dissenting. In my view, the trial court clearly abused its discretion by prohibiting Michael Leippe, a professor of psychology and an expert in the recognized field of eyewitness behavior, from testifying before the jury about the unreliability of eyewitness identification. The jury should have had the benefit of Leippe’s testimony because this case turned entirely on the seemingly credible identification testimony of Darlene Hale, whose identification of the defendant was made two years after the event and was not supported by any corroborating evidence. For a panoply of reasons, Leippe would have cautioned the jurors to regard this testimony with a substantial quantum of suspicion. See United States v. Butler, 636 F.2d 727, 732 (D.C. Cir. 1980), cert. denied, 451 U.S. 1019, 101 S. Ct. 3010, 69 L. Ed. 2d 392 (1981) (Bazelon, J., dissenting) (“many experts have concluded that convictions based solely on one eyewitness identifications represent conceivably the greatest single threat to the achievement of our ideal that no innocent man shall be punished” [internal quotation marks omitted]).
*591I
In the present appeal, the reliability of Hale’s identification is the central issue. In fact, it was the only evidence presented by the state to prove that the defendant had any involvement in the crime. No credible forensic evidence connected the defendant to the murder: the ballistics evidence offered by the state was, at best, equivocal,1 and a partial latent palm print lifted from the employment application that the defendant allegedly completed at the crime scene did not match the defendant’s prints. Although the defendant was observed near the scene of the crime, this was of no significance because he lived in the neighborhood and was walking in the direction of his home.2
*592“The dangers of misidentification are well known and have been widely recognized by this court and other courts throughout the United States.” State v. Tatum, 219 Conn. 721, 733, 595 A.2d 322 (1991). “[T]he experience of law and psychology has been that eyewitness testimony may sometimes be the least trustworthy means to identify the guilty. See 3 J. Wigmore, Evidence § 786a (1970). [Indeed, as early as 1904] England’s Royal Committee of Inquiry, established ... to investigate the case of Adolf Beck, twice convicted of crimes he did not commit, adjudged that evidence as to identity based on personal impressions ... is perhaps of all classes of evidence the least to be relied upon, and therefore, unless supported by other facts, an unsafe basis for the verdict of a jury. Parliamentary Papers (Commons) 1905, vol. 62 at 465 (Accounts and Papers) Cmd. 2315 Cd., 1904, Report of Committee of Inquiry into the Case of Mr. Adolf Beck, at vii . . . .” (Internal quotation marks omitted.) Kampshoff v. Smith, 698 F.2d 581, 585 (2d Cir. 1983). More recently, the United States Supreme Court agreed: “The vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification. . . . The identification of strangers is proverbially untrustworthy. The hazards of such testimony are established by a formidable number of instances in the records of English and American trials.” United States v. Wade, 388 U.S. 218, 228, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967).
Moreover, there is undisputed evidence of the propensity of jurors to be swayed by eyewitness testimony, despite its unreliability. In one of the seminal works in the field of eyewitness testimony, Professor Elizabeth F. Loftus documents impressive statistical evidence that, even “when no other evidence is available, the testimony of . . . eyewitnesses can be overwhelmingly influential. A jury seems to find it proof enough when *593a single person implicates another with a remark such as ‘I am certain that’s the man.’ ” E. Loftus, Eyewitness Testimony (1979) p. 9.
Furthermore, this case involves cross-racial identification, the perils of which three members of today’s majority expressly recognized in State v. Reddick, 224 Conn. 445, 470 n.25, 619 A.2d 453 (1993) (“[t]he dissent does not own the comer on the recognition of ‘the problems of cross-racial identification’ ”).3 As I discussed in my dissent in Reddick, “[i]t is well documented that cross-racial identification is less reliable than identification of one person by another of the same race. [Considerable evidence indicates that people are poorer at identifying members of another race than of their own. . . . Moreover, counterintuitively, the ability to perceive the physical characteristics of a person from another racial group apparently does not improve significantly upon increased contact with other members of that race. . . . Note, ‘Did Your Eyes Deceive You? Expert Psychological Testimony on the Unreliability of Eyewitness Identification,’ 29 Stan. L. Rev. 969, 982 (1977). [In her treatise, Professor Loftus] wrote that [i]t seems to be a fact — it has been observed so many times — that people are better at recognizing faces of persons of their own race than a different race. E. Lof-tus, [supra] pp. 136-37. She concluded, on the basis of an examination of studies made on the subject, that [p]eople have greater difficulty in recognizing faces of another race than faces of their own race. Id., p. 139.
“Concern for the problems of cross-racial identification is well documented in our case law as well as by other social scientists. We are painfully aware of *594miscarriages of justice caused by wrongful identification. Those experienced in criminal trial work or familiar with the administration of justice understand that one of the great problems of proof is posed by eyewitness identification, especially in cross-racial identification. . . . Brown v. Davis, 752 F.2d 1142, 1146 (6th Cir. 1985) .... [An] important factor to be considered in assessing the reliability of an identification is whether the witness and the person identified are of the same or different races. In general, there is a much greater possibility of error where the races are different than where they are the same. ... P. Wall, Eye-Witness Identification in Criminal Cases (2d Ed.) p. 122; A. Yar-mey, The Psychology of Eyewitness Testimony (1979) pp. 130-31.” (Citations omitted; internal quotation marks omitted.) State v. Reddick, supra, 224 Conn. 471-72 n.l (Berdon, J., dissenting).
