People v. LeGrand

OPINION OF THE COURT

Bernard J. Fried, J.

Seeking the admission of expert testimony concerning eyewitness identification, the defendant has requested a pretrial ruling allowing such evidence at his retrial, following the declaration of a mistrial, on a charge of murder in the second degree. Specifically, what is sought to be introduced is the testimony of a psychologist, professor Roy S. Malpass, Department of Psychology, University of Texas at El Paso, with regard to (1) the confidence-accuracy correlation; (2) postevent information and confidence malleability; and (3) weapon focus.¹ For the reasons set forth below, this application is denied.

At the outset, in order to avoid any misapprehensions, I believe it is necessary to recognize that these subjects or concepts are commonly defined as follows:

(1) Confidence-accuracy correlation refers to “the relation between the accuracy of an eyewitness’s identification and the confidence that [the] eyewitness expresses in the identification” (Gary L. Wells, Mark Small, Steven Penrod, Roy S. Malpass, Solomon M. Fulero and C.A.E. Brimacombe, Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads, 22 L & Hum Behav [No. 6] 1, 14 [1998]).

(2) Postevent information refers to the proposition that “[e]ye-witness testimony about an event often reflects not only what [the eyewitness] actually saw but information they obtained later on” (Saul M. Kassin, V. Anne Tubb, Harmon M. Hosch and Amina Memon, On the “General Acceptance” of Eyewitness Testimony Research: A New Survey of the Experts, 56 Am Psychologist [No. 5] 405 [in press at *13] [2001]; see also Saul M. Kassin, Phoebe C. Ellsworth and Vicki L. Smith, The “General Acceptance” of Psychological Research on Eyewitness Testimony: A Survey of Experts, 44 Am Psychologist [No. 8] 1089, 1091 [1989]).

(3) Confidence malleability refers to the proposition that “[a]n eyewitness’s confidence can be influenced by factors that are unrelated to identification accuracy” (id.).

(4) Weapon focus refers to “the visual attention that eyewitnesses give to a perpetrator’s weapon during the course of a crime” (Nancy Mehrkens Steblay, A Meta-Analytic Review of the Weapon Focus Effect, 16 L & Hum Behav [No. 16] 413, 414 [1992]).

Factual and Procedural Background

At approximately 7:00 a.m. on June 15, 1991, a taxicab driver, Joaquin Liar ano, was stabbed to death. Witnesses had heard the cab crash into a parked automobile, observed the victim and his assailant fighting both inside and outside the cab, saw the victim repeatedly being stabbed, and watched as the assailant then took property from the cab and fled the area. Almost immediately they called 911, police officers responded to the scene, and the victim was taken to a hospital, where later that morning he died.

That same day one of the witnesses was taken to the 26th Precinct and shown photographs; however, he did not make an identification. The next day, that witness, and another witness, were shown additional photographs, but neither witness made an identification. Five days later, on June 20th, at the Police Artist Unit, four witnesses — Pazmino, Foote, Gonzalez and Gomez — were present and a composite sketch was prepared. This sketch was incorporated into a wanted poster and circulated; however, for the next seven years no suspects were identified.

On April 5, 1998, the defendant was arrested in connection with burglary charges within the confines of the 26th Precinct, leading to a detective reopening the 1991 homicide investigation. A lineup was held on the burglary arrest; however, counsel objected to having a witness from a 1991 incident view the lineup and therefore there was no such viewing. Thereafter, on May 9, 1998, Pazmino viewed a photo array, containing the defendant’s photograph, and identified the defendant as the person who had committed the 1991 homicide. Then, on August 17, 1998, the detective traveled to Florida, where he displayed the photo array to Foote and Gonzalez. Each selected the defendant’s photograph — to Foote it was a “close match”; to Gonzalez, it was similar.

A lineup was held on February 3, 1999, where the defendant was identified by Pazmino, which led to the defendant’s indictment in April 1999. Just prior to the April 2001 trial, there was a defense motion seeking permission to call an eyewitness expert. The trial justice reserved decision until after Pazmino’s testimony, and then after argument denied the application. At that trial, Pazmino identified the defendant, as did Foote, noting that he was a “striking match.” Gonzalez pointed out the defendant, stating that “it looks like he gained more weight.” Other witnesses, including Gomez, did not identify the defendant. Because of the jury’s inability to reach a unanimous verdict, a mistrial was declared, and the case was ultimately sent to me for retrial.

Before retrial could commence, the New York Court of Appeals issued its unanimous decision in People v Lee (96 NY2d 157, 160 [2001]), holding that while expert testimony on eyewitness identification is “not inadmissible per se, the decision whether to admit it rests in the sound discretion of the trial court.” This led to a renewed application to admit the expert testimony. Following papers submitted by the defense and the prosecution, I ordered an evidentiary hearing to determine (1) whether the proposed witness could be qualified as an expert, (2) the admissibility of the proposed testimony, and (3) the relevancy of the proffered evidence.

This hearing was held at which the defense called professor Roy S. Malpass, who testified over two days.² To rebut Dr. Malpass, the People called professor Ebbe B. Ebbesen, Department of Psychology, University of California, San Diego, whose testimony was concluded in one day. Following this testimony, I received extensive and comprehensive papers in support of, and in opposition to, the admission of professor Malpass. In addition, the People have provided counsel and this court with copies of numerous scholarly articles, dealing with the subject of eyewitness identification.

Frye or Daubert?

There currently are two distinct methodologies for the judicial determination of the admissibility of expert testimony. Traditionally, it had been evaluated under the landmark 1923 decision in Frye v United States (293 F 1013 [DC Cir 1923]), from which evolved the so-called Frye test. However, in 1993, the United States Supreme Court, in Daubert v Merrell Dow Pharms., Inc. (509 US 579 [1993]), a case which arose under the Federal Rules of Evidence, abandoned the Frye test in favor of a new approach.

While the New York Court of Appeals has recently adhered to the Frye test (see People v Wesley, 83 NY2d 417 [1994]), there is an energized debate, among various courts and commentators, as to whether we will one day adopt the Daubert approach to the admissibility of expert evidence. (E.g., Hofmann v Toys “R” Us, 272 AD2d 296 [2d Dept 2000]; Wahl v American Honda Motor Corp., 181 Misc 2d 396 [Sup Ct, Suffolk County 1999]; Clemente v Blumenberg, 183 Misc 2d 923 [Sup Ct, Richmond County 1999]; Castrichini v Rivera, 175 Misc 2d 530 [Sup Ct, Monroe County 1997]; see also Justice Walter J. Relihan, Jr., Outside Counsel, Considering the Frye Rule in New York Justice, NYLJ, Sept. 15, 2000, at 1, col 1.) This is a debate, however, which I need not enter, since both the People and the defense have agreed that analysis of the admissibility of the proffered evidence be under the Frye test.

The Necessity For an Evidentiary Hearing

Recognizing that expert identification testimony “may involve novel scientific theories and techniques,” Lee stated that “a trial court may need to determine whether the proffered expert testimony is generally accepted by the relevant scientific community.” (Lee, supra at 162.) Moreover, in her opinion for a unanimous Court, Judge Graffeo observed that “[although the issue of admissibility of expert testimony regarding the reliability of eyewitness identification has been addressed in other jurisdictions (see generally, 1 McCormick, Evidence § 206, at 776-779 [5th ed 1999]), it remains an unresolved issue in this State (see People v Alexander, 94 NY2d 382; People v Mooney, 76 NY2d 827).” (Lee, supra at 162; but see People v Mooney, supra at 829-830 [Kaye, Ch. J., dissenting] [“the emerging trend today is to find expert psychological testimony on eyewitness identification sufficiently reliable to be admitted, and the vast majority of academic commentators have urged its acceptance”].)

I am aware of two trial court opinions, decided after Lee, which have split on the issue of requisite procedures and necessary information to decide the admissibility of expert identification testimony. In People v Radcliffe (191 Misc 2d 545, 548 [Sup Ct, Bronx County 2002]), Judge William C. Donnino concluded that “until the law of New York settles on the parameters of expert identification testimony, the application for the admission of such testimony should be made pretrial because the trial court may need to conduct a hearing” to determine the issues. He further explained that the proponent of such evidence should:

“(1) to the extent known set forth the pertinent alleged facts of the identification and any corroborative evidence; (2) set forth the name and qualifications of the witness and the ‘proffered’ testimony; (3) correlate the proffered testimony with the facts of the case to demonstrate the relevance of the expert testimony; (4) explain whether the testimony involves ‘novel scientific theories and techniques,’ and if it does, include an offer of proof as to its general acceptance by the relevant scientific community; and (5) explain why the testimony is warranted if an existing standard jury instruction (CJI[NY]2d Identification — One Witness [2001]) would appear to cover the area of the proffered expert testimony.”

