely to be consistent therewith,
regardless of any actual memory of the underlying events.
In Stevens, we affirmed the exclusion of expert testimony
on this topic because we found the testimony "rather
pedestrian" and "susceptible of elucidation without
specialized knowledge." Id. at 1399-1400. We are bound to
reach a parallel conclusion here; hence, we affirm the
exclusion of Dr. Loftus's testimony on this subject.

Another portion of Dr. Loftus's discussion of "double
identification," however, was significantly more complex
than the testimony at issue in Stevens. In particular, Dr.
Loftus described studies in which subjects remembered
having seen an image, but, upon being asked to associate
that image with one of two times they might possibly have
seen it, the subjects often "remembered" inaccurately. Dr.
Loftus further explained how Sergeant Gubbei might have
clearly remembered Mr. Mathis's face from having
previously seen a photograph of Mr. Mathis, but might have
confidently and incorrectly associated that memory with the
_________________________________________________________________

8. We do not mean to bar, by this determination, the strong possibility
that parts of Dr. Loftus's account of memory formation might have been
properly admissible as a necessary precursor, or ancillary, to other
admissible testimony regarding memory. Our ruling merely indicates
that such testimony, in and of itself, does not necessarily surmount the
barrier of Rule 403.

29
previous day's fleeing suspect. Dr. Loftus indicated that
such a misidentification was peculiarly likely in
circumstances where one possible event-memory
association (the previously viewed photograph) involved
favorable conditions for memory formation and involved the
same visual medium as the image to be identified (a
photograph), whereas the other event-memory association
(the robbers' escape) involved unfavorable memory
conditions and a different visual medium (three-
dimensional observation). We believe that testimony of this
sort, with its accompanying level of scientific detail, would
not simply duplicate jurors' intuitions or common sense,
and such principles seem difficult to establish indirectly
through cross-examination. Accordingly, we find that the
District Court abused its discretion in excluding such
testimony.

The third type of testimony analyzed in Stevens
concerned the relationship between an eyewitness's
confidence and the accuracy of a resultant identification.
Dr. Loftus opined on this same topic in the present case,
stating that the correspondence between confidence and
accuracy is "at best, about 25 percent" and, under
circumstances like these, is "essentially zero." In Stevens,
we held that such information was sufficiently illuminative
and susceptible to scientific expertise that exclusion was
unacceptable under "the liberal standard of admissibility of
Rule 702" and Rule 403. Id. at 1400-01. In the present
case, the government objects that Stevens cannot "mean
that [any time] an eyewitness has confidence in his
identification, a defendant is automatically entitled to
present testimony that confidence has little or no
correlation to accuracy." Appellee's Br. at 43. We agree that
Stevens did not establish such a universally applicable rule,
and we also decline to do so here. As the government
correctly notes, analysis under Rules 702 and 403 is fact-
intensive and case-specific. That said, however, Stevens did
hold that, when expert testimony of this character satisfies
the reliability and fit requirements of Rule 702, and when
there is no countervailing rationale for excluding the
evidence under Rule 403, the evidence must be admitted.
In light of our findings regarding this evidence's probative
value and our inability to discover any offsetting prejudice,

30
confusion, or delay, we hold that the District Court abused
its discretion in failing to admit Dr. Loftus's testimony on
the relationship between eyewitness confidence and
accuracy.