In my view, our judicial system cannot fulfill its constitutional duty to ensure the fairness of criminal convictions that are predicated on the eyewitness identification of a lone, uncorroborated witness unless expert testimony such as that offered by Leippe is available to aid the jurors. Having reviewed the record in this case, I conclude that the exclusion of Leippe’s testimony in this case violated the defendant’s right to a fair trial.
II
It is now generally accepted in other jurisdictions that “the admission of expert testimony regarding eyewitness identifications is proper”; United States v. Moore, 786 F.2d 1308, 1312 (5th Cir. 1986); and that its exclusion can constitute reversible error. Several federal and state courts have held explicitly that trial courts have abused their discretion by excluding this type of testimony in cases in which identification was the central issue. See, e.g., United States v. Harris, 995 *595F.2d 532, 535 (4th Cir. 1993) (reliability of identification causes greater concern where case “involves one identification, by one witness, under stress”); United States v. Sebetich, 776 F.2d 412, 418-19 (3d Cir. 1985), cert. denied, 484 U.S. 1017, 108 S. Ct. 725, 98 L. Ed. 2d 673 (1988) (error to exclude expert testimony where identification, made nineteen months after robbery, was made under stressful circumstances, and was derived from one person’s testimony); United States v. Norwood, 939 F. Sup. 1132, 1137-41 (D.N.J. 1996) (proffered testimony of expert not necessarily within common knowledge of average juror and hence should have been admitted); United States v. Jordan, 924 F. Sup. 443, 449 (W.D.N.Y. 1996) (error to exclude expert testimony where state’s case “depends largely on the testimony of a single eyewitness who was the victim”); United States v. Brown, 45 M.J. 514, 516 (C.A.A.F. 1996) (error to exclude expert testimony where, upon viewing photographic array, sole eyewitness to truck theft selected defendant’s photograph and asked police, “Was that him?”); State v. Chapple, 135 Ariz. 281, 293-97, 660 P.2d 1208 (1983) (error to exclude expert testimony on witness reliability when facts are close); Nations v. State, 944 S.W.2d 795, 798 (Tex. App. 1997) (error to exclude expert testimony where “most highly contested issue was the victim’s identification of her assailant”; victim viewed photographic array and stated one photograph “could be” assailant and, when pressed by police for assurance, “responded that she thought it was him”).
Our own opinion in a recent case led me to believe that this court was headed in the same direction. In the closely analogous case of State v. Barletta, 238 Conn. 313, 321, 680 A.2d 1284 (1996), we unanimously held that the trial court abused its discretion by excluding expert testimony on the unreliability of an eyewitness’ perception and memory.4 More specifically, the expert *596in Barletta would have testified that “it was more likely than not that [the eyewitness’] visual observations at [the time of the crime] were unreliable due to her cocaine ingestion.” Id. If exclusion of that testimony constitutes an abuse of discretion, then no principled ground exists for maintaining that the trial court in the present case did not, similarly, abuse its discretion.5 In light of our unanimous opinion in Barletta, I find the majority opinion in the present appeal inexplicable. Even if Barletta were not part of our jurisprudence, however, I would still disagree with the result that the majority has reached. In State v. Kemp, 199 Conn. 473, 476, 507 A.2d 1387 (1986), this court established the following standard for assessing whether expert testimony is admissible: Generally, expert testimony is *597admissible if (1) “the witness has a special skill or knowledge directly applicable to a matter in issue,” (2) “the testimony would be helpful to the court or jury in considering the issues,” and (3) “[the witness’] skill or knowledge is not common to the average person . . . .” Upon careful consideration of these three factors, I conclude that the trial court abused its discretion by excluding the testimony of Leippe. I address each aspect of the Kemp analysis seriatim.
A
I begin by considering Leippe’s “special knowledge” in the field of eyewitness behavior, a recognized discipline of scientific inquiry. Id. Both his undergraduate and doctoral degrees are in psychology. For more than ten years, he has been a professor of psychology at Adelphi University. He received a research grant in conjunction with the Law and Sciences Division of the National Science Foundation entitled, “The Accuracy, Credibility and Impact of Children’s and Adult’s Eyewitness Testimony.” In the area of eyewitness identification, Leippe has been involved in numerous studies over the past twenty years, including studies of the memory of eyewitnesses to crime-like events. He has coauthored a book in the field, and has published twenty-five articles in treatises and scientific journals, such as Law and Human Behavior, The Journal of Applied Psychology and The Journal of Applied Social Psychology. Many of these publications concerned the subject of eyewitness recall. He is also a member of the Psy-chonomics Society, “which is comprised mainly of people who do research on human memory and human thought.”