Certainly, Judge Donnino has outlined a procedure, which undoubtedly anticipates “a hearing to amplify or verify any of the [proponent’s] assertions.” This is not only consistent with Judge Graffeo’s decision in Lee, that this issue remains “unresolved in” New York, but it is cognizant of Lee’s caution that “in recognition that expert testimony of this nature may involve novel scientific theories and techniques, a trial court may need to determine whether the proffered expert testimony

is generally accepted by the relevant scientific community.” (Lee, supra at 162.)

To the contrary, in People v Smith (191 Misc 2d 765, 766 [Sup Ct, NY County 2002]), Judge James A. Yates denied a prosecution request for a Frye hearing to determine if the proffered expert eyewitness testimony, concerning “stress, event violence, cross-racial impact, and exposure time that affect the reliability and accuracy of eyewitness identification,” was generally accepted in the relevant scientific community. In order to reach this conclusion, Smith relied solely on articles which had urged such testimony, and did not address the numerous studies, whose methodologies, measurements and conclusions had become hotly contested amongst the experts in the relevant scientific community. Nevertheless, Smith concluded that there was nothing novel about the proposed testimony, implicitly accepting its general acceptability. This seems to have ignored Lee’s observation, that “[although the issue * * * has been addressed in other jurisdictions * * * it remains an unresolved issue in this State.” (Lee, supra at 162.) While the New York decisions relied on by Smith included cases which had admitted such testimony as beyond the ken of the jury or for other reasons, perhaps implicitly finding general acceptability, there were no explicit findings concerning general acceptability.³ Moreover, the previous New York cases which had declined to admit such testimony generally did so on the ground that such evidence was within the “ken,” thereby failing to make any findings on general acceptability.⁴

To me, then, a hearing was required, and it was held as described above. Indeed, in a pre-Lee decision, People v Jeter (80 NY2d 818, 820-821 [1992]), concerning the proffer of expert testimony on spectrographic analysis, the Court of Appeals held that the trial court could not have properly “determined that voice spectrography is generally accepted as reliable” due to the “marked conflict in the judicial and legal authorities as to the reliability of the procedure,” and “the legal scholarship on the admissibility of voice spectrography [which was] likewise conflicting.” Accordingly, the Court (at 821) concluded that “the trial court lacked a proper basis to admit the voice spectrographic evidence without a preliminary inquiry into reliability.”

The Frye Hearing

Once a pretrial hearing has been determined to be necessary, the trial judge must apply the fourfold test for the admissibility of scientific expert evidence, as set forth in Frye v United States (supra). The first requirement concerning the admissibility of scientific expert evidence under the Frye test provides that the witness be competent in the field of expertise that he purports to address at trial. Thus, “the expert should be possessed of the requisite skill, training, education, knowledge or experience from which it can be assumed that the information imparted or the opinion rendered is reliable.” (Matott v Ward, 48 NY2d 455, 455-456 [1979]; see Prince, Richardson on Evidence § 7-304, at 463 [Farrell 11th ed].)

The second requirement is that such “expert testimony be based on a scientific principle or procedure which has been ‘sufficiently established to have gained general acceptance in the particular field in which it belongs’.” (People v Wemick, 89 NY2d 111, 115 [1996].) Thus, under Frye, it is not the responsibility of the trial judge to determine whether a novel expert testimony is reliable. (Wesley, supra at 439 [Kaye, Ch. J., concurring].) Rather, the trial judge is only required to determine whether there is a general consensus among scientists in the relevant community, i.e., general acceptability, that the proposed expert testimony is reliable. (Id.) Therefore, “the particular procedure need not be ‘unanimously indorsed’ by the scientific community but must be considered ‘generally accepted as reliable’.” (Wesley, supra at 423, quoting People v Middleton, 54 NY2d 42, 49 [1981]; see also Selig v Pfizer, Inc., 185 Misc 2d 600, 605 [Sup Ct, NY County 2000].) Consequently, the Frye test has been criticized for “counting scientists’ votes, rather than on verifying the soundness of a scientific conclusion.” (Wesley, supra at 439 [Kaye, Ch. J., concurring], quoting Jones v United States, 548 A2d 35, 42 [DC Ct App].)

In addressing the general acceptability of the proffered expert testimony, it is not necessary that I determine whether the eyewitness identification discipline, in its entirety, is generally accepted. Rather, the issue that I must decide is whether the proposed testimony of the defense’s witness, Dr. Malpass, which, specifically addresses the confidence-accuracy correlation, postevent information and confidence malleability, and weapon focus, is generally accepted as reliable within the relevant psychological community. “It is incumbent upon the proponent of expert testimony to lay a proper foundation establishing that the processes and methods employed by the expert in formulating his or her opinions adhere to accepted standards of reliability within the field” (People v Fortin, 184 Misc 2d 10, 13 [Nassau County Ct 2000]), “through judicial opinions, scientific or legal writings, or expert opinion other than that of the proffered expert” (Cameron v Knapp, 137 Misc 2d 373, 375 [Sup Ct, NY County 1987]).

The third requirement under the Frye test involves an inquiry into whether the proffered expert testimony is beyond the ken of the jury. (See People v Cronin, 60 NY2d 430, 433 [1983].) “The guiding principle is that expert opinion is proper when it would help to clarify an issue calling for professional or technical knowledge, possessed by the expert and beyond the ken of the typical juror.” (De Long v County of Erie, 60 NY2d 296, 307 [1983]; see also People v Taylor, 75 NY2d 277, 288 [1990].) As a result, expert testimony may be admissible “where the conclusions to be drawn from the facts ‘depend upon professional or scientific knowledge or skill not within the range of ordinary training or intelligence’ ” of a jury. (Cronin, supra at 432, quoting Dougherty v Milliken, 163 NY 527, 533 [1900]; see also People v Miller, 91 NY2d 372, 379 [1998].) Thus, the trial judge must determine “ ‘when jurors are able to draw conclusions from the evidence based on their day-to-day experience, their common observation and their knowledge, and when they would be benefited by the specialized knowledge of an expert witness’.” (Lee, supra at 162, quoting Cronin, supra at 433.) However, “even if the proposed testimony was not beyond the jury’s ken, [the Court of Appeals has repeatedly] upheld the admission of expert testimony for the purpose of clarifying an area of which the jurors have a general awareness.” (Mooney, supra at 832 [Kaye, Ch. J., dissenting].) Therefore, trial judges are to be wary “not to exclude such testimony merely because, to some degree, it invades the jury’s province.” (Lee, supra at 162.)

The fourth and final requirement concerning the admissibility of scientific expert evidence under the Frye test is that the expert’s opinion be relevant to the issue and facts of the individual case. (People v Allweiss, 48 NY2d 40, 50 [1979].) “Evidence is relevant if it has any tendency in reason to prove the existence of any material fact, i.e., it makes determination of the action more probable or less probable than it would be without the evidence.” (People v Scarola, 71 NY2d 769, 777 [1988].) Moreover, even if the court determines that the evidence is relevant to the issue and facts of the individual case, it may still exclude the evidence “if its probative value is substantially outweighed by the danger that it will unfairly prejudice the other side or mislead the jury.” (Id.)

At the hearing, Dr. Malpass established his qualifications as an expert in the field of eyewitness identification: He currently serves as a professor of psychology and director of the criminal justice program, University of Texas at El Paso. In addition, he is a member of both, the editorial board of Law and Human Behavior, and of the technical working group for eyewitness evidence at the United States Department of Justice. Moreover, he has published numerous writings and has given many presentations on this subject. Similarly, the witness called by the People, Dr. Ebbesen, established his qualifications as an expert in this discipline: He is presently a professor of psychology at the University of California, San Diego; has served on many occasions as a consultant or witness to various state prosecutors throughout the country; has published numerous writings; and has been an invited participant and speaker at various conferences and presentations on the subject of eyewitness identification.⁵ Thus, it is evident that each witness — Dr. Malpass and Dr. Ebbesen — is an expert in the area of eyewitness identification.

There is also no question as to whether the proffered expert testimony is relevant to the issues and facts of this case. For instance, Dr. Malpass has proposed to testify on the reliability of eyewitness identification, which, unlike Lee, is a central part of the People’s case-in-chief, as there is no independent evidence to corroborate the eyewitness identification testimony. More significantly, the eyewitness identification testimony consists solely of four eyewitnesses, whose identification of the defendant occurred seven years after the alleged criminal activity. Thus, the proffered expert testimony constitutes a good fit with the issues and facts of this case, as it pertains to the central question before the jury, i.e., the eyewitness identifications.

