Craig Brownlee was convicted by a jury of carjacking (18 U.S.C. § 2119), using a firearm in relation to a federal crime of violence (18 U.S.C. § 924(c)(1)(A)(ii)), and possession of a firearm by a convicted felon (18 U.S.C. § 922(g)(1)). He appeals his conviction and sentence and, for the reasons provided below, we reverse and remand for a new trial.1

On the morning of June 13, 2003, Virginia Daly stopped on her way home from work at the K-Mart located in New Kensington, Pennsylvania. After making her purchases, Daly left the store and proceeded toward her parked Jeep. As she began to get into her vehicle, a man approached her from behind and said "[H]ey." Daly turned around "and saw that [the person] had a gun in his hand." Now face to face with the man, Daly "told him to get away from [her and] he told [her] to get out of the car." The man then aimed the gun at Daly's chest, prompting her to get out of the car, turn over her keys, and run back to the K-Mart where she called the police. According to Daly, the suspect was black and wearing a dark t-shirt and a baseball cap.

Mary Ulizio, who had also stopped at the K-Mart to shop, viewed the entire incident and her version of events was similar to Daly's. As Ulizio was approaching her car in the parking lot, she saw a black male dressed in a dark navy blue t-shirt and a baseball cap "very quickly ... walk[]over towards [a] Jeep Grand Cherokee." Ulizio saw the man approach Daly and heard her say, "[L]eave me alone. Leave me alone." Daly and the man engaged in what was "basically a fight. She was trying to get him away from her. Then she started screaming, [`H]elp me.'" Ulizio also witnessed the man drive Daly's vehicle from the lot, and reported that "he pretty much pealed out of there pretty fast." Ulizio then returned to the K-Mart and awaited the police.

The carjacker drove Daly's car from the K-Mart lot toward Tarentum, a small town located across the Allegheny River from New Kensington. Daniel Spangler was traveling on the Tarentum Bridge (which connects Tarentum and New Kensington) when Daly's "vehicle ... passed [him] on the right-hand side ... at a very high rate of speed." The suspect "lost control of the vehicle ... and [it] fishtailed a couple of times and ... rolled over a number of times .. . and came to rest against a utility po[le]." Spangler got out of his car and, as he was approaching the scene of the accident, saw a man run from the wrecked Jeep toward downtown Tarentum. According to Spangler, the person was wearing "dark clothing" and was "[r]unning just fine." Spangler reported the accident to the police.

Scott Thomson was also driving his car in the vicinity of the Tarentum Bridge when the carjacker wrecked Daly's Jeep. Thomson was idling at a red light when he "saw a vehicle that was speeding ... across the bridge. Then, all of a sudden... it los[t] control right at the intersection. Rolled around a few times and wrapped around the utility pole...." Thomson left his car and

... started walking over to the scene and then I see someone get out [of] the vehicle and [he] just started running down Sixth Avenue. He stumbled to the ground. Just got up, took off running down Sixth Avenue.

Thomson remained at the scene in order to report the accident to the police.

Robert Walker was also in the vicinity of the bridge on the morning of June 13 when "he heard a loud noise." He turned to his right and saw "a car flip, hit the pole." Walker approached the wreck and "noticed a guy crawling out of the back door of the vehicle." According to Walker, as the man was extricating himself from Daly's Jeep, his baseball cap fell off of his head. He then ran from the scene at a "[p]retty good" clip.

By this time, the local police had issued a BOLO (Be On Lookout) broadcast concerning the Daly carjacking. In that broadcast, the suspect was described as "a black male with a dark blue shirt and ball cap." Daniel Glock, an officer with the East Deer Township Police Department, received the broadcast and drove to the scene of the accident to assist the police already there. Once at the scene, Glock received a report that the suspect had been observed "around First Avenue." This information prompted Glock to drive to First Avenue where he spoke with Constable Timothy Dzugan.

Dzugan, who lives in Tarentum, had been on his way to work when he received a radio report detailing the carjacking. As he approached the accident scene, he heard that the suspect "ran from the vehicle, heading north on East Sixth." This new information prompted Dzugan to go to this area, where he saw Brownlee — a thirty year old "black male" wearing a "dark shirt" — run across Second Avenue toward First Avenue. Dzugan notified the dispatcher concerning his observations and continued to follow Brownlee. At one point, Brownlee walked directly in front of Dzugan's vehicle in the direction of a house located at 329 First Avenue.

Brownlee was acquainted with the residents at this address, John and Arlene Boush. He knocked on the Boushes' door and awoke Arlene, who answered the door. Brownlee asked her if her husband was home and left after learning that he was not. He proceeded to walk through the Boushes' backyard. By this time, Dzugan and Glock got out of their cars and approached the Boushes' yard where they arrested Brownlee.

Brownlee then was taken by police cruiser to the accident scene, where Walker stated that he was the individual who had wrecked Daly's Jeep. Thomson also identified Brownlee as the man he witnessed crawling from the wrecked vehicle. Brownlee was handcuffed and in the back seat of the police cruiser during these identifications. According to Thomson,

I recognized him. I kind of went ... to see who was in the back seat of the police car and I was one hundred percent sure the guy in the back seat of the police car was the guy that crawled out of the vehicle.