B
In my view, Leippe established the relevance of his proposed testimony with unambiguous clarity; he certainly established it with sufficient clarity to satisfy the *598second prong of the Kemp analysis. During the voir dire examination conducted by defense counsel, Leippe explained research findings relevant to the fohowing issues: (1) the crippling effect of stress on memory; (2) the very weak correlation between accuracy of eyewitness testimony and expressions of confidence; (3) the unreliability of aural identifications; and (4) the phenomenon of “unconscious transference.” Without hearing Leippe’s testimony on these four topics, the jury could not intelligently evaluate Hale’s ability to perceive and recall. For this reason, the jury was not competent to resolve the central issue — the reliability of Hale’s identification — without the benefit of Leippe’s expert testimony. I consider each portion of Leippe’s testimony in turn.
First, Hale’s identification of the defendant as the shooter was based entirely on an observation that she made while under extreme stress. Indeed, shortly after the murders occurred, police officers observed that Hale was frightened to the point of hysteria.6 The state even emphasized this fact in its closing argument to the jury: “Here we have a [witness] who’s several months pregnant, scared out of her mind, in a situation she’d never been in before .... That was a traumatic, shocking event for her . . . .” Leippe testified that “the more elevated the stress . . . the more negative the impact of the stress on the reliability of recall.” It is apparent that this testimony was relevant to the jury’s evaluation of the probable accuracy of the memories formed by a terrified pregnant woman being held at gunpoint.
*599Second, Hale testified that, at the time she made the identification, she was 100 percent certain that the defendant was the murderer. The state emphasized this in its final closing argument: “I suggest to you [that Hale did not identify the shooter] until she was sure— I don’t pretend to know all the dynamics of the human mind to tell when she was sure and what made her sure, all I know is she did not make an identification until she felt sure. The point is she felt 100 percent sure at that time. [Y]ou could contrast that with the way she felt before when she wouldn’t make an identification and find that she was convinced and she was accurate when she said [she] was 100 percent sure.” (Emphasis added.) In short, the state urged the jury to believe Hale because she was confident. Leippe would have cautioned the jury, however, that “dozens of studies demonstrate] that there is, at best, a very weak, and often insignificant, relationship between accuracy of, for example, an identification from a lineup, and the confidence with which that is expressed.”7 Leippe’s testimony on this particular issue was clearly relevant to the jury’s determination of whether Hale’s identification was accurate.
Third, Hale’s identification of the defendant was based, in part, on the sound of his voice. She heard the applicant utter a few words when he picked up the employment form, and (assuming arguendo that it was *600the same person) a few more words when he stuck a gun to her head. While speaking with the police immediately after the event, Hale characterized this voice as “plain.” After watching a videotaped lineup in which each suspect recited a few words, Hale purported to recognize the defendant’s voice as the voice of the shooter. Leippe testified at voir dire that “it would be very difficult— very unlikely that a person — that a voice heard one time . . . only one day of a person’s life, that had no distinguishing features, that was of another race, and two years had gone by since that voice had been heard, that the likelihood that that person would recognize that voice would be very low. And in fact, the evidence suggests that voice identification is highly subject to false positive identifications. That is, mistaken identifications.” This testimony on aural identification was obviously relevant to the jury’s assessment of whether Hale accurately recalled the voice of the man who placed a gun to her head.
Fourth, Leippe’s testimony regarding the phenomenon of unconscious transference, which occurs when a witness confuses a person seen in one context with one seen in another, would have been relevant to the jury’s determination of Hale’s reliability. Leippe explained that a witness exposed to an innocent person in a photographic array is — when confronted with a subsequent array — likely to identify that person as the one who committed the crime. Leippe testified that an eyewitness who identifies a defendant after viewing merely two or three prior photographic arrays will probably be affected by this phenomenon. In this case, Hale viewed at least eight photographic arrays that included pictures of the defendant before purporting to identify the defendant as the shooter. It is beyond dispute that Leippe’s testimony would have been relevant to the issue of whether Hale mistook the defendant — whom *601she may have recognized from exposure to his photograph in eight prior photographic arrays — for the shooter.
C
Turning to the final prong of the Kemp analysis, Leippe testified that the subject of his testimony is not part of the common knowledge of laypersons, most of whom dramatically “overestimate the accuracy of memory in general.” See United States v. Moore, supra, 786 F.2d 1312 (“conclusions of the psychological studies .. . serve to explode common myths about an individual’s capacity for perception” [internal quotation marks omitted]); United States v. Norwood, supra, 939 F. Sup. 1137-41 (proffered testimony of expert not necessarily within common knowledge of average juror and should have been admitted); People v. Campbell, 847 P.2d 228, 233 (Colo. App. 1992) (eyewitness identification testimony is “contrary to common wisdom”); Engberg v. Meyer, 820 P.2d 70, 79 (Wyo. 1991) (“recent research seems to demonstrate that the process [of eyewitness identification] is more complex than earlier assumed, and some of the research findings are contrary to intuitive perceptions”); J. Brigham & M. Wolfskeil, “Opinions of Attorneys and Law Enforcement Personnel on the Accuracy of Eyewitness Identifications,” 7 Law & Hum. Behav. 337, 347 (1983) (judges and jurors “strongly believe in eyewitness evidence which may often be erroneous”).