Moreover, there is no need for me to determine whether the evidence is beyond the ken of the jury, since this question apparently has been put to rest by the statement in Lee, that although “jurors may be familiar from their own experience with factors relevant to the reliability of eyewitness observations and identification, it cannot be said that psychological studies regarding the accuracy of an identification are within the ken of the typical juror.” {Lee, supra at 162; see, e.g., Brian Cutler, Steven D. Penrod and Hedy R. Dexter, The Eyewitness, the Expert Psychologist, and the Jury, 13 L & Hum Behav [No. 3] 311 [1989].) Thus, having affirmatively dealt with three parts of the Frye test, I will now turn to the last outstanding issue under Frye, which is whether the proffered expert testimony has gained general acceptance within the relevant psychological community.

I. Confidence-Accuracy Correlation

First turning to the proffered expert testimony on confidence-accuracy, Dr. Malpass testified that 25 years of research has led psychologists to conclude that “there is a statistically significan [t], but very small correlation” between the confidence eyewitnesses express in the accuracy of their identifications, and the actual accuracy of those identifications. Dr. Malpass further testified that, according to a survey conducted by Saul M. Kassin, V. Anne Tubb, Harmon M. Hosch and Amina Memon, entitled On the “General Acceptance” of Eyewitness Testimony Research: A New Survey of the Experts (Am Psychologist, supra [2001]), this relationship between confidence and accuracy is generally accepted within the relevant psychological community. (See also S. Kassin et al., supra [1989].) This is based on the finding that 86.7% of the respondents found the proposition that “[a]n eyewitness’s confidence is not a good predictor of his or her identification accuracy” to be reliable. (S. Kassin et al., supra at *17 [2001].)

However, the proposition that the confidence-accuracy correlation is generally accepted was disputed by Dr. Ebbesen, who testified that “there’s an extremely strong relationship between confidence and accuracy.” Moreover, Dr. Ebbesen and other psychologists, including Dr. Malpass, have discussed the relationship between the confidence-accuracy correlation and other variables, such as the variance among witnesses and conditions in the eyewitness identification.⁶ But, downplaying the claim that the relationship is dependent upon the effects of other variables (see D. Lindsay et al., supra [1998]; J. Read et al., supra [1998]), professor Michael Leippe has written that even under the “most pristine conditions,” the confidence-accuracy correlation “only tends to be about .40.” (Michael R. Leippe, The Case For Expert Testimony About Eyewitness Memory, 1 Psychol, Pub Pol’y, & L [No. 4] 909, 927 [1995] [emphasis in original].) However, the value of professor Leippe’s writings on the confidence-accuracy correlation is diminished by the Connecticut Supreme Court’s decision in State v McClendon (248 Conn 572, 586, 730 A2d 1107, 1115 [1999]), to exclude his testimony on the subject because, in part, he was unable to state his opinion with a “reasonable degree of scientific certainty.” Moreover, professor Leippe testified “that ‘we don’t always know what factors are influencing’ an eyewitness,” and that “a controversy existed in the area of the statistical probability of false identification, the one kind of information inaccessible to the average juror.” (248 Conn at 587, 730 A2d at 1115.) Thus, it is evident that there is no consensus on the confidence-accuracy correlation. To further illustrate this point, “Deffenbacher reviewed 25 studies that collectively involved 43 assessments of the confidence/accuracy relationship. In 22 of the assessments a positive correlation between confidence and accuracy was found, whereas in the remaining 21 assessments either no correlation or a slight negative correlation was obtained.” (J. McKenna et al., supra at 288 [1992]; see also J. Read et al., supra at 107 [1998].) Accordingly, “it is misleading to claim that, in general, confidence eyewitness identifications are little more likely to be accurate than nonconfident ones.” (D. Lindsay et al., supra at 218 [1998].)

The search for a general consensus on confidence-accuracy is further undermined by the lack of an accepted standard of measurement among psychologists. Dr. Ebbesen testified to the existence of three methods to calculate the confidence-accuracy correlation, each revealing a very different result. (See E. Ebbesen, supra [2000].) Thus, the strength and weakness of the relationship between confidence and accuracy is highly dependent on the method of measurement chosen by the individual psychologist. Moreover, Dr. Ebbesen pointed out that the Kassin survey, on which the defense relies, represented only a “very small sample of the thousands of researchers who do research on human memory,” thereby excluding a “highly published minority” of researchers in the eyewitness identification discipline from participating in the survey (e.g., Cynthia Reinick, Vladimir Konecni, John Wixted, Rogers Elliott, Michael McCloskey and John Yuille). Certainly, “[i]f the field is too narrowly defined, the judgment of the scientific community will devolve into the opinion of a few experts. [However] [t]he field must still include scientists who would be expected to be familiar with the particular use of the evidence at issue * * * whether through actual or theoretical research.” (Wesley, supra at 438 [Kaye, Ch. J., concurring].) Notably, there is an outstanding question as to who actually makes up the relevant psychological community, even though both Dr. Malpass and Dr. Ebbesen agreed that the relevant community consisted of psychologists, who had conducted research in the broad area of memory, beyond Dr. Malpass’s testimony of a core group of 36 experts in eyewitness memory. In addition, the authors of the Kassin survey failed to adequately describe the makeup of its respondents, and their particular expertise within the eyewitness identification discipline.

More significantly, the authors of the survey acknowledged that the survey was mailed to a total of 197 prospective respondents, of which only 64 responded to the questionnaire, yielding “a 34.4% response rate.” (S. Kassin et al., supra at *3 [2001].) And, although 86.7% of the respondents found the proposition that “[a]n eyewitness’s confidence is not a good predictor of his or her identification accuracy,” this finding is not determinative in view of the fact that only 24 respondents found the proposition to be “very reliable,” while a mere 16 respondents found the proposition to be “generally reliable,” leaving the implication that the 21 remaining respondents did not find the proposition to be “reliable.” {Id. at *13., 16-17.) Moreover, questions as to the legitimacy of the survey’s findings are raised by the fact that only 73.3% of the survey’s respondents stated that they would testify on the proposition of confidence-accuracy, despite their high response rate concerning the reliability of the proposition and their prior history of testimony on the proposition. {Id. at *17.) Thus, the Kassin survey does not warrant a conclusion that there is “general acceptance” within the psychological community concerning the confidence-accuracy correlation. To decide otherwise would, in essence, have the effect of accepting a survey, which claims that there is general acceptability in the area of confidence-accuracy, when the survey failed to both obtain a response rate of greater than 35% and include researchers, who were reputed for disagreeing with the survey’s proposition on the confidence-accuracy correlation. It is also apparent that the survey was designed expressly for litigation, inasmuch as the authors have written: “[b]y providing empirical evidence of the consensus within the community of experts, this survey has proved useful to judges ruling on the admissibility of expert witnesses.” {Id. at *1.)

Moreover, Dr. Malpass acknowledged that he had previously testified, in the Supreme Court, New York County, on January 12, 1993, that there was no relationship between confidence and accuracy, although he now contends that there is, in fact, a relationship as a result of further research on the subject. Thus, it would be inappropriate to admit testimony on confidence-accuracy when the research is inconclusive, and it cannot be stated with a “reasonable degree of scientific certainty,” particularly when the relevant psychological community has demanded further research on the subject. {See H. Egeth, supra at 161-162 [1995]; J. McKenna et al., supra at 291-292 [1992].) It is also not the duty of the jury to take on the role of the scientist, and determine the reliability of the conclusions on confidence-accuracy, when the relevant psychologists themselves have been unable to do so.

For the foregoing reasons, I conclude that the confidence-accuracy correlation has not yet achieved general acceptability within the relevant scientific community.

II. Postevent Information and Confidence Malleability

Turning now to the proffered expert testimony on postevent information and confidence malleability: The defense seeks to introduce evidence “regarding studies investigating the impact of various types of suggestion, such as that involved in lineup and photo array procedures and postidentification feedback suggesting the identification was correct, on the witness’s recollection of a perpetrator’s appearance.” The defense further seeks to introduce evidence “that witnesses exposed to this type of post-event information tend to incorporate it into their memory of the actual event, that witnesses’ confidence levels in an identification are higher when they chose a person pointed to by suggestion in an identification procedure, and that witnesses’ confidence levels are enhanced by their receipt of post-event information which suggests to them that their initial identification was correct.” This includes the testimony of Dr. Malpass, who testified concerning his own experiments relating to this case, which he described as a “Mock Witness Paradigm” experiment. The design of this experiment is set forth in the footnote.⁷

The experiment itself represented a photographic reenactment of the police lineup that was used to identify the defendant in this case. Dr. Malpass concluded from the experimental data that the “lineup [was] biased against [the defendant] Mr. LeGrand, but only for persons who [had] seen the drawing of the offender constructed with the assistance of certain of the eyewitnesses.” According to Dr. Malpass, the lineup fillers were inadequate:

“For the condition where the participants were shown the artist sketch, it is apparent that lineup fillers numbers 1, 2, 4 and 5 were not appropriate fillers because they were chosen very infrequently * * * For the condition in which participants were shown the verbal descriptions of the offender, two of the lineup ‘fillers’ were inadequate (numbers 1 and 4) * * * For the condition in which participants were asked only if anyone stands out in the lineup, three of the lineup ‘fillers’ were inadequate (numbers 1 & 4 were chosen at about ²/s of expectation, while number 6 [the defendant] was not chosen at all).”