The final topic discussed by Dr. Loftus was "weapons
focus." In Stevens, this issue was not before us because the
District Court had admitted expert testimony on the
subject. In light of the principles decided in Stevens, as well
as those discussed supra, we find that Dr. Loftus's
proposed testimony regarding weapons focus should have
been admitted into evidence. The government argues that a
jury does not need expert testimony to learn "the idea that
one might be distracted by a weapon." Appellee's Br. at 34.
But Dr. Loftus's testimony was not limited to such a narrow
scope. In combination, the studies he discussed indicated
that the presence of weapons weakened memories of other
aspects of an observed scene, even when such weapons
posed no immediate threat, and even when those weapons'
images might be located quite near the object later to be
identified. We find that the degree and scope of memory
distortion that, according to Dr. Loftus, a weapon typically
causes for eyewitnesses are not matters that would
necessarily be apparent to jurors. And, just as with the
other admissible aspects of Dr. Loftus's testimony, it is
difficult to comprehend how weapons' destructive effect on
memory might be elucidated through cross-examination.
Especially if Dr. Loftus is correct that witnesses
remain confident of their identifications even when those
identifications are inaccurate, eyewitnesses cannot be
expected--even under the most skilled
questioning--explicitly to recall or nonverbally to reveal the
extent to which their remembered impressions might have
been distorted or undermined. We find that, in this case,
and on this record, it was an abuse of discretion not to
admit such testimony into evidence.

D. Harmless Error

Our final step is to consider whether the District Court's
error in excluding parts of Dr. Loftus's testimony was
legally harmless. See 28 U.S.C.A. S 2111 ("In the hearing of
any appeal[,] . . . the court shall give judgment after an
examination of the record without regard to errors or

31
defects which do not affect the substantial rights of the
parties."); Fed. R. Crim. P. 52(a); Fed. R. Evid. 103(a);
Kotteakos v. United States, 328 U.S. 750 (1946). This court
has held that a non-constitutional error committed at trial
does not warrant reversal in circumstances where"it is
highly probable that the error did not contribute to the
judgment." United States v. Helbling, 209 F.3d 226, 241 (3d
Cir. 2000) (internal quotation omitted). And the applicable
standard for " `[h]igh probability' requires that we have a
sure conviction that the error did not prejudice the
defendants. We may not simply conclude that it is more
likely than not that the error was harmless. On the other
hand, we may be firmly convinced that the error was
harmless without disproving every `reasonable possibility' of
prejudice." United States v. Jannotti, 729 F.2d 213, 220 n.2
(3d Cir. 1984) (internal citations omitted). The burden of
demonstrating such high probability lies with the
government, United States v. Adams, 2001 WL 543711 at
*4 (3d Cir. May 23, 2001), but we retain our established
authority to affirm on any ground supported by the record,
Nicini v. Morra, 212 F.3d 798, 805 (3d Cir. 2000) (en banc);
Alexander Hamilton Life Ins. Co. of Am. v. Gov't of Virgin
Islands, 757 F.2d 534, 547-58 (3d Cir. 1985).

The government contends that, even if (as we now have
found) the District Court wrongly failed to admit certain
testimony from Dr. Loftus, Mr. Mathis's conviction should
be affirmed because the government presented other
evidence that constituted more "compelling evidence of
Mathis's guilt than Sergeant Gubbei's identification."
Appellee's Br. at 47. In particular, the government relies on
testimony from Mr. Gantt, who described his cooperation
with Mr. Mathis in robbing the Sun National Bank and
eleven other banks. According to the government, Mr.
Gantt's "tremendous opportunity to observe, speak to, and
identify Mathis was damning and far outweighed Sergeant
Gubbei's testimony in importance." Id. Mr. Mathis replies
by challenging Mr. Gantt's credibility, both noting that Mr.
Gantt lied to police after his arrest and arguing that Mr.
Gantt's plea agreement provided a motive to lie if doing so
might advance Mr. Mathis's prosecution. In Mr. Mathis's
view, the government wrongly urges this court to"speculate

32
that the jury would have convicted defendant . . . had it
. . . heard Dr. Loftus' testimony . . . . [T]he jury may well
have concluded that the government had failed to bear its
burden of proving beyond a reasonable doubt that . ..
[Mathis] participated in the Sun National Bank robbery."
Reply Br. at 23. In essence, the government asks this court,
apparently as a matter of first instance, to weigh the import
of Mr. Gantt's testimony against that of Sergeant Gubbei's
testimony, while Mr. Mathis asks us not to "speculate"
about what the jury would have done if Dr. Loftus had
testified.