Although most people believe that they possess a strong intuitive grasp of the workings of memory, numerous studies demonstrate that “ ‘common sense’ notions may not accurately reflect the true state of affairs.”8 J. Brigham & R. Bothwell, “The Ability of Prospective Jurors to Estimate the Accuracy of Eyewitness *602Identifications,” 7 Law & Hum. Behav. 19,19-20 (1983). Leippe testified that “[many] well established findings in the field of memory . . . could not be predicted . . . by asking people . . . how they think memory works.” For example, Leippe testified that a substantial percentage of laypersons mistakenly believes that, “the more stressed out a person is, the better the person’s memory’s [going to] be for [a] particular episode . . . .” According to Leippe, although jurors may “be tempted [to accept] . . . the statement, ‘I was so scared . . . there was just no way I could forget that face,’ ” the fact of the matter is that “the greater the anxiety a person experiences, the less likely the person is . . . to have a good memory [of the event experienced while] under that stress. That’s a well documented finding.” A jury’s commonsense beliefs about stress and memory are thus inconsistent with the empirical evidence that stress reduces one’s ability to perceive and recall identifying characteristics.
In addition, Leippe indicated that scientific studies demonstrate that laypersons believe that the accuracy of a witness’ recollection can be gauged by evaluating the witness’ apparent confidence. As previously discussed, Leippe would have testified that this belief is simply not true.9
*603Leippe further testified that many courts wrongly believe that the unreliability of eyewitness testimony is “common knowledge.” He explained to the trial court during voir dire that, “after I tell you that a relationship does not exist, or that a relationship exists, that the natural human tendency . . . is to say, ‘Well, of course. I knew that. That makes sense. That’s obvious.’ ” Nevertheless, without the testimony of Leippe, the jury may have fundamentally misunderstood what it may subsequently have characterized as “obvious” after listening to the testimony of an expert witness. If there were any doubts about this conclusion, the majority has dispelled them by concluding that Leippe’s testimony “should come as no surprise to the average juror.”10
Notwithstanding the very real danger that the jury may have dramatically overestimated the reliability of Hale’s testimony, the majority today mimics the erroneous conclusion this court reached twelve years ago in Kemp. The court in Kemp posited the following hypothesis: “If [1] the reliability of the identifications can be adequately questioned by [cross-examination and closing argument] and [2] the jury is capable of understanding the reasons why they may be unreliable, [then] the introduction of expert testimony would be a superfluous attempt to put the gloss of expertise, like a bit of frosting, upon inferences which lay persons were equally capable of drawing from the evidence.” (Emphasis added; internal quotation marks omitted.) State v. Kemp, supra, 199 Conn. 478-79. In Kemp, this court concluded that counsel could adequately impeach eyewitness identifications without recourse to expert testimony. Id., 479. The empirical evidence discussed thus *604far conclusively demonstrates that this court’s faith in the acumen of laypersons was misplaced. It is apparent that defense counsel cannot pierce the armor of uncorroborated eyewitness accounts unless armed with expert testimony. In the language of the Kemp hypothesis, it is clear (1) that the reliability of eyewitness identifications cannot be adequately questioned through either cross-examination11 or closing argument12 and (2) that jurors are not capable of understanding apparent indicia of reliability without the aid of expert testimony. For these reasons, the defendant in a case such as the present one is entitled to inform the jurors about the issue that lies at the core of their deliberations; such testimony cannot be characterized as “superfluous frosting.” This court’s decision in State v. Barletta, supra, 238 Conn. 313, reflects this fact;13 a contrary result in the present case cannot be supported either by our own precedent14 or by well established findings in the field of memory.15
*605III
Although the majority asserts that it “continuejs] to follow the rationale of [Kemp],” it misapplies Kemp by focusing on whether Leippe established the principles upon which he relied with a “ ‘reasonable degree of scientific certainty.’ ” The “scientific certainty” test comes from the case of Frye v. United States, 293 F. 1013 (D.C. Cir. 1923). The United States Supreme Court :rejected the Frye test in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 588-89, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993); we did likewise in our decision in State v. Porter, 241 Conn. 57, 75-76, 698 A.2d 739 (1997), cert. denied, 523 U.S. 1058, 118 S. Ct. 1384, 140 L. Ed. 2d 645 (1998). Accordingly, the “scientific certainty” test — upon which the majority relies in order to affirm the judgment of the trial court — is no longer good law. Even before Daubert/Porter, moreover, this court had rejected the use of the Frye test in the context of appeals such as the one presently before us.16 See *606State v. Kemp, supra, 199 Conn. 476 (setting forth the three-prong test discussed previously). Whether or not Daubert/Porter trumps Kemp, one matter is certain: the Frye “scientific certainty” test has nothing to do with the present appeal. For these reasons, the majority is absolutely wrong to rely on this test in order to avoid reversing the defendant’s conviction and ordering a new trial.