Passing whether such an experiment, and its hypothesis, would ever be admissible at a criminal trial, the experiment appears to suffer from serious deficiencies. Turning first to experiment No. 1, Dr. Malpass testified that the results from the experiment revealed that “the line-up was not biased toward the defendant; it was biased away from the defendant. In fact, not one subject picked the defendant’s picture as standing out.” Next, turning to experiment No. 2, it is evident on the face of the experiment that it contained various errors. For instance, the subjects were instructed under the heading, “Verbal Instructions,” that “Four witnesses observed a man stab and kill a taxi driver with a large knife in an argument that began in the Taxi, and continued outside,” even though the experimental data contained five eyewitness descriptions of the defendant. Also, the description by “Witness 1” — Pazmino, who had identified the defendant from both the photo array and at trial — was inaccurate since it stated that he had described the defendant as having “short kinky white hair,” when actually Pazmino never stated that the defendant had “white” hair. A further review of experiment No. 2 revealed that it failed to include the verbal description given by witness Gonzalez, even though he had identified the defendant from both the photo array and at trial. Instead, the experimental data included the descriptions of “Witness 4” and “Witness 5,” despite the fact that these two witnesses were never present at either the photo array or trial.

Moreover, unlike the police lineup in this case, experiment No. 2 employed a “forced choice” option, in that the subjects of the experiment were told, “One of these individuals is the person the police believe is the killer. Which of these is the person you think is the man the police have arrested?” The subjects were precluded from being able to not select an individual, as would an actual eyewitness. Rather, they were forced to identify a suspect, even though they may have believed that the suspect was not present in the photo lineup. In this regard, it is noteworthy that, as one study concluded:

“Failure to warn the eyewitness that the culprit might not be in the lineup resulted in 78% of the eyewitnesses attempting an identification from the culprit-absent lineup. This false identification rate dropped to 33% when the eyewitnesses were explicitly warned that the culprit might not be in the lineup. Importantly, warning the eyewitnesses that the culprit might not be in the lineup still resulted in 87% of the eyewitnesses making accurate identifications when the culprit was in the lineup, indicating that this instruction does not merely reduce eyewitnesses’ willingness to identify someone. Results of this type reveal that eyewitnesses will simply select the person in the lineup whom they perceive is relatively more similar to the culprit than are the other lineup members if they approach the lineup with the presumption that the culprit is among the set.” (G. Wells et al., supra at 11 [1998]; see also M. Leippe, supra at 916 [1995].)

Although this research has caused Dr. Malpass to conclude that in conducting such experiments, the “[e\yewitnesses should he told explicitly that the person in question might not he in the lineup or photospread and therefore should not feel that they must make an identification,” there was nevertheless a “forced choice” in the “Mock Witness Paradigm.” (G. Wells et al., supra at 23 [1998] [emphasis in original].)

Dr. Malpass, who acknowledged that experiments should not utilize a “forced choice” option, also employed a “forced choice” in experiment No. 3, where the subjects were instructed under the “Artists Sketch Instructions” that “The police believe one of these individuals is the person who committed the offense the witnesses saw. Which one of the persons in the lineup is the police suspect?” Consequently, the subjects were forced to identify a suspect, even though they may have believed that the suspect was not present in the photo lineup. Moreover, in addition to the utilization of a “forced choice,” there are real and unaccounted for distinctions between the “Mock Witness Paradigm” experiment and real life criminal events. For example, as Dr. Malpass himself recognized, the subjects in his experiment possessed a greater universe of information, in that they had the benefit of possessing five separate eyewitness descriptions of the defendant. This is contrary to real life lineups, where the eyewitness only has information pertaining to his or her own individual observation of the defendant.

Turning specifically to experiment No. 3 (“Artists Sketch Instructions”), when asked whether the inability to replicate the seven-year lag period, between the witness’ visualization of the police officer’s sketch of the suspect and the witness’ visualization of the photo array, undermined the reliability of the experiment, Dr. Malpass replied:

“it’s difficult to answer. The witnesses who participated and constructed the artist sketch went through * * * a process of welding the sketch, and that’s very different from looking at a sketch. So we do have some immeasurable things here, I’ll grant. I can pose, but I can’t answer the question of how crisp the image is in the minds of the original witnesses.”

Most astonishingly, Dr. Malpass testified that to compare the “Mock Witness Paradigm” experiment to a real life criminal event was like comparing “apples and oranges,” since the purpose of the experiment, unlike that of a real life criminal event, was narrowly designed “to diagnose the degrees to which the suspect stands out or doesn’t stand out in the photo spread.” This distinction underscores the lack of relevance of the research findings, to the tasks undertaken by a real world witness. Indeed, critics of the “Mock Witness Paradigm,” according to Dr. Ebbesen, have stated that it is not understood “how to take this mock witness paradigm and apply it to the real world because we really don’t know what an appropriate measure of lineup bias is.” More remarkably, Dr. Malpass has written that the technology for the evaluation of the fairness of a lineup is “relatively young, and current research is changing our preferences for techniques and the way they are applied.” (R. Malpass and R. Lindsay, Measuring Lineup Fairness, 13 Applied Cognitive Psychol SI, S6 [1999].) Thus, it is evident that Dr. Malpass’s proposed testimony concerning the “Mock Witness Paradigm” experiment has not achieved general acceptability.

Despite these flaws associated with the “Mock Witness Paradigm” experiment, it is argued that the proposed testimony on postevent information and confidence malleability is generally accepted as reliable within the relevant psychological community. The defense relies on the Kassin survey, according to which 93.7% of the respondents found the proposition on postevent information, that “[e]yewitness testimony about an event often reflects not only what they actually saw but information they obtained later on,” to be reliable. (S. Kassin et al., supra at *13, 17 [2001].) Moreover, according to the survey, 95.2% of the respondents found the proposition on confidence malleability, that “[a]n eyewitness’s confidence can be influenced by factors that are unrelated to identification accuracy” (at *13), to be reliable. However, questions as to the legitimacy of the survey^s findings are raised by the fact that only 82.5% of the survey’s respondents stated that they would testify on the proposition concerning postevent information, while 79.0% stated that they would testify on the proposition concerning confidence malleability, despite their high response rate concerning the reliability of the two propositions, and their prior history of testimony on each proposition. More significantly, this apparent general consensus, as to the reliability of these propositions on postevent information and confidence malleability, is undermined by the survey’s failure to adequately represent the relevant community of researchers in the eyewitness identification discipline, and the low response rate, of only 34.4.% from the people to whom the questionnaire was sent. And, of course, as noted earlier, the survey failed to give an adequate description of who the respondents were, and their particular position in the forensic eyewitness identification debate.

For the foregoing reasons, I conclude that this proposed testimony on postevent information and confidence malleability, which includes the “Mock Witness Paradigm” experiment, has not been generally accepted.

III. Weapon Focus

Finally, turning to weapon focus: Dr. Malpass testified concerning the effect of “weapon focus” on the accuracy of eyewitness identifications. Relying solely on Steblay’s meta-analysis,⁸ he testified that it was generally accepted within the relevant psychological community that “witnesses to a crime in which the perpetrator is displaying a weapon tend to focus their attention on the weapon, and therefore are less accurate in identifying the perpetrator, relative to witnesses in equivalent situations where no weapon is displayed.” (See N. Steblay, supra [1992].)

However, the allegedly general acceptability of this I conclusion, which was drawn from Steblay’s meta-analysis, does not adequately represent the majority of studies on weapon focus. According to the study, only 6 of the 19 tests, which made up the meta-analysis, found a weapon focus effect, even though the mean effect size for the group of tests indicated otherwise. (See N. Steblay, supra [1992].) This clearly raises into question the issue of whether the mean effect size computed from the meta-analysis is truly representative of the group of studies that made up the meta-analysis. Since a majority of the tests that made up the meta-analysis showed that there was no weapon focus effect, which was in direct contrast to the mean effect size computed from the combined tests, it is evident that the variability among the individual tests serves to discredit the conclusion that weapon focus is generally accepted.