The identifications occurred "approximately twenty-five minutes" after the accident involving Daly's vehicle.

Ulizio and Daly were taken from the K-Mart to the scene of the accident by a police officer. There both women identified Brownlee as the man who had taken Daly's car. According to Ulizio, "the policeman asked ... can we identify anyone. And the man was standing there, and we did." Daly remarked that "[t]here was no doubt in [her] mind" that it was Brownlee who had taken her vehicle. Brownlee was handcuffed, surrounded by police and standing beside the police cruiser at the time of Ulizio's and Daly's identifications.

Brownlee then was taken to the police station, read his Miranda2 rights, informed of the charges to be filed, and questioned by detectives. He told the police that he could not recall most of the prior evening. He did remember that had been at his girlfriend's place, but they had an argument, the police were called and he was asked to leave. Brownlee also noted that his father had picked him up and he remembered walking up to his home in Natrona Heights at approximately 4:00 a.m. He said he could not recall anything that had occurred between 4:00 a.m. and the time of his arrest.

At the scene of the accident, the police found a Yankees baseball cap on one side of the Jeep and, on the other side, a damaged, but operative, firearm on the ground or floorboard. Neither the car nor its contents were tested for fingerprints, and the car was subsequently destroyed. The firearm and cartridge were tested for comparable latent prints, but none were found.

The Government's principal evidence against Brownlee was the testimony of the four witnesses who provided on-the-scene identifications shortly after the accident. Brownlee moved to suppress each of the identifications as the product of unnecessarily suggestive procedures, but the District Court denied that motion. The Government bolstered the identification testimony presented at trial with the testimony of Constable Dzugan, who claimed that Brownlee had made various admissions to him while in custody arrest at the accident scene. Brownlee had moved to suppress those statements pretrial on the grounds that they were obtained in violation of his Miranda rights and his Fifth Amendment right against self-incrimination, but the District Court denied that motion, ruling that Dzugan did not subject Brownlee to "interrogation."

At trial, Brownlee presented a mistaken identity defense. In support of this theory, he sought to present the opinions of Dr. Jonathan Wolf Schooler, an expert in the field of human perception and memory. Brownlee offered this testimony to address the circumstances surrounding each of the Government's identification witnesses, specifically (1) show-up identification procedures and how they can influence a witness' accuracy, (2) a comparison between the show-up and other identification procedures, (3) the tendency of a witness to focus on a weapon, (4) the lack of correlation between witness confidence in identification and the accuracy of that identification, (5) the effect of exposure to multiple witnesses, (6) the effect of hair covering on eyewitness recognition ability, (7) the phenomena of confidence malleability (i.e., the effect of post-event information on a witness' confidence in the accuracy of an identification), (8) time delay on identification, (9) the effect of post-event suggesting, and (10) cross-racial identification. After a Daubert3 hearing, the District Court allowed Dr. Schooler to testify about cross-racial identification, the effects of hair covering, weapons focus, and exposure to multiple witnesses, but refused to allow expert testimony as to the other categories.

After a three-charge indictment was filed against Brownlee, a jury found him guilty of each charge. The District Court sentenced him after the Supreme Court decided Blakely v. Washington, 542 U.S. 296, 124 S. Ct. 2531, 159 L. Ed. 2d 403 (2004), but prior to United States v. Booker, 543 U.S. 220, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005). At sentencing, Brownlee argued that he could not be sentenced pursuant to the federal Sentencing Guidelines because they were unconstitutional under Blakely. The District Court agreed that the Sentencing Guidelines were unconstitutional and, as a result, "sentence[d] [Brownlee] according to the statutory range, without regard to the Guidelines," to 37 years (444 months) imprisonment and three years of supervised release. This appeal followed.

A. Did the District Court err by refusing to suppress the identification testimony of four witnesses due to unnecessarily suggestive identification procedures?

"As with many evidentiary rulings, we review a decision to admit identification testimony over an objection for abuse of discretion." United States v. Emanuele, 51 F.3d 1123, 1127 (3d Cir. 1995). Where a motion to suppress has been denied, we review the order "for clear error as to the underlying facts, but exercise plenary review as to its legality in the light of the court's properly found facts." United States v. Inigo, 925 F.2d 641, 656 (3d Cir.1991). If the admission of identification testimony violated due process, as Brownlee contends, we then consider whether this constitutional error was harmless. Foster v. California, 394 U.S. 440, 444, 89 S. Ct. 1127, 22 L. Ed. 2d 402 (1969).