Although the majority invokes the discredited Frye test, it expressly declines to analyze Leippe’s testimony through the lens supplied by Daubert/Porter, on the ground that neither party has asked us to do so. This simply is not true. The defendant both raised the Daub-ert/Porter issue before the trial court and devotes an *607entire section of his brief to Daubert/Porter, arguments which he further emphasizes in his reply brief. See, e.g., United States v. Rincon, 28 F.3d 921, 923-26 (9th Cir. 1994) (analyzing expert testimony on reliability of eyewitness identification under Daubert).
Turning to the merits of the issue that the majority has elected to ignore, Leippe’s proposed testimony is clearly admissible under Daubert/Porter. In Porter, I recently explained the Daubert test as follows: “[T]he scientific evidence need not be generally accepted by the scientific community; rather, the scientific evidence must be relevant and reliable. ... In Daubert, the court delineated certain considerations, although not an exhaustive checklist, in order to guide trial courts through their evaluation of the admissibility of scientific evidence: (1) whether the theory or technique can be, or has been, tested; (2) whether the theory or technique has been subjected to peer review and publication, although publication (which is one element of peer review) is not an indispensable condition; (3) the known or potential rate of error and the existence and maintenance of standards controlling the technique’s operation; and (4) whether the theory or technique possesses general acceptance within the scientific community. . . . The Supreme Court also stated that [t]he inquiry ... is, we emphasize, a flexible one. Its overarching subject is the scientific validity — and thus the eviden-tiary relevance and reliability — of the principles that underlie a proposed submission. The focus, of course, must be solely on principles and methodology, not on the conclusions that they generate. ... [A]s the majority states it, a trial judge should deem scientific evidence inadmissible only when the methodology underlying such evidence is sufficiently invalid to render the evidence incapable of helping the fact finder determine a fact in dispute.” (Citations omitted; internal quotation marks omitted.) State v. Porter, supra, 241 *608Corm. 145-46 (Berdon, J., concurring and dissenting). In the present case, the trial court held that “memory, as a branch of the social sciences, is a generally accepted field of expertise,” and that Leippe was a qualified expert in the field. After questioning Leippe at length about this scientific discipline, the trial court commented that, if the defendant needed to satisfy Daubert/Porter, he had done so.
IV
Given the prodigious and undisputed empirical research marshaled by Leippe, it is apparent that his testimony should not have been deemed inadmissible under any test.17 I conclude, therefore, that Leippe’s testimony — which would have dispelled commonly held myths by informing the jury that many apparent indicia of reliability in this case are, in fact, highly suspect — would have helped the jury evaluate the probable accuracy of Hale’s identification of the defendant as the shooter. See United States v. Smith, 736 F.2d 1103, 1106 (6th Cir.), cert. denied, 469 U.S. 868, 105 S. Ct. 213, 83 L. Ed. 2d 143 (1984) (determining that jury could be helped by testimony on: [1] unconscious transference between photospread three weeks after robbery *609and lineup four months after robbery; [2] relationship between stress and identification; and [3] cross-racial transference).
The trial court’s exclusion of the evidence — based, in part, on its assumption that scientific theory regarding the working of human memory could be developed on cross-examination and effectively argued without evidentiary foundation — is simply incorrect.18 As previously discussed, this assumption is contradicted by undisputed empirical evidence. In short, Leippe’s testimony would have helped the jury resolve the single most important issue in this case: whether Hale’s identification was reliable. Accordingly, the trial court’s order excluding Leippe’s testimony was neither supported by the record nor correct as a matter of law.
By upholding the trial court’s refusal to admit the testimony of Leippe as a proper exercise of its discretion, the majority insulates a flagrant and dangerous error from review. “Discretion means a legal discretion, to be exercised in conformity with the spirit of the law and in a manner to subserve and not to impede or defeat the ends of substantial justice. In a plain case this discretion has no office to perform, and its exercise is limited to doubtful cases, where an impartial mind hesitates.” (Internal quotation marks omitted.) State v. Onofrio, 179 Conn. 23, 29, 425 A.2d 560 (1979). Discretion does not mean, as the majority seems to believe, an untrammeled right to do as the trial court wishes. See United States v. Brien, 59 F.3d 274, 277 (1st Cir.), cert. denied, 516 U.S. 953, 116 S. Ct. 401, 133 L. Ed. 2d 320 (1995) (trial court’s “discretion is not carte blanche”); State v. Chapple, supra, 135 Ariz. 296 (discretion “does not mean that the court is free to reach any conclusion it wishes”). If discretion were as unbridled as the majority seems to believe, a criminal defendant *610would be playing Russian roulette whenever his case was assigned to a judge: justice would depend upon which chambers he happened to draw. This is not justice. Rather, a trial court’s discretion must be based upon a reasoned consideration of the unique circumstances of each particular case. In this capital case, the defendant’s fate turned entirely upon the jury’s determination of the credibility of an identification made by one terrified eyewitness who viewed the assailant for thirty seconds. The empirical evidence marshaled by Leippe casts substantial doubt upon the reliability of this identification. Under these circumstances, the trial court’s exclusion of Leippe’s testimony constituted an abuse of discretion.