Indeed, the present data on weapon focus is mixed.⁹ To illustrate this point, professor Rogers Elliott has explained that some experiments have found that the accuracy of an eyewitness identification can be reduced by 20% when a weapon is present. However, he also explained that there are other experiments that have shown that there is no difference in accuracy, irrespective of the presence or absence of a weapon. Moreover, on the basis of experiments, “it appears that the presence of a weapon did not hinder the ability of eyewitnesses to offer a detailed description. [Rather e]yewitnesses in crimes involving a weapon provided more detailed descriptions than those in weaponless crimes and victims offered more detailed descriptions than witnesses.” (P. Tollestrup et al., supra at 156 [1994].) According to professor Elliott, the effect of a weapon on the accuracy of a witness’ identification varies as a function of the individual; however, “the research does not yet tell us the circumstances under which presence should produce a moderate or small effect or which people are more susceptible to a weapon focus.” (See S. Kassin, P. Ellsworth and V. Smith, Deja vu All Over Again: Elliott’s Critique of Eyewitness Experts, 18 L & Hum Behav [No. 2] 203, 205 [1994] [the authors of the earlier, 1989, Kassin survey, which was subsequently updated in 2001, wrote that “no expert ha(d) ever claimed that the mere presence of a weapon, or any other single factor for that matter, (was) sufficient on its own to discredit an identification”].) Thus, there is no consensus as to how the presence of a weapon may affect a particular witness’ memory.

Moreover, Dr. Malpass stated in his testimony that “there are studies out there that are inconsistent with the findings [he] claim [s] are conclusive,” since the conclusions from Steblay’s meta-analysis, on which his proposed testimony relies, is not “finished knowledge,” but rather, and most tellingly, it is a “work in progress.” He also stated that there is insufficient data on weapon focus, to the point that he himself would not even “speculate” on the subject. However, despite these reservations, the defense claims that the proposed testimony on weapon focus is generally accepted as reliable within the relevant psychological community, based on the Kassin survey, according to which 86.7% of the respondents found the proposition, that “[t]he presence of a weapon impairs an eyewitness’s ability to accurately identify the perpetrator’s face,” to be reliable. (S. Kassin et al., supra at *13, 17 [2001].) But, it is notable that this finding is compromised by the fact that only 15 respondents found the proposition on weapon focus to be “very reliable,” while 27 respondents found it to be “generally reliable,” implying that the remaining 21 respondents did not find the proposition to be “reliable.” (Id. at *16.) More significantly, questions as to the legitimacy of the survey are raised by the fact that only 77.1% of the survey’s respondents stated that they would testify on the proposition concerning weapon focus, despite their high response rate concerning the reliability of the proposition, and their prior history of testimony on the proposition. (Id. at *17.) Furthermore, not only did the survey fail to give an adequate description of the respondents and their eyewitness identification expertise, but it did not adequately represent the researchers in the eyewitness discipline, and it received a low response rate (34.4%) from those who actually received the questionnaire.

For the foregoing reasons, I conclude that this proposed testimony on weapon focus is not generally accepted.

Laboratory vs. Reality

One of the fundamental underpinnings of the disagreement concerns the question of whether it is generally acceptable for experts to extrapolate their research findings from laboratory studies to the testimonies of actual eyewitnesses in real life criminal events.¹⁰ Each witness — Dr. Malpass and Dr. Ebbesen — testified on the issue of external validity and whether experimental research findings in eyewitness identification can be generalized to real court cases. Methodological critics, such as Dr. Ebbesen, claim that experimental results “cannot be generalized to real court cases due to a lack of realism of experiment procedures.” (A. Maass, supra at 290 [1996].) Specifically, Dr. Ebbesen testified that researchers, by not employing a parametric research design,¹¹ had failed to examine the extent to which the relationship between a specific variable and memory was affected by the existence of other variables present in a real life criminal event. In other words, the variables that affect accuracy, according to Dr. Ebbesen, are highly interactive, and therefore, “the size of the effect of one variable on accuracy, depends on the level of other variables.” And, since multiple variables interact to produce differing effects, one cannot state, with any definitiveness, that the effect will apply to all circumstances. This problem is exacerbated by the requirement that experimental research reduce the number of “variables to a manageable size,” thereby excluding many variables from the design. (A. Maass, supra at 290 [1996].) Moreover, it is vulnerable to the possibility that “the examined variables [will] interact with other important factors in an unknown way.” (Id.) Thus, Dr. Ebbesen notes that psychologists are unable to give an exact percentage of the confidence-accuracy correlation, the effects of postevent information and confidence malleability on accuracy, and the weapon focus effect, because such evidence is dependent on so many variables. Accordingly, this defect, which has produced greater control and internal validity, has caused experimental psychology in eyewitness identification to become generally associated with a loss of external validity and generalizability.

Professor Nancy Steblay has acknowledged this problem relating to variables, when she wrote, “[a] desirable goal for future research is to clarify the interactive effect of arousal and attention in the weapon focus phenomenon.” (N. Steblay, supra at 422 [1992].) She also noted that “[i]t may be argued that real-life crime events include so many stimuli that the hypothesized weapon focus effect becomes irrelevant or insignificant in magnitude.” (Id.) Moreover, professor Michael Leippe has acknowledged this problem, when stating that researchers had shown themselves to be unable to effectively test the accuracy and arousal levels experienced by witnesses in real life criminal events (see M. Leippe, supra at 919 [1995]; J. Yuille et al., supra at 157-160 [1998]), since, “[flor ethical reasons, participants in research experiments cannot be subjected to levels of arousal as extreme as those present when a person is in a life-threatening situation.” (Otto H. MacLin, Kimberly M. MacLin and Roy S. Malpass, Race, Arousal, Attention, Exposure, And Delay: An Examination of Factors Moderating Face Recognition, 7 Psychol, Pub Pol’y, & L [No. 1] 134, 149 [2001]; see also J. Yuille et al., supra at 931 [1994].) Thus, the real life eyewitness has been “tainted” by the presence of exceptional psychological states and motivations, while the laboratory witness has remained “untainted” by such factors. (M. Leippe, supra at 919 [1995].)

Psychologists have recognized this lack of understanding of the psychological state of witnesses in real life criminal events, and the influential effect of such factors on memory itself, as such suspect qualities are dependent on the nature of the event, as well as the witness and victim, and their exposure to the event. (See M. Leippe, supra at 918-919 [1995]; J. Yuille et al., supra at 213 [1992]; John Shepherd, Social Factors in Face Recognition, Perceiving and Remembering Faces, at 55-79 [1981].) Consequently, whereas research on real life events suggests that traumatic events are well preserved in memory, laboratory simulation studies on the effects of stress with regard to a witness’ level of accuracy have revealed mixed results.¹² For instance:

“In Deffenbacher’s (1983) review of 21 studies examining the impact of stress on eyewitness memory, he found that the results of 10 of the studies suggested that stress either increases eyewitness accuracy or has no effect, whereas the findings from 11 studies led to the conclusion that stress decreases eyewitness accuracy.

“One possible explanation for these contradictory results relates to differences in the methodology employed in these studies. Of the 10 studies that reported a facilitative or neutral effect of stress on memorial accuracy, the majority involved staged live events similar to that pioneered by von Listz. Of the 11 studies suggesting that arousal or stress decreases memorial accuracy, only 1 study employed a live event.” (J. Yuille et al., supra at 157 [1998].)

Thus, it is evident, from this mixed data, that the effects of stress in any given situation are unclear. Accordingly, experts have cautioned others against the idea of generalizing their research findings to criminal scenarios where certain psychological qualities, such as stress and arousal, are suspected. (See M. Leippe, supra at 919 [1995]; J. Yuille, supra at 573 [1993].)