An identification procedure that is both (1) unnecessarily suggestive and (2) creates a substantial risk of misidentification violates due process. Manson v. Brathwaite, 432 U.S. 98, 107, 97 S. Ct. 2243, 53 L. Ed. 2d 140 (1977). Unnecessary suggestiveness "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." United States v. Stevens, 935 F.2d 1380, 1389 (3d Cir.1991) (internal quotations and emphases omitted). An impermissibly suggestive identification procedure can occur in four settings: a show-up, a photo array, a line-up and in court. Identifications, 34 Geo. L.J. Ann. Rev.Crim. Proc. 149, 153 n. 496 (2005). The first setting (the one presented in this case) is a "show-up," in which a single individual arguably fitting a witness's description is presented to that witness for identification.

As the Supreme Court has acknowledged, a show-up procedure is inherently suggestive because, by its very nature, it suggests that the police think they have caught the perpetrator of the crime. Stovall v. Denno, 388 U.S. 293, 302, 87 S. Ct. 1967, 18 L. Ed. 2d 1199 (1967) (recognizing that "[t]he practice of showing suspects singly to persons for the purpose of identification, and not as part of a line-up, has been widely condemned"). Brownlee was handcuffed and seated in the back seat of a police cruiser when identified by Thomson and Walker, and he was handcuffed and pulled out of the police cruiser when Daly and Ulizio identified him. Not only was Brownlee handcuffed, surrounded by police officers, and either seated inside or standing beside a police cruiser at the time of the identifications, he was also at the scene of the accident — a condition that creates the impression the police had caught him in the stolen Jeep.

Three other points are noteworthy and exacerbate the suggestiveness of the show-up in this case. First, no "suspect" save Brownlee was presented to any of the eyewitnesses at any time. Second, all four eyewitnesses were allowed to make identifications while exposed to the suggestive influence of others. See Emanuele, 51 F.3d at 1131 (holding that the witness' inability to recognize defendant in photo array, "coupled with the highly suggestive viewing of the defendant in conditions reeking of criminality, bolstered by the comments of another witness, render[ed] the in-court identification unreliable" (emphasis added)). Thomson and Walker were left at the scene to talk with bystanders as well as police from the time of the crash until Brownlee was transported to their location; Ulizio and Daly were not only questioned together but were taken to identify Brownlee together. Finally, there is no reason evident why Brownlee and the witnesses could not have been taken to the police station for a less suggestive line-up or photo array.4 See United States v. Sebetich, 776 F.2d 412, 420 (3d Cir.1985) (stating that line-up or similar procedure should "be employed whenever necessary to ensure the accuracy and reliability of identifications"), cert. denied, 484 U.S. 1017, 108 S. Ct. 725, 98 L. Ed. 2d 673 (1988); see also Stovall, 388 U.S. at 302, 87 S. Ct. 1967 (recognizing show-up procedure inherently suggestive though imperative where witness was physically unable to leave hospital and it was uncertain how long she would live). Thus, we agree with Brownlee that the show-up procedure here was unnecessarily suggestive.

But unnecessary suggestiveness alone does not require the exclusion of evidence. Neil v. Biggers, 409 U.S. 188, 198-99, 93 S. Ct. 375, 34 L. Ed. 2d 401 (1972). A "suggestive and unnecessary identification procedure does not violate due process so long as the identification possesses sufficient aspects of reliability," Brathwaite, 432 U.S. at 106, 97 S. Ct. 2243, for reliability is the "linchpin in determining the admissibility of identification testimony," id. at 114, 97 S. Ct. 2243. As the Supreme Court explained in Biggers, in order to determine whether an identification was reliable even though the confrontation procedure was suggestive, we must look to the totality of the circumstances. 409 U.S. at 199, 93 S. Ct. 375. The Court considers factors that include: (1) the opportunity of the witness to view the criminal at the time of the crime; (2) the witness' degree of attention; (3) the accuracy of the witness' prior description of the criminal; (4) the level of certainty demonstrated by the witness at the confrontation; and (5) the length of time between the crime and confrontation. Id. at 199, 93 S. Ct. 375.

In Biggers, the Court set out and applied each of these factors, noting that the witness had ample opportunity to view the defendant, paid a high degree of attention, gave a detailed description of the defendant, and was unequivocal in her identification. Id. at 200, 93 S. Ct. 375. The Court pointed out that several months had passed between the time of the crime and the identification, but reasoned that, weighing all the factors, there was no substantial likelihood of misidentification and that the evidence was properly allowed to go to the jury. Id. at 201, 93 S. Ct. 375.

Similarly, in Brathwaite, the Court enumerated and applied each of the Biggers factors to determine whether an identification from a single-photograph display was unreliable. 432 U.S. at 114-16, 97 S. Ct. 2243. There the witness looked directly at the defendant (who was in close proximity), paid a high degree of attention, gave a detailed and accurate description of the defendant within minutes of the encounter and unequivocally identified the defendant's photograph two days later. Id. at 114-15, 97 S. Ct. 2243. The Court concluded that — given these factors as well as the absence of any coercive pressure positively to identify the photograph — there was not a "very substantial likelihood of irreparable misidentification." Id. at 116, 97 S. Ct. 2243 (internal quotations omitted).