V
Finally, the abuse of discretion in the present appeal was far from harmless. By excluding Leippe’s testimony, the trial court deprived the “jurors of the benefit of scientific research on eyewitness testimony [and] force [d] them to search for the truth without full knowledge and opportunity to evaluate the strength of the evidence. In short, this deprivation prevented the] jurors from having the best possible degree of understanding of the subject toward which the law of evidence strives.” (Internal quotation marks omitted.) Note, supra, 29 Stan. L. Rev. 1017-18. It is likely that depriving the jury of the excluded testimony prejudiced the defendant.
Other courts have warned that jurors may be swayed by unreliable eyewitness testimony. “Juries . . . [are] naturally desirous to punish a vicious crime, [and] may well be unschooled in the effects that the subtle compound of suggestion, anxiety, and forgetfulness in the face of the need to recall often has on witnesses. Accordingly, doubts over the strength of the evidence of a defendant’s guilt may be resolved on the basis of *611the eyewitness’ seeming certainty when he points to the defendant and exclaims with conviction that veils all doubt, ‘[T]hat’s the man!’ United States v. Wade, [supra, 388 U.S. 235-36] . . . .” (Citation omitted.) Kampshoff v. Smith, supra, 698 F.2d 585. This danger was particularly grave in this case, because there was not a shred of forensic or other evidence corroborating the claim of the lone eyewitness.
This was a close case: the palm print on the application did not match that of the defendant, and the state’s own ballistic expert could not state that the three fatal bullets had been fired from the defendant’s gun.19 The jury had serious doubts about the defendant’s guilt. Indeed, the jury deliberated for two weeks before reaching a verdict.20 Moreover, during that time period, the jurors asked the court to replay both Hale’s testimony and the videotape of the lineup from which she selected the defendant. They also asked the court if Hale had ever given a written statement to the police, and, if so, whether they could read the statement. It is apparent that the jurors regarded these pieces of evidence as important to their determination of the defendant’s guilt. See State v. Hoeplinger, 206 Conn. 278, 298, 537 A.2d 1010 (1988). It is equally apparent that the jurors may very well have reached a different verdict if the trial court had not prevented them from scrutinizing the purported identification through the lens supplied by the testimony of Leippe.21
*612I would reverse the judgment of the trial court, vacate the defendant’s effective sentence of 140 years, and remand the case for a new trial at which the jurors would not be permitted to evaluate the credibility of the lone eyewitness — a terrified pregnant woman being held at gunpoint — without the expert testimony of Leippe.
Accordingly, I dissent.
Indeed, it is misleading for the majority to quote, from the Appellate Court opinion, the testimony of James McDonald. A forensic firearms examiner for the Hartford police department, McDonald testified at trial that the fired bullets of the .22 caliber revolver found in the defendant’s parents’ apartment matched those removed from the bodies of the two victims. The majority does not disclose that other testimony from the state’s own witnesses substantially undermined the credibility of McDonald, who was discharged by the Hartford police department after disagreeing with the Federal Bureau of Investigation over the results of forensic tests in another criminal case. David Gibbs, a ballistics expert who is a firearms and toolmark examiner with the Connecticut state police forensic laboratory, previously had testified as an expert fifty to fifty-five times (usually as a state’s witness) and had conducted thousands of firearms examinations each year. After firing the defendant’s revolver and analyzing the bullet fragments from the victims’ bodies, Gibbs concluded that he could not determine that the three fatal bullets had been fired from the defendant’s gun. In addition, Detective Richard Omicienski of the Hartford police department testified that Hale examined the defendant’s weapon three days after the event and stated “that it did not look like the weapon used during the commission of the [murder].”
Even the state did not credit McDonald’s testimony. An affidavit executed by McDonald would have connected the defendant’s pistol with the fatal bullets. Accordingly, such an affidavit would have supplied evidence of probable cause sufficient to justify a lineup and the taking of a voice sample. Significantly, the state’s attorney opted not to include as part of the warrant application an affidavit from McDonald. From this fact I infer that even the state’s attorney was not persuaded by McDonald’s conclusion.
The majority is thus incorrect when it asserts that “the state produced other eyewitnesses who saw the defendant leaving the murder scene . . . at the time 1he crimes had been committed . . . .”
I have no desire to “own the corner”; I would be content if the members of the majority recognized both the problems of cross-racial identification and the need for expert testimony concerning the subject.