However, Dr. Malpass and others have noted an exception here for the uninvolved bystander, whom they claim experiences a comparable level of arousal and accuracy to the laboratory witness. (See O. MacLin, supra at 149 [2001]; J. Yuille, supra at 572-573 [1993] [professor Yuille does not support, as the defense alleges, Dr. Malpass’s viewpoint here concerning the uninvolved bystander. Rather, he only points out his disagreement with “Wells and most of the eyewitness identification literature (which) treats the bystander witness as the model for all eyewitnesses”].) This claim follows suit with some psychologists’ findings that the accuracy level of a victim and bystander are similar. (See P. Woolnough et al., supra at 405-407 [2001]; M. Leippe, supra at 919 [1995].) Professor Leippe points to staged crime studies, which showed either similar (Hosch and Cooper [1982]) or higher (Kassin [1984]) rates of false identification between bystander witnesses and mildly victimized witnesses. Moreover, he writes that “field and archival studies of bystander and mildly victimized witnesses to real crimes do not generally contradict the results of laboratory studies.” But, this exception has been refuted by other psychologists, i.e., the 1986 archival research of professor Yuille, and the 1986 field study of professors Yuille and Cut-shall, who have found that “real witnesses are typically not unaffected and generally are not bystanders.” (J. Yuille et al., supra at 204-205 [1992].) Dr. Ebbesen also testified as to the variability between the uninvolved bystander witness in the laboratory and the real life witness, noting that the laboratory witness, who is aware of the experimental character of his or her identification, and who is not personally affected by the crime, behaves differently from the real life witness, whose identification may pose serious consequences to himself or herself, and the offender. (See G. Koehnken, Eyewitness Testimony: False Alarms on Biased Instructions?, 73 J Applied Psychol [No. 3] 363 [19983; A. Maass, supra at 291 [1996]; J. Yuille et al., supra at 1276 [1995]; J. Yuille et al., supra at 208 [1992]; C. Krafka et al., supra at 66-67 [1985].) As a result, “real witnesses appear to often recall more details, often have a high level of accuracy, and often show little loss of either detail or accuracy with the passage of time.” (J. Yuille et al., supra at 208 [1992].) Accordingly, other researchers have found that there is a real discrepancy between the accuracy rates of actual eyewitnesses and those of experimental eyewitnesses, as the former is superior to the latter. (See J. Yuille et al., supra at 291, 299 [1986].) This discrepancy between the real life witness and the laboratory witness has caused these psychologists to conclude that research, particularly on the impact of stress or trauma on memory, must be conducted in an actual criminal context for the findings on eyewitness identification to be generalized to the testimony of actual eyewitnesses of real life criminal events. (J. Yuille et al., supra at 160 [1998].)

Moreover, Dr. Ebbesen has written that “not much is known about how these factors [stress, unconscious transference, confidence, retention interval, exposure duration, lineup fairness, racial similarity, and weapon focus] do interact” to affect eyewitness identification, which has, in turn, led to the question, “if not much is known, one wonders how defense experts can draw the sweeping conclusions that they do?” (E. Ebbesen et al., supra at 17 [1996]; see also E. Ebbesen, supra at 22 [2000].) Even Dr. Malpass testified that the research on the psychological state of witnesses (e.g., stress) is “messy,” and therefore, he is unable to determine its application to a given situation. Thus, not only do the stress or arousal levels in the experimental studies dramatically differ from the real world, but there is apparently a lack of understanding of how that difference in stress or arousal levels will effect a witness’s accuracy.

The claim that psychologists generally accept the external validity of the research findings in eyewitness identification, i.e., that it would allow in-court testimony, is also undermined by the research methodologies. One illustration focuses on the “forced choice” rule, which I previously discussed with respect to the “Mock Witness Paradigm” experiment. Another focuses on experimental psychology’s use of slide sequences, filmed events and staged criminal events to replicate a real life crime scene. Professors Yuille and Cutshall have noted that the usage of these mechanisms “does not qualify as a ‘forensically relevant paradigm’ and may be of limited value for generalizing to witness situations in the real world (cf Clifford, 1978; Yuille and Cutshall, 1984).” (J. Yuille et al., supra at 932 [1994]; see J. Yuille et al., supra at 291 [1986]; V. Konecni et al., supra at 123-124 [1986].) This is because “[laboratory studies typically involve relatively low levels of stress induced by means such as the presentation of an ‘upsetting’ videotape or a series of slides. These contrived situations are necessarily quite different from a situation in which someone actually witnesses a crime.” (J. McKenna et al., supra at 286 [1992].) Moreover, “[t]he typical introductory psychology student who serves as a volunteer subject, looks at a series of photos or a film depicting a simulated crime and then attempts an identification differs from a real witness not only in demographical characteristics but also because he or she is not personally affected by the crime (Malpass & Devine, 1980).” (A. Maass, supra at 290 [1996]; see also J. Yuille et al., supra at 157-158 [1998].) For instance, an armed target presented on slides or video films is not generalizable to a real crime that involves the presence of a weapon since, unlike most real life crime situations, the “subjects are never actually threatened by the presence of the weapon.” (A. Maass et al., supra at 399 [1989].) However, in an article that Dr. Malpass coauthored, he wrote that “studying [the level of arousal] in a laboratory setting can and does provide empirical evidence that can lead insight into how these factors affect a witness to a crime.” (O. MacLin, supra at 149 [2001].) But, in his testimony, he stated that it was a “best guess,” when he wrote, “[although it can be argued that the level of arousal derived from a pointed gun or a threatening knife is not the same as the level of arousal induced in a laboratory setting, it is unlikely that participants are using cognitive systems in the laboratory that are separate from what they use during an actual criminal event.” (Id.) Thus, there is much controversy with regard to the value of the existing literature concerning the impact of trauma on eyewitness memory, as some psychologists contend that “[v]irtually none of the studies that have been conducted to date reflect the nature or degree of impact” of stress or trauma on a witness’s memory that can be found in a real life criminal context. (See J. Yuille et al., supra at 159-160 [1998].) Clearly, it is evident that the external validity of the proposed evidence, particularly with respect to Steblay’s meta-analysis in which 17 of the 19 tests used either a slide or video presentation, is impaired by the use of such nonemotionally charged events that are staged in the controlled setting of a researcher’s laboratory.

Moreover, even advocates of archival research appear to have not yet endorsed the generalization of their research findings to a forensic context. Professor Yuille and other experts have stated that, despite the dramatic difference of results between archival and experimental research, it is too soon to generalize the archival studies to a specific forensic context. Rather, they recognize that more research needs to be done, using a variety of methods (i.e., field studies, archival studies, experimental studies, etc.) to determine the variation of results before the conclusions can be generalized to a forensic context. (See B. Behrman et al., supra at 475-476, 489 [2001]; J. Yuille et al., supra at 935 [1994]; P. Tollestrup et al., supra at 159 [1994]; J. Yuille, supra at 572 [1993]; J. Yuille et al., supra at 207 [1992].) And then, “[o]nly when we have converging results from such diverse methods of inquiry should we feel confident in generalizing to the forensic context.” (J. Yuille et al., supra at 935-936 [1994].)

The research findings in eyewitness identification are also affected by the lack of standardization and uniformity in the measurement of an individual’s accuracy rate. For example, Dr. Ebbesen testified that the method, which measures how many details (or facts) are correctly remembered out of a total number of possible details that could have been remembered, represents a “completely irrelevant” calculation, since it assumes that the researcher possesses knowledge of the total number of items in an event, when “in the real world * * * we have no idea what the actual total [number] of events to be remembered or details to be remembered is * * * because we didn’t have a video camera taking a picture.” Thus, the test is based on the subjective assessment of the relevant details by the researcher, excluding many of the details that the witness may actually remember. There are other concerns regarding the methodology used to examine the relationship between accuracy and confidence: The confidence-accuracy correlation may be misleading, depending upon the method of measurement employed, since there is no uniform numerical scaling system to rate a subject’s confidence level; therefore, the particular scoring method is dependent upon the subjectivity of the individual psychologist, and the data cannot be translated or generalized to a real world context.

There is also an issue as to whether experiments on the confidence-accuracy correlation should include both “choosers” and “nonchoosers” in the equation, since the real world is only concerned with the eyewitness who was able to make an identification. The effect of nonchoosers in the equation has proven itself to be inconclusive, as some psychologists have concluded that the confidence-accuracy correlation is higher for witnesses who are able to make an identification (see G. Wells et al., supra at 17-18 [1998]), while other psychologists claim that there is no difference. (See C. Krafka et al., supra at 65 [1985].) Additionally, experts have questioned the inclusion of “known errors” (an eyewitness identification of a foil or filler, not the suspect, from the lineup), since it fails to mirror police lineups in the real world, and it inflates the statistics, which support the conclusion that a large percentage of people are wrongly convicted based on erroneous eyewitness identifications. Thus, the measurements used in the studies pertaining to confidence-accuracy are not agreed upon as realistic measurements of the means by which the legal system evaluates an eyewitness identification.

There are three distinct methods of measuring the confidence-accuracy correlation, all of which reveal strikingly different results. The first method, which supports Dr. Malpass’s proffered testimony, calculates the average confidence level for the entire group of subjects, finding a correlation coefficient of only 0.2. The second method calculates the average confidence and accuracy level for each subject out of the total number of possible suspects, resulting in a correlation coefficient of 0.6 to 0.7. And finally, the third method calculates the correlation coefficient by using a calibration curve, resulting in a correlation coefficient of 0.85 to 0.9. The latter method demonstrates a very strong relationship between a witness’s confidence and accuracy. Thus, the absence of a standard method of measuring the confidence-accuracy correlation undermines the generalizability of the proffered testimony, as the correlation coefficient is highly dependent on the subjective tastes of the individual expert.