We reached the opposite conclusion on the facts in United States v. Emanuele. The two witnesses in that case were bank tellers, each of whom had observed a robbery at her place of employment. 51 F.3d at 1126-27. Neither was able to identify the robber from a photo array. Id. These witnesses were subpoenaed by the Government to testify. While sitting outside the courtroom, they "saw the defendant being led from the courtroom in manacles by U.S. Marshals." Id. at 1127. They then spoke with one another, saying "it has to be him." Id. We first determined that the confrontation between the witnesses and the manacled defendant was impermissibly suggestive. Id. at 1129-30. We then concluded that the witnesses' failure to pick the defendant out of the photo array, coupled with the impermissibly suggestive "viewing of [him] in conditions reeking of criminality, bolstered by the comments of another witness, rendered the in-court identification unreliable." Id. at 1131.

Returning to our facts, the critical question is whether the circumstances surrounding the identifications at issue here are more like Biggers and Brathwaite, or Emanuele. To answer that inquiry, we turn to the Biggers factors. Certainly, some of the circumstances presented here weaken the reliability of the eyewitnesses' identifications. For instance, Daly conceded that the entire carjacking lasted only thirty seconds, and that she spent a predominant amount of that time focused on the weapon (which, incidentally, she misidentified). Moreover, Daly at first told the 911 dispatcher that her assailant was wearing shorts (whereas Brownlee wore blue jeans). Ulizio testified that she initially believed the carjacker was a young kid (while Brownlee was 30 at the time the crime was committed), and both Thomson and Walker saw more of the suspect's back than his front as he ran away from them down the street. Finally, none of the witnesses could describe the suspect's facial features or provide the police with more than a relatively general description of him.

These facts notwithstanding, the totality of the circumstances establish that the identifications were reliable. The evidence provided at the suppression hearing indicates that (1) the witnesses' opportunity to observe the perpetrator at the time of the crime was sufficient, at fairly close range, and in broad daylight; (2) their degree of attention was substantial; (3) their prior descriptions, while rather general, were fairly accurate; (4) their degree of certainty was absolute; and (5) relatively little time passed between the crime and confrontations (approximately 25 minutes). The generality of the witnesses' descriptions of the suspect, the relatively short period of time they saw him, and the other shortcomings pertaining to their identifications, go more to the weight of the evidence than the reliability of their identifications, and thus were issues for the jury. Accordingly, we conclude that the identifications were properly admitted at trial despite the fact that the show-up procedure was unnecessarily suggestive.

B. Did the District Court err by refusing to allow the defendant's expert witness in the field of human perception and memory to testify regarding the reliability of the identifications?

Brownlee contends the District Court erred in restricting the testimony of Dr. Schooler, a professor of psychology at the University of Pittsburgh and an expert in human memory and perception. In a pre-trial pleading, Brownlee reported that he intended to call Dr. Schooler to testify about "issues of cross-racial identification and the reliability of identifications made under a stressful environment."5 As noted earlier, the District Court allowed expert testimony concerning cross-racial identification, the effects of hair covering, weapons focus, and exposure to multiple witnesses, but refused to allow expert testimony in the other categories.

We review the District Court's decision to exclude expert testimony for an abuse of discretion. In re Paoli R.R. Yard PCB Litig., 916 F.2d 829, 856 n. 33 (3d Cir.1990), cert. denied, 499 U.S. 961, 111 S. Ct. 1584, 113 L. Ed. 2d 649 (1991).6

The Government argues that the District Court properly excluded testimony regarding (1) the comparison between the show-up and other identification procedures because "it held the potential for confusion, was irrelevant, and not helpful to the factfinder," and (2) the suggestiveness of the show-up involved in this case and the effect it potentially played in the identifications because "the jury could have determined for itself without expert opinion whether the show-up in this case was capable of influencing the witnesses' identification." As for (3) confidence malleability, (4) post-event suggestiveness, and (5) confidence of accuracy, although the District Court ruled such testimony excluded, the Government points out that "Brownlee ... managed to elicit testimony concerning" those three categories from his expert at trial.

We are not persuaded by the Government's arguments concerning the exclusion of these five categories of excluded testimony. This case was primarily about the accuracy and reliability of the identifications.7 The District Court's rulings, specifically with regard to confidence of accuracy, significantly undermined Brownlee's ability to challenge effectively the witnesses' certainty and confidence in their identifications — a point the Government used to its benefit both in presenting testimony and arguing to the jury in its closing at trial. Moreover, the record belies the Government's contention that Brownlee managed to elicit any expert testimony concerning confidence of accuracy.