After engaging in the fact-specific analysis of harm to the defendant as a result of the exclusion of the evidence, we declined to order a new trial *596in Barletta, on the ground that the suspect identification of the defendant “was substantially corroborated by other eyewitnesses, each of whom was subject to a sequestration order.” State v. Barletta, supra, 238 Conn. 323. The testimony of the lone eyewitness in the present appeal, in contrast, was the only piece of evidence that suggested that the defendant had any involvement in the crime. I consider the issue of the harm of the trial court’s exclusion of the evidence more fully in part V of this dissent.
The majority provides no support whatsoever for its conclusory assertion that “our decision in [Barletta] . . . [is] a far cry from Leippe’s proposed testimony . . . .” This inability to invoke either law or logic stems from the fact that the result the majority reaches today cannot possibly be reconciled with Barletta. The bare fact that the unreliability of the identification in Barletta stemmed from cocaine ingestion is a distinction without a difference. As discussed more fully in this dissent, laypersons embrace wholly erroneous myths regarding the reliability of memory, whether or not such reliability is impaired by the effects of controlled substances. If anything, there was a greater need for expert testimony in this case than there was in Barletta. Even without expert testimony, a jury is likely to view with skepticism an identification made by a witness under the influence of cocaine, because it is common knowledge that use of narcotics impairs perception. This simply is not true of the identification in this case, the unreliability of which is highly counterintuitive.
I can only conclude that the dispositive distinction between this case and Barletta inheres in the desired result: in Barletta, this court avoided reversing the conviction because the trial court’s abuse of discretion was harmless; State v. Barletta, supra, 238 Conn. 324-25; whereas, in the present case, the abuse of discretion would require reversal of the conviction.
Incredibly, the majority argues that Hale’s identification was not affected by stress because she observed the defendant before the onset of stress. This argument is flawed because it ignores the critical fact that Hale never would have identified the defendant as the shooter unless she had determined — while under extreme stress — that the shooter resembled the person who filled out an employment application.
“Numerous investigations of this phenomenon have been conducted: the majority of recent studies have found no statistically significant correlation between confidence and accuracy, and in a number of instances the correlation is negative — i.e., the more certain the witness, the more likely he is [to be] mistaken. . . . Indeed, the closer a study comes to reproducing the circumstances of an actual criminal investigation, the lower is that correlation . . . leading [the researchers who conducted one study] to conclude that the eyewitness accuracy-confidence relationship is weak under good laboratory conditions and functionally useless in forensically representative settings.” (Citations omitted; internal quotation marks omitted.) People v. McDonald, 37 Cal. 3d 351, 369, 690 P.2d 709, 208 Cal. Rptr. 236 (1984).
“In one study, 500 subjects answered questionnaires designed to test their knowledge about eyewitness identification. . . . While ninety percent knew that the wording of a question affects the response, only sixty-six *602percent knew that stress decreases one’s ability to process information. . . . [Only] thirty-three percent knew that violence impairs one’s ability to perceive and recall; and a scant eighteen percent knew that a violent event is harder to recall than a nonviolent one.” R. Handberg, “Expert Testimony on Eyewitness Identification: A New Pair of Glasses for the Jury,” 32 Am. Crim. L. Rev. 1013, 1035 (1995).
The Third Circuit Court of Appeals has observed that the “correlation between confidence and accuracy in eyewitness identifications is far lower than people probably would expect.” United States v. Stevens, 935 F.2d 1380, 1401 (3d Cir. 1991). The California Supreme Court further points out that “[a] number of researchers using a variety of methods have found that people intuitively believe that eyewitness confidence is a valid predictor of eyewitness accuracy.” (Internal quotation marks omitted.) People v. McDonald, 37 Cal. 3d 351, 369, 690 P.2d 709, 208 Cal. Rptr. 236 (1984).
Even if my colleagues in the majority were empirically correct, I am unable to comprehend the force of their argument. If indeed “the average juror” were savvy about the problems of eyewitness identification, what harm could possibly flow from the court’s decision to admit expert testimony on the subject? It would do no more than confirm the justifiable skepticism of some jurors while casting doubt upon the dangerous misconceptions embraced by others.
I agree with the Third Circuit Court of Appeals that, “[t]o 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 [that] witness’ recollection of [the] event.” United States v. Downing, 753 F.2d 1224, 1230 n.6 (3d Cir. 1985); see State v. Taylor, 50 Wash. App. 481,489,749 P.2d 181 (1988) (“expert testimony on the unreliability of eyewitness identification can provide significant assistance to the jury beyond that obtained through cross-examination and common sense”).
Because many of the empirical findings marshaled by Leippe are highly counterintuitive, I fail to understand how defense counsel could effectively impeach an eyewitness’ identification during closing argument without recourse to expert testimony.
See the beginning of part II of this dissent.
To the extent that the majority relies on Kemp rather than Barletta— which was decided a decade after Kemp — this reliance cannot withstand scrutiny.