Nevertheless, despite these demonstrated difficulties in transposing the experimental findings to a real world courtroom setting, the defense claims that its contention concerning general acceptability of these principles is supported by a 1999 United States Department of Justice research report: “Eyewitness Evidence: A Guide for Law Enforcement” (Guide) (reprinted at <http://www.courts.state.ny.us/reporter/webdocs/ eyewitness_evidence.pdf >) arguing that

“The fact that the Department of Justice, certainly a respected governmental organization, asked for psychologists in the field of eyewitness identification, including Professor Malpass, to help them promulgate guidelines based on their research findings, is a definitive pronouncement that endorses these findings; thereby indicating a high level of general acceptance of these principles.”

However, in her introduction to the Guide, Attorney General Janet Reno explained that a technical working group of law enforcement, legal practitioners and psychologists had been convened “to explore the development of improved procedures for the collection and preservation of eyewitness evidence within the criminal justice system” (at iii). And the Guide itself made clear that it used “psychological findings, either by including them in the procedures themselves or by using them to point the way to the design and development of further improvements in procedures and practices for possible inclusion in future amendments or revisions to this document” (at 1-2). This is hardly a statement that the United States Department of Justice has determined that there is general acceptability of the principles urged upon this court. And indeed, nowhere in the Guide is there such a statement.

A further problem in admitting testimony, based on a science that lacks general acceptability, is that the testimony may subsequently be found to be erroneous based on further research. This problem was highlighted by Dr. Malpass, who had previously testified before the Supreme Court in New York County, that there was no relationship between confidence and accuracy, which is contrary to his position today.

It is undeniable that there has been a raging controversy among the experts in eyewitness identification. More significantly, Dr. Ebbesen and Dr. Malpass have both agreed that the experts do not exactly understand why various research results have been reached. This is not a debate among experts about a generally accepted principle. Rather, it is a real controversy among the relevant experts concerning whether these principles are generally accepted. Thus, it is the current state of disagreement, and inconclusiveness, and the problems of external validity associated with the research, i.e., transposition to a courtroom setting, that leads me to conclude that this proffered evidence has not been generally accepted within the relevant psychological community. Consequently, to allow this testimony to be received at trial would require jurors to determine whether the expert evidence is generally accepted as reliable, which is the role of the court, not the jury. And absent general acceptability, as Chief Judge Kaye once noted, “[i]t is not for a court to take pioneering risks on promising new scientific techniques, because premature admission both prejudices litigants and short-circuits debate necessary to determination of the accuracy of a technique.” (Wesley, supra at 437 n 4.)

Conclusion

Accordingly, for the foregoing reasons, the defense motion to admit the proffered expert testimony on eyewitness identification is, in all respects, denied.

1 . The defense has stated that it is not seeking to present evidence on cross-racial identification, even though Dr. Malpass addressed the topic at the pretrial hearing.

2 . Actually, several months before professor Malpass testified, on July 31, 2001, the defense called a different witness: professor Justin Anderson of Baruch College. However, on October 15, 2001, the defense withdrew professor Anderson’s testimony, and requested that the hearing be continued in order to call professor Malpass. Professor Malpass then testified in November 2001, followed by professor Ebbesen in December 2001.

3 . E.g., People v Drake, 188 Misc 2d 210, 215 (Sup Ct, NY County 2001); People v Beckford, 141 Misc 2d 71, 75 (Sup Ct, NY County 1988); People v Lewis, 137 Misc 2d 84 (NY County Ct 1987); People v Brooks, 128 Misc 2d 608, 610 (Westchester County Ct 1985); see also Mooney, supra (Kaye, Ch. J., dissenting).

4 . E.g., Lee, supra; Mooney, supra; People v Robinson, 267 AD2d 981 (4th Dept 1999); People v Perry, 251 AD2d 895 (3d Dept 1998); People v Fennell, 237 AD2d 459 (2d Dept 1997); People v Greene, 223 AD2d 474 (1st Dept 1996); People v Doros, 217 AD2d 415 (1st Dept 1995); People v Kelley, 220 AD2d 456, 457 [2d Dept 1995]; People v Maddox, 159 AD2d 954 (4th Dept 1990); People v Wright, 161 AD2d 743 (2d Dept 1990); People v Gibbs, 157 AD2d 799 (2d Dept 1990); People v Knighton, 165 AD2d 904 (3d Dept 1990); People v Barnwell, 155 AD2d 886 (4th Dept 1989); People v Brown, 136 AD2d 1 (2d Dept 1988); People v Foulks, 143 AD2d 1038 (2d Dept 1988); People v Mitchell, 129 AD2d 589 (2d Dept 1987); People v Slack, 131 AD2d 610 (2d Dept 1987); People v Valentine, 53 AD2d 832 (1st Dept 1976); People v Wong, 150 Misc 2d 554, 556 (Sup Ct, Queens County 1991); People v Krel, 154 Misc 2d 820 (Grim Ct, Kings County 1991); People v Schor, 135 Misc 2d 636, 639 (Nassau Dist Ct 1987); People v Brown, 124 Misc 2d 938, 940 (Westchester County Ct 1984).

5 . Professor Malpass’s curriculum vitae runs some 16 pages; professor Ebbesen’s is 19 pages.

6 . E.g., Ebbe B. Ebbesen and John T. Wixted, A signal detection analysis of the relationship between confidence and accuracy in face recognition memory: Implications for eyewitness identification, J Experimental Psychol <http://www-psy.ucsd.edu/~eebbesen/ConfidAccur.html>; Bruce W. Behrman and Sherrie L. Davey, Eyewitness Identification in Actual Criminal Cases: An Archival Analysis, 25 L & Hum Behav (No. 5) 475 (2001); E. Ebbesen, Some thoughts about generalizing the role that confidence plays in the accuracy of eyewitness memory <http://www-psy.ucsd.edu/~eebbesen/ confidence.htm> (2000); D. Stephen Lindsay, J. Don Read and Kusum Sharma, Accuracy And Confidence in Person Identification: The Relationship is Strong when Witnessing Conditions Vary Widely, 9 Psychol Sci (No. 3) 215 (1998); G. Wells et al., supra at 15 (1998); D. Lindsay et al., supra (1998); J. Don Read, D. Stephen Lindsay and Tonia Nicholls, The Relationship Between Accuracy and Confidence in Eyewitness Studies: Is the Conclusion Changing?, Eyewitness Memory: Theoretical and Applied Perspectives, at 107-130 (1998); M.R. Kebbell, G.F. Wagstaff and J.A. Covey, The Influence of Item Difficulty on the Relationship Between Eyewitness Confidence and Accuracy, 87 Brit J of Psychol (No. 4) 653 (1996); Howard Egeth, Expert Psychological Testimony About Eyewitnesses: An Update, Psychology, Science, and Human Affairs: Essays in Honor of William Bevan, 151,162 (1995); Judith McKenna, Molly Treadway and Michael E. McCloskey, Expert Psychological Testimony on Eyewitness Reliability: Selling Psychology Before Its Time, Psychology And Social Policy at 283, 288 (1992).

7 . The “Mock Witness Paradigm” was conducted by his students at the University of Texas at El Paso as follows: they approached subjects of Anglo and Hispanic descent and in the age range of 20 years and older, and asked them, individually, to read the following instructions:

“ ‘Standout’ Instructions [Experiment #1]

‘We will show you a police photo lineup containing six persons.

“Please examine the lineup carefully.

“Does one of the persons in the lineup stand out from the others? If so, please write his number in the lineup in the space below.

“Verbal Description Instructions [Experiment #2]

“Four witnesses observed a man stab and kill a taxi driver with a large knife in an argument that began in the Taxi, and continued outside. They gave these descriptions of the man with the knife:

“Witness 1. Male black or Hispanic, light skin, approximately 6'2", muscular build, clean shaven, short kinky white hair.

“Second description-.

“Mustache, more than 6' tall, kinky hair — short, short — really short. Sideburns — like Elvis. Both sides. And a mustache. Light skinned, African American, 25-28 yrs old.

“Witness 2. Male black, light skin, mustache, clean shaven, short black hair (balding), about 6' tall.

Witness 3. Male Hispanic or black, late 20s, about 5T1" tall, short dark hair, athletic build.

“Second description.

“Male black 5'll"-6'0", medium dark skin, thin but athletic, clean cut, short black hair, early 30s.

“Third description.