It is widely accepted by courts, psychologists and commentators that "[t]he identification of strangers is proverbially untrustworthy." Felix Frankfurter, The Case of Sacco and Vanzetti: A Critical Analysis for Lawyers and Laymen 30 (Universal Library ed., Grosset & Dunlap 1962) (1927) ("What is the worth of identification testimony even when uncontradicted?... The hazards of such testimony are established by a formidable number of instances in the records of English and American trials. These instances are recent — not due to the brutalities of ancient criminal procedure."); see also United States v. Wade, 388 U.S. 218, 228, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967) (stating that "[t]he vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification"); C. Ronald Huff et al., Guilty Until Proven Innocent: Wrongful Conviction and Public Policy, 32 Crime & Delinq. 518, 524 (1986) ("the single most important factor leading to wrongful conviction in the United States ... is eyewitness misidentification"). The recent availability of post-conviction DNA tests demonstrate that there have been an overwhelming number of false convictions stemming from uninformed reliance on eyewitness misidentifications. In 209 out of 328 cases (64%) of wrongful convictions identified by a recent exoneration study, at least one eyewitness misidentified the defendant. Samuel R. Gross et al., Exonerations in the United States: 1989-2003 95 J.Crim. L. & Criminology 523, 542 (2004). In fact, "mistaken eyewitness identifications are responsible for more wrongful convictions than all other causes combined." A. Daniel Yarmey, Expert Testimony: Does Eyewitness Memory Research Have Probative Value for the Courts?, 42 Canadian Psychology 92, 93 (May 2001). "[E]yewitness evidence presented from well-meaning and confident citizens is highly persuasive but, at the same time, is among the least reliable forms of evidence." Id. (Emphasis added.)

Even more problematic, "jurors seldom enter a courtroom with the knowledge that eyewitness identifications are unreliable." Rudolph Koch, Note, Process v. Outcome: The Proper Role of Corroberative Evidence in Due Process Analysis of Eyewitness Identification Testimony, 88 Cornell L.Rev. 1097, 1099 n. 7 (2003). Thus, while science has firmly established the "inherent unreliability of human perception and memory," id. at 1102 (internal quotations omitted), this reality is outside "the jury's common knowledge," and often contradicts jurors' "commonsense" understandings, id. at 1105 n. 48 (internal quotations omitted). To a jury, "there is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says[,] `That's the one!'" Watkins v. Sowders, 449 U.S. 341, 352, 101 S. Ct. 654, 66 L. Ed. 2d 549 (1981) (Brennan, J., dissenting) (emphasis in original).

Faced with "[t]he tragic irony of eyewitness testimony," Koch, Process v. Outcome, supra, at 1098 n. 6 (quoting Lawrence Taylor, Eyewitness Identification 1 (1982)), and no physical scientific means of exonerating himself,8 Brownlee sought to present expert scientific evidence to establish the inherent unreliability of human perception and memory by demonstrating that the correlation between confidence and accuracy is weak.9 Federal Rule of Evidence 702 "authorizes the admission of expert testimony so long as it is rendered by a qualified expert and is helpful to the trier of fact." DeLuca v. Merrell Dow Pharm., Inc., 911 F.2d 941, 954 (3d Cir.1990). Application of this Rule to Dr. Schooler's proposed testimony required the District Court to apply United States v. Downing, 753 F.2d 1224 (3d Cir. 1985). There we recognized that Rule 70210 may permit a defendant "to adduce, from an expert in the field of human perception and memory, testimony concerning the reliability of eyewitness identifications." Id. at 1226. The test outlined in Downing instructs the trial court, after conducting a preliminary hearing, to balance two factors:

(1) the reliability of the scientific principles upon which the expert testimony rests, hence the potential of the testimony to aid the jury in reaching an accurate resolution of a disputed issue; and (2) the likelihood that introduction of the testimony may in some way overwhelm or mislead the jury.

Id. In addition, "admission depends upon the `fit,' i.e., a specific proffer that the testimony will focus on particular characteristics of the eyewitness identification at issue and discuss how those characteristics call into question the reliability of the identification." Sebetich, 776 F.2d at 419. More specifically,

a defendant who seeks the admission of expert testimony must make an on-the-record detailed proffer to the court, including an explanation of precisely how the expert's testimony is relevant to the eyewitness identifications under consideration. The offer of proof should establish the presence of factors (e.g., stress, or differences in race or age as between the eyewitness and the defendant) which have been found by researchers to impair the accuracy of eyewitness identifications.

Downing, 753 F.2d at 1242.

In Downing (and unlike this case), no specific proffer was made in the District Court. Id. Nonetheless, we remanded the case. In doing so, we cited with approval the admission of expert psychological testimony concerning, inter alia, "the fact that studies demonstrate the absence of a relationship between the confidence a witness has in his or her identification and the actual accuracy of that identification...." Id. at 1230-31; see also id. at 1242 & n. 23 (noting "the proliferation of empirical research demonstrating the pitfalls of eyewitness identification," "the [impressive] consistency of the results of these studies," and agreeing that "the science of eyewitness perception has achieved the level of exactness, methodology and reliability of any psychological research" (internal citations omitted)).11

Subsequent to Downing, we reaffirmed in Stevens the role of expert testimony regarding the lack of confidence-accuracy correlation. 935 F.2d at 1384. In that case, we reviewed the District Court's decision to exclude expert testimony on confidence of accuracy studies because it found no "fit" between the proffered testimony and the facts of that case. Id. at 1398. We reversed, pointing out the weak correlation (or "fit") between confidence of the witness and his/her accuracy. Exclusion of the expert testimony, we determined, was error under, inter alia, Federal Rule of Evidence 702:

We think that the district court misapprehended Downing's "fit" requirement. Both [eyewitnesses] expressed high confidence in their identifications of [the defendant] as the perpetrator. To rebut the natural assumption that such a strong expression of confidence indicates an unusually reliable identification, [the defendant] sought to admit [expert] testimony that there is a low correlation between confidence and accuracy. We believe that [the expert's] proposed testimony "is sufficiently tied to the facts of the case that it will aid the jury in resolving a factual dispute." Downing, 753 F.2d at 1242.