Furthermore, the trial court’s jury instructions did not accomplish the defendant’s goal of focusing the jury’s attention on factors that may have impaired Hale’s ability to identify the defendant. The trial court instructed the jury as follows with respect to eyewitness identification: “On the issue of eyewitness identification you should take into account the following. You should consider whether or not witnesses who identified the defendant had the opportunity to observe the features which they described in court, the *605time they had to view the perpetrator of the crime, the conditions under which they made those observations, [and] the distance between them and the perpetrator of the crime. You should also consider the description which they gave to the police and the length of time between the giving of that description and the incident in question. You should take into account the circumstances under which they first viewed and identified the defendant, the suggestibility or lack of suggestibility of the procedures used in that viewing, and all other facts which you find relating to the reliability or lack of reliability of the identification of the defendant. You may take into account any occasions in which a witness failed to make an identification that was consistent with his or her identification at trial.”
Although this instruction discusses several general factors that may influence perception and memory, it does not mention the particular factors that Leippe addressed in the defendant’s offer of proof, namely: (1) that stress cripples reliable memory; (2) that an eyewitness’ assertion is not likely to be accurate merely because it is confident; (3) that aural identifications are unreliable; and (d) that unconscious transference may create mistaken identifications. The majority’s reliance upon the court’s instruction is, accordingly, severely misplaced.
Even if the case law of either this court or the United States Supreme Court supported the use of the Frye test in the present appeal, such an analysis would be wholly inappropriate. Nearly ten years before Daubert, *606the California Supreme Court cogently explained why it had concluded that expert testimony on the reliability of eyewitnesses should not be subject to the “scientific certainty” test set forth in Frye: “It is important to distinguish . . . between expert testimony and scientific evidence. When a witness gives his personal opinion on the stand — even if he qualifies as an expert — the jurors may temper their acceptance of his testimony with a healthy skepticism bom of their knowledge that all human beings axe fallible. But the opposite may be true when the evidence is produced by a machine: like many laypersons, jurors tend to ascribe an inordinately high degree of certainty to proof derived from an apparently ‘scientific’ mechanism, instrument, or procedure. Yet the aura of infallibility that often surrounds such evidence may well conceal the fact that it remains experimental and tentative.. . . For this reason, courts have invoked the [Frye] rule primarily in cases involving novel devices or processes ....
“Here, by contrast, no such methods are in issue. We have never applied the [Frye] rule to expert medical testimony, even when the witness is a psychiatrist and the subject matter is as esoteric as the reconstitution of a past state of mind or the prediction of future dangerousness, or even the diagnosis of an unusual form of mental illness not listed in the diagnostic manual of the American Psychiatric Association .... We see no reason to require a greater foundation when the witness is a qualified psychologist who will simply explain to the jury how certain aspects of everyday experience shown by the record can affect human perception and memory, and through them, the accuracy of eyewitness identification testimony. Indeed, it would be ironic to exclude such testimony on [Frye] grounds on the theory that jurors tend to be unduly impressed by it, when jurors are far more likely to be unduly impressed by the eyewitness testimony itself.” (Citations omitted; emphasis added; internal quotation marks omitted.) People v. McDonald, 37 Cal. 3d 351, 372-73, 690 P.2d 709, 208 Cal Rptr. 236 (1984).
In Porter, this court explained that scientific evidence is admissible if it “tend[s] to support a relevant fact even to a slight degree, so long as it is not prejudicial or merely cumulative.” (Emphasis in original; internal quotation marks omitted.) State v. Porter, supra, 241 Conn. 87. As previously discussed, Leippe’s testimony would have helped the jury evaluate the central issue in this case.
Indeed, we always have employed a liberal standard when evaluating the admissibility of expert testimony on general findings in a field of expertise when such testimony is relevant to a determination of the credibility of witnesses. See State v. Ali, 233 Conn. 403, 428-35, 660 A.2d 337 (1995) (upholding admission of expert testimony “regarding the general characteristics of women who delay reporting sexual assault” where expert’s “general description of characteristics that are common among” rape victims “narrowly focused on a subject not familiar to the average person”); State v. Borrelli, 227 Conn. 153, 168, 629 A.2d 1105 (1993) (expert “presented a general description of battered woman’s syndrome”).
See footnotes 16 and 17 of this dissent.
See fool note 1 of this dissent.
Jury deliberations began on December 17, 1992, and continued for ten days, until January 4, 1993.
Moreover, the harm of not allowing the jury the benefit of Leippe’s testimony was further compounded by the trial court’s failure to allow into evidence a report — made by Detective Ronald Faggaini of the Hartford police department, who died prior to trial — that “Hale stated [that] since she didn’t see the culprit’s face at the time of the robbery/murder, she wouldn’t be able to make a[n] identification.” At trial, Hale denied the substance of this statement. In my view, the trial court should have allowed the defendant to introduce Faggaini’s statement under the residual exception to the hearsay rule; I would also reverse on this ground.
*612“Few rights are more fundamental than that of an accused to present witnesses in his own defense. ... In these circumstances, where constitutional rights directly affecting the ascertainment of guilt are implicated, the hearsay rule may not be applied mechanistically to defeat the ends of justice.” Chambers v. Mississippi, 410 U.S. 284, 302, 93 S. Ct. 1038, 35 L. Ed. 2d 297 (1973).