We find that neither of these suggestions, viewed in
isolation, is well-suited to our role in this appeal. As the
Supreme Court held in Kotteakos:

[I]t is not the appellate court's function to determine
guilt or innocence. Nor is it to speculate upon probable
reconviction and decide according to how the
speculation comes out. . . . Those judgments are
exclusively for the jury . . . . But this does not mean
that the appellate court can escape altogether taking
account of the outcome. . . . In criminal causes that
outcome is . . . . guilt in law, established by the
judgment of laymen. And the question is, not were they
right in their judgment . . . . It is rather what effect the
error had or reasonably may have had upon the jury's
decision. The crucial thing is the impact of the thing
done wrong on the minds of other men, not on one's
own, in the total setting.

Kotteakos, 328 U.S. at 763-64. Our inquiry is not whether
this court's members would convict Mr. Mathis on the
evidence presented, or whether we would do so based on
that evidence plus Dr. Loftus's wrongly excluded testimony.
Rather, we begin with the guilty verdict the jury has already
rendered, and we determine only whether, "after an
examination of the record," 28 U.S.C.A. S 2111, "it is highly
probable" that the jury would have reached the same
decision absent the identified error. Helbling , 209 F.3d at
241.

Although we believe that portions of Dr. Loftus's proffered
testimony should have been admitted, we also find that, in

33
the context of the record as a whole, his testimony was
highly unlikely to have caused a different result. The matter
is not resolved simply by noting--as the government
does--that Mr. Gantt's testimony could, if believed, have
had great probative force and might have constituted more
"compelling evidence" than Sergeant Gubbei's testimony. It
is certainly possible that Mr. Mathis's jury would have
convicted on the basis of Mr. Gantt's testimony alone, but
we would hesitate before terming such a result "highly
probable." Nor is it enough to find--as Mr. Mathis
urges--that skeptical jurors might, with reasonable
justification, have disbelieved both Mr. Gantt and Sergeant
Gubbei. Reasonable doubt is the proper test for harmless
error analysis concerning constitutional rights, e.g., Neder
v. United States, 527 U.S. 1, 16 (1999), but a less stringent
standard regulates non-constitutional errors such as this
one, e.g., Kotteakos, 328 U.S. at 764-65; Jannotti, 729 F.2d
at 220 n.2.

The crux of our analysis derives, not from any piece of
testimony in isolation, but from the interlocking
correspondence of evidence throughout the record. For
example, it is not disputed that the testimony of several
eyewitnesses, located inside and outside the bank building,
confirmed Mr. Gantt's basic description of how the Sun
National Bank robbery occurred. Photographs produced
from the bank's video security cameras were similarly
corroborative of this testimony. Also, Mr. Gantt's account of
the robbers' foiled attempt to escape in the Jeep was
virtually identical to Sergeant Gubbei's description of the
same events from the perspective of his police cruiser. Mr.
Gantt's identification of Mr. Seaberry as one of the co-
conspirators was confirmed by an arresting officer's
testimony that Mr. Seaberry was apprehended near the
stalled Jeep, and the arrest of Mr. Seaberry also tended to
corroborate Mr. Gantt's account of how he and Mr. Mathis
(but not Mr. Seaberry) finally escaped.

Giving Dr. Loftus's testimony its full weight, a jury might
have had reasonable cause to doubt Sergeant Gubbei's
description of events, if such description were viewed in
isolation. But here, where the disputed eyewitness account
was repeatedly confirmed by a co-conspirator's testimony,

34
and where various aspects of that co-conspirator's
testimony were in turn confirmed by other available
evidence, we think it highly improbable that a jury that
voted to convict without the aid of Dr. Loftus's testimony,
would, had it heard that testimony, have reached a
different verdict.

III. Conclusion

For the foregoing reasons, the District Court's judgment
of conviction is affirmed.

A True Copy:
Teste:

Clerk of the United States Court of Appeals
for the Third Circuit

35