* * *

Moreover, ... [the expert's] explication of the confidence/accuracy studies could prove helpful to the jury in assessing the reliability of [the eyewitnesses'] identifications. That witnesses ofttimes profess considerable confidence in erroneous identifications is fairly counterintuitive.

See id. at 1230 n. 6 ("To the extent that a mistaken witness may retain great confidence in an inaccurate identification, cross-examination can hardly be seen as an effective way to reveal the weakness in a witness' recollection of an event."). In fact, [the expert] opined at the preliminary hearing that the correlation between confidence and accuracy in eyewitness identifications is far lower than people probably would expect. Given this potential for helpfulness and "the liberal standard of admissibility mandated by Rule 702," id. at 1230, we hold that the district court abused its discretion in barring [the expert's] tendered testimony on the confidence/accuracy factor.

Stevens, 935 F.2d at 1406-07.

The same analysis necessarily controls here. Given that "witnesses ofttimes profess considerable confidence in erroneous identifications," expert testimony was the only method of imparting the knowledge concerning confidence-accuracy correlation to the jury. Due to the nature of the Government's evidence12 and Brownlee's defense (mistaken identity), the primary issue before the jury was the reliability of the Government's four eyewitnesses. "[I]t would seem anomalous to hold that the probative value of expert opinion offered to show the unreliability of eyewitness testimony so wastes time or confuses the issue that it cannot be considered even when the putative effect is to vitiate the [primary] evidence offered by the government." Downing, 753 F.2d at 1243. In light of these considerations, we hold it was wrong to exclude expert testimony regarding the reliability of the very eyewitness identification evidence on which Brownlee was convicted, and remand the case for a new trial.13

C. Did the District Court err by refusing to suppress incriminating statements Brownlee allegedly made to a police officer whom he knew at the time he was brought to the scene of the accident?

At trial, the Government bolstered its eyewitness identification evidence with the testimony of Constable Dzugan, who claimed that Brownlee made various confessions to him while in police custody at the scene of the accident. Dzugan testified that he recognized Brownlee from playing neighborhood basketball with Brownlee's older brother approximately twenty years previously, but that he had not seen Brownlee for about ten years. According to Dzugan, while seated in the police cruiser at the accident scene, Brownlee struck up a conversation with him by yelling, "Hey, Dzugan, can you turn the air conditioning on?" "Then he asked me if I could call his father." Dzugan obtained a piece of paper and wrote down the telephone number of Brownlee's father. The Constable indicated that, at this point, he and Brownlee proceeded to talk for "about a good fifteen minutes." Dzugan was asked whether "[a]t any point... Brownlee sa[id] anything related to the accident?" Dzugan replied "Yes," and testified as follows:

I told him, I said, "[D]id you look up there at that Jeep?" I said, "[H]ow crushed is it. You could have been killed. How did you get out of here?" He said, "I climbed through the back window."

According to Dzugan, the following exchange also occurred:

Well, I told him, I says, "you know, you just were in trouble." I said, "you just got out of jail. Why would you do something dumb, like this? With a gun?" And he told me, he says, "[I]t wasn't my gun. It didn't work. I got it from a friend in New Kensington."

Brownlee moved in the District Court to suppress this testimony on the basis that the allegedly inculpatory statements were obtained in violation of Miranda v. Arizona, 384 U.S. 436, 86 S. Ct. 1602, 16 L. Ed. 2d 694 (1966). In response, the Government conceded that (1) Brownlee was in custody at the time he made the alleged statements to Dzugan, (2) who was a law enforcement officer for the purposes of Miranda, and (3) who failed to provide Miranda warnings.14 Despite this, it argued that suppression of the statements should be denied on the ground that the incriminating answers were not the product of an "interrogation." The District Court adopted the Government's position in toto, stating:

Clearly, the defendant was in custody when he was in the back of the police car. When Constable Dzugan was talking to him at the scene of the accident, without deciding, the Court will assume that the constable is a law enforcement officer for the purpose of conducting an custodial interrogation following a felony arrest. Therefore, Miranda warnings would be appropriate before he should conduct a custodial interview.

However, the Court finds that the constable had a conversation with Mr. Brownlee, who[m] he was aware of and knew for more than twenty years. That Mr. Brownlee instigated and initiated the conversation when he asked the constable to call his father. And the constable did not use any actions or words which he knew or should have known were reasonably likely to elicit an incriminating response from the defendant.

The constable did not, therefore, engage in interrogation of the suspect, and his statement from the defendant to the constable will be admissible evidence.