“Thin but athletic build, clean cut. In his 30s, 6 ft. tall, short black hair.

“Witness 4. Spanish or light skinned black, 30-35 yrs. old. Tall, husky build, clean shaven, kinky short hair.

‘Witness 5. Male light skin black or Hispanic, 6'0" tall, 190 pounds, in his 30s, short kinky hair, clean shaven, slant eyes, muscular build.

“Please look carefully at the 6-person police photo lineup we will show you.

One of these individuals is the person the police believe is the killer.

“Which of these is the person you think is the man the police have arrested?

“Please write the number of his photograph in the space below.

“Artists Sketch Instructions [Experiment #3]

“Please look carefully at the artists sketch of a person recently arrested by the police in New York City.

We will show you a photo lineup containing six persons. The police believe one of these individuals is the person who committed the offense the witnesses saw.

“Which one of the persons in the lineup is the police suspect? Write the number of his photograph in the space below.”

After they read the instructions, each subject was shown a photo lineup, which consisted of the defendant and five fillers, at which point the subject was instructed to make an identification.

8 . The term “meta-analysis” refers to “[t]he process or technique of synthesizing research results by using various statistical methods to retrieve, select, and combine results from previous separate but related studies.” (See The American Heritage Dictionary of the English Language [4th ed 2000], found at www.dictionary.com.) “With the help of a relatively new procedure, the meta-analysis, we are now in a position to statistically compare the results of a large number of experiments.” (Anne Maass, Logic and Methodology of Experimental Research in Eyewitness Psychology, Psychological Issues in Eyewitness Identification, at 279, 289 [1996].)

9 . E.g., Patricia A. Tollestrup, John W. Turtle and John C. Yuille, Actual Victims and Witnesses to Robbery and Fraud: An Archival Analysis, Adult Eyewitness Testimony, Current Trends and Developments, at 144, 156 (1994); H. Egeth, Emotion and the Eyewitness, The Heart’s Eye: Emotional Influences in Perception & Attention, at 245-267 (1994); Rogers Elliott, Expert Testimony About Eyewitness Identification: A Critique, 17 L & Hum Behav (No. 4) 423, 427-428 (1993); Thomas H. Kramer, Robert Buckhout and Paul Eugenio, Weapon Focus, Arousal, and Eyewitness Memory: Attention Must Be Paid, 14 L & Hum Behav (No. 2) 167 (1990); J. Yuille, Expert Evidence by Psychologists: Sometimes Problematic and Often Premature, 7 Behav Sci & L (No. 2) 181, 187 (1989); Richard O. Lempert, Social Science in Court: On “Eyewitness Experts” and Other Issues, 10 L & Hum Behav (Nos. 1, 2) 167, 169 (1986).

10 . E.g., E. Ebbesen and J. Wixted, supra; E. Ebbesen, supra (2000); B. Behrman et al., supra (2001); Penny S. Woolnough and Malcolm D. MacLeod, Watching the Birdie Watching You: Eyewitness Memory for Actions using CCTV Recordings of Actual Crimes, 15 Applied Cognitive Psychol 395 (2001); E. Ebbesen and Heather D. Flowe, Simultaneous v. Sequential Lineups: What Do We Really Know?, <http://psy.ucsd.edu/~eebbesen/SimSeq.htm> (2001 [accessed Sept. 4, 2003]); E. Ebbesen, supra (2000); E. Ebbesen, Why We Cannot Generalize Conclusions from Source Misattribution (or Misleading Post-Event Information) Studies to Actual Crime Situations, <http:// psy.ucsd.edu/~eebbesen/Misleading.html> (2000); E. Ebbesen and Cynthia B. Rienick, Retention Interval and Eyewitness Memory for Events and Personal Identifying Attributes, 83 J of Applied Psychol [No. 5] 745 (1998); J. Yuille and Judith Daylen, The Impact of Traumatic Events on Eyewitness Memory, Eyewitness Memory: Theoretical and Applied Perspectives, at 155-178 (1998); J. Read et al., supra (1998); E. Ebbesen and Vladimir J. Konecni, Eyewitness Memory Research: Probative v. Prejudicial Value, 5 Intl Dig of Hum Behav, Sci & L 2 (1996); A. Maass, supra (1996); M. Leippe, supra (1995); J. Yuille, J. Daylen, Stephen Porter and David Marxsen, Challenging the Eyewitness Expert, 2 Coping with Psychiatric and Psychological Testimony, at 1268-1298 (1995); H. Egeth, supra (1995); J. Yuille, Graham Davies, Felicity Gibling, D. Marxsen and S. Porter, Eyewitness Memory of Police Trainees for Realistic Role Plays, 79 J of Applied Psychol [No. 6] 931 (1994); P. Tollestrup et al., supra (1994); H. Egeth, supra (1994); J. Yuille, We Must Study Forensic Eyewitnesses to Know About Them, 48 Am Psychol 572 (1993); R. Elliott, supra (1993); D.A. Bekerian, In Search of the Typical Eyewitness, 48 Am Psychol 574 (1993); J. McKenna et al., supra (1992); J. Yuille and P. Tollestrup, A Model of the Diverse Effects of Emotion on Eyewitness Memory, The Handbook Of Emotion and Memory: Research and Theory, at 201-215 (1992); Sven-Ake Christianson, Remembering Emotional Events: Potential Mechanisms, The Handbook Of Emotion and Memory: Research and Theory, at 307-340 (1992); A. Maass and G. Kohnken, Eyewitness Identification: Simulating the “Weapon Effect,” 13 L & Hum Behav [No. 4] 397 (1989); J. Yuille, supra (1989); M. McCloskey, H. Egeth and J. McKenna, The Experimental Psychologist in Court: The Ethics of Expert Testimony, 10 L & Hum Behav [Nos. 1, 2] 1 (1986); V. Konecni and E. Ebbesen, Courtroom Testimony by Psychologists on Eyewitness Identification Issues: Critical Notes and Reflections, 10 L & Hum Behav [Nos. 1, 2] 117 (1986); J. Yuille and J. Cutshall, A Case Study of Eyewitness Memory of a Crime, 71 J Applied Psychol [No. 2] 291 (1986); R. Lempert, supra (1986); Robert G. Pachella, Personal Values and the Value of Expert Testimony, 10 L & Hum Behav [Nos. 1, 2] 145 (1986); Carol Krafka and S. Penrod, Reinstatement of Context in a Field Experiment on Eyewitness Identification, 49 J of Personality & Soc Psychol [No. 1] 58 (1985); V. Konecni and E. Ebbesen, Social Psychology and the Law: The Choice of Research Problems, Settings, and Methodology, The Criminal Justice System: A Social-Psychological Analysis, at 25-44 (1982); V. Konecni and E. Ebbesen, A Critique of Theory and Method in Social-Psychological Approaches to Legal Issues, 2 Perspectives in Law and Psychology: The Trial Process, at 481 (1981).

11 . The term “parameter” refers to “1 ..Math. a. a constant or variable term in a function that determines the specific form of the function but not its general nature, as a in f(x) = ax, where a determines only the slope of the line described by f(x). b. One of the independent variables in a set of parametric equations. 2. Statistics, a variable entering into the mathematical form of any distribution such that the possible values of the variable correspond to different distributions. 3. Computers, a variable that must be given a specific value during the execution of a program or of a procedure within a program.” (See Random House Dictionary of English Language 1408 [2d ed 2000]).

12 . E.g., S. Kassin et al., supra (2001); E. Ebbesen et al., supra at 8-12 (1996); H. Egeth, supra at 153-154 (1995); J. Yuille et al., supra at 934 (1994); H. Egeth, supra (1994); H. Egeth, What Do We Not Know About Eyewitness Identification?, 48 Am Psychol 577, 578 (1993); George Handler, Memory, Arousal, and Mood: A Theoretical Integration, The Handbook Of Emotion and Memory: Research and Theory, at 100-105 (1992); J. McKenna et al., supra at 286 [1992]; S. Christianson, supra (1992); Alafair Burke, Friderike Heuer and Daniel Reisberg, Remembering Emotional Events, 20 Memory & Cognition [No. 3] 277 (1992); J. Read, J. Yuille & P. Tollestrup, Recollections of a Robbery: Effects of Arousal and Alcohol upon Recall and Person Identification, 16 L & Hum Behav [No. 4] 425 (1992); J. Yuille et al., supra at 207-208 (1992); F. Heuer et al., supra at 496 (1990); F. Heuer and D. Reisberg, Vivid Memories of Emotional Events: The Accuracy of Remembered Minutiae, 18 Memory & Cognition [No. 5] 496, 496 (1990); S. Kassin et al., supra (1989); J. Yuille et al., supra at 299-300 (1986).