Thus, the question before us is whether Dzugan "interrogated" Brownlee,15 an inquiry that we answer in the affirmative for the reasons that follow.

Under the prophylactic rules announced in Miranda, a statement made by a suspect in response to custodial interrogation after he or she has elected to remain silent is inadmissible at trial. 384 U.S. at 478-79, 86 S. Ct. 1602. As the Supreme Court held in Rhode Island v. Innis, 446 U.S. 291, 100 S. Ct. 1682, 64 L. Ed. 2d 297 (1980), this rule comes "into play whenever a person in custody is subjected to either express questioning or its functional equivalent. That is to say, the term `interrogation' under Miranda refers... to any words or actions on the part of the police (other than those normally attendant to arrest and custody) that the police should know are reasonably likely to elicit an incriminating response from the suspect." Id. at 300-01, 100 S. Ct. 1682 (internal footnote omitted). An incriminating response is "any response — whether inculpatory or exculpatory — that the prosecution may seek to introduce at trial." Id. at 301 n. 5, 100 S. Ct. 1682 (emphasis in original). Police may not, however, "be held accountable for the unforeseeable results of their words or actions[,]" id. at 302, 100 S. Ct. 1682, and to constitute an interrogation their conduct "must reflect a measure of compulsion above and beyond that inherent in custody itself." Id. at 300, 100 S. Ct. 1682.

While Dzugan alleges that Brownlee initiated the "conversation" that took place shortly after his arrest by asking the officer to adjust the air conditioning and, later, to call his father, Dzugan concedes that it was he who took the opportunity to bring up the subject of the crime. Indeed, Dzugan admitted that he initiated the conversation concerning the stolen car, the gun and the carjacking, and expressly asked Brownlee:

(1) How did you get out of there?

(2) Did you get hurt? (3) Why would you do something dumb like this?

(4) With a gun?

It is difficult to imagine questions that are more likely to evoke an incriminating response — that is, a "statement[] ... amount[ing] to `admissions' of part or all of the offense" — from a suspect than those posed by Dzugan to Brownlee. Id. at 301 n. 5, 100 S. Ct. 1682 (quoting Miranda, 384 U.S. at 476-77, 86 S. Ct. 1602).

The Government emphasizes that Dzugan was not attempting to elicit incriminating statements from Brownlee. The Supreme Court made clear in Innis, however, that the interrogation analysis focuses "primarily upon the perceptions of the suspect, rather than the intent of the police." Id. at 301, 100 S. Ct. 1682; see also Illinois v. Perkins, 496 U.S. 292, 296, 110 S. Ct. 2394, 110 L. Ed. 2d 243 (1990) (stating "[c]oercion is determined from the perspective of the suspect"). The focus on a suspect's perceptions "reflects ... that the Miranda safeguards were designed to vest a suspect in custody with an added measure of protection against coercive police practices, without regard to objective proof of the underlying intent of the police." Id.

While the "focus" of the Innis test is on the suspect's perceptions, the intent of a police officer is nonetheless relevant. See id. at 301 n. 7, 100 S. Ct. 1682. The intent of the officer, particularly when "a police practice is designed to elicit an incriminating response," may bear on the question of "whether the police should have known that their words or actions were reasonably likely to evoke an incriminating response." Id. Additionally, "[a]ny knowledge the police may have had concerning the unusual susceptibility of a defendant to a particular form of persuasion might be an important factor in determining whether the police should have known that their words or actions were reasonably likely to elicit an incriminating response from the suspect." Id. at 301 n. 8, 100 S. Ct. 1682.

The premise of Miranda is that a suspect speaking with those whom he knows to be law enforcement officers "will feel compelled to speak by the fear of reprisal for remaining silent or in the hope of more lenient treatment should he confess." Perkins, 496 U.S. at 296-97, 110 S. Ct. 2394 (emphasis added). In this connection, it is likely that the acquaintanceship between Dzugan and Brownlee as described by Dzugan increased rather than decreased the importance of Miranda warnings. Indeed, "[t]he Miranda warnings are intended to `warn' a suspect that the police have interests that are antagonistic to his, and that they can use anything he says against him in court." United States v. Mesa, 638 F.2d 582, 588 n. 5 (3d Cir.1980). No suspect needs Miranda warnings more than one questioned by a law enforcement officer that the suspect assumes is a quasiconfidante. Our Court has recognized that an agent's relationship with a suspect is a factor in the coercion analysis, even absent any direct inquiry or deceptive intent on the part of the agent. See United States v. Walton, 10 F.3d 1024, 1028 n. 1 (3d Cir.1993) (after acknowledging that conversation between agent and suspect constituted "interrogation" within the meaning of Innis, we stated "[w]e believe it self-evident that an assurance to a suspect that an agent has known him `for a long time' [—] and that if he desires, he `can tell us what happened off the cuff' [—] is the functional equivalent of questioning"); Miller v. Fenton, 796 F.2d 598, 607 (3d Cir.1986) ("Excessive friendliness on the part of an interrogator can be deceptive," potentially creating "an atmosphere in which a suspect forgets that his questioner is in an adversarial role, and thereby prompt admissions that the suspect would ordinarily make only to a friend, not to the police."); see also Miranda, 384 U.S. at 469, 86 S. Ct. 1602 (stating that the warnings it requires are intended "to make the individual more acutely aware that he is faced with a phase of the adversary system — that he is not in the presence of persons acting solely in his interests"). Moreover, that Brownlee and Dzugan were friendly made Dzugan acutely aware of the unusual susceptibility of Brownlee to his inquiries.

Simply stated, we conclude that Dzugan subjected Brownlee to an "interrogation" without providing the warnings demanded by Miranda. Because the District Court committed constitutional error, we must reverse unless the Government establishes that the improper admission of Brownlee's statements was "harmless beyond a reasonable doubt," i.e., proves beyond a reasonable doubt that the inculpatory statements "did not contribute to" Brownlee's conviction. Walton, 10 F.3d at 1032. As noted above, the evidence linking Brownlee to the carjacking was (1) his presence in the area wearing dark clothing, (2) eyewitness testimony (which Brownlee was unable to attack as he wished via expert testimony), and (3) the statements he allegedly made to Dzugan. The record makes clear that the Government used Brownlee's statements to bolster its eyewitness testimony. Moreover, it is difficult for the Government to argue with effect that the admission of the confession did not contribute to Brownlee's conviction when it submitted just the opposite view to the jury during the trial. As the Supreme Court has recognized,

[a] confession is like no other evidence. Indeed, the defendant's own confession is probably the most probative and damaging evidence that can be admitted against him .... [T]he admissions of a defendant come from the actor himself, the most knowledgeable and unimpeachable source of information about his past conduct. Certainly, confessions have profound impact on the jury, so much so that we may justifiably doubt its ability to put them out of mind even if told to do so.

Arizona v. Fulminante, 499 U.S. 279, 296, 111 S. Ct. 1246, 113 L. Ed. 2d 302 (1991) (internal quotations omitted).

For these reasons, we believe admission of the confession was not harmless in this case. Therefore, we reverse the District Court's erroneous admission of Brownlee's inculpatory statements and remand the case for a new trial.

D. Was Brownlee unconstitutionally prosecuted for intrastate crimes having no substantial relationship to interstate commerce?

Brownlee also contends that his federal prosecution for the crimes of carjacking (18 U.S.C. § 2119), using a firearm in relation to a federal crime of violence (18 U.S.C. § 924(c)(1)(A)(ii)), and possession of a firearm by a convicted felon (18 U.S.C. § 922(g)(1)), was unconstitutional. More specifically, he argues that the statutes upon which his convictions are based are unconstitutional, both facially and as applied, because the convictions were for intrastate crimes, and thus those statutes exceed the regulatory authority granted Congress under the Constitution's Commerce Clause. U.S. Const. art. I, § 8, cl. 3. Brownlee concedes that separate panels of our Court have previously addressed the constitutionality of the felon-in-possession statute, United States v. Singletary, 268 F.3d 196 (3d Cir.2001), and the carjacking statute, United States v. Bishop, 66 F.3d 569 (3d Cir.1995), and found them to be constitutional. He nonetheless requests reconsideration of those issues.

In Singletary, a panel of this Court scrutinized the same line of Commerce Clause decisions of the Supreme Court to which Brownlee directs our attention, and ruled that § 922(g) was constitutional. See 268 F.3d at 200-205 (analyzing Jones v. United States, 529 U.S. 848, 120 S. Ct. 1904, 146 L. Ed. 2d 902 (2000); United States v. Morrison, 529 U.S. 598, 120 S. Ct. 1740, 146 L. Ed. 2d 658 (2000); United States v. Lopez, 514 U.S. 549, 115 S. Ct. 1624, 131 L. Ed. 2d 626 (1995); Scarborough v. United States, 431 U.S. 563, 97 S. Ct. 1963, 52 L. Ed. 2d 582 (1977)). A similar analysis was provided — albeit without the benefit of the Jones and Morrison decisions — by the Bishop panel, which held that § 2119 was constitutional. In light of the binding effect we give to precedential opinions of panels of this Court, and because we discern no principled distinction between the statutes already ruled on by our Court and § 924(c)(1)(A)(ii), we must reject Brownlee's argument and hold that the statutes he challenges survive constitutional scrutiny. See 3d Cir. Internal Operating P. 9.1.16

We affirm the District Court's ruling that the eyewitness identifications were reliable and admissible at trial despite the fact that the show-up procedure was unnecessarily suggestive, and reject Brownlee's claim that he was unconstitutionally prosecuted for intrastate crimes having no substantial relationship to interstate commerce. We reverse, however, the District Court's (1) exclusion of expert testimony regarding the reliability of the eyewitness identification evidence upon which Brownlee was convicted and (2) admission of Brownlee's inculpatory statements to Constable Dzugan because they were the product of a custodial interrogation without Miranda warnings. As a result of those determinations, we remand this case for a new trial.