Opinion by JUDGE BRIGGS.

Defendant, Earl William Campbell, appeals the judgments of conviction entered on jury verdicts finding him guilty of criminal attempt, aggravated robbery, felony theft, and being an habitual criminal. He contends the trial court erred in excluding expert testimony on the reliability of eyewitness identifications. We reverse the convictions and remand for a new trial.

Defendant has previously appealed his robbery and related convictions in this and a companion case. At trial he presented the expert's testimony as an offer of proof and relied upon the same offer in the companion case.

The expert was prepared to testify about studies which indicate that there is a weak correlation between a witness' expression of confidence in his or her identification of a suspect and the accuracy of that witness' recollection; that contrary to popular belief, high levels of stress impair rather than improve memory; that the tendency of victims to focus on the weapon used during the crime reduces their ability to recall details about the offender's appearance; and that a witness' viewing of a composite drawing or photograph can influence the witness' memory of the criminal. In both cases the trial court excluded the testimony because the evidence did not meet the test of trustworthiness set out in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).

This court in separate opinions upheld the exclusion of the evidence in both cases without addressing the issue of trustworthiness on the basis that the subject was within the ken of the jury and not appropriate for expert testimony. We affirmed the conviction in this case, People v. Campbell, No. 87CA1956 (Colo.App. Nov. 30, 1989) (Not Selected for Official Publication), but in the companion case reversed on other grounds and remanded for a new trial. People v. Campbell, 785 P.2d 153 (Colo.App. 1989)

The supreme court granted certiorari in both proceedings and consolidated the cases. It recognized that historically federal and state courts had almost uniformly upheld the trial court's exclusion of this type of testimony. However, after reviewing recent decisions allowing this type of evidence, the court concluded that admission of expert testimony on the reliability of eyewitness identification may be proper in some cases. Campbell v. People, 814 P.2d 1 (Colo. 1991).

Both cases were remanded to the trial court to evaluate the proffered testimony under CRE 702 and 403. The trial court was instructed to grant a new trial in this case if it determined that the testimony was admissible. On remand, the trial court reviewed the prior offer of proof and found the expert testimony admissible only in the companion case. Defendant's convictions in this case were therefore affirmed without a new trial.

The sole issue before us is whether the offer of proof relied upon by the trial court to admit the evidence in the companion case was also a sufficient basis to admit the evidence in this case. We conclude that it was.

I.

An employee on break in a grocery store storage room noticed a man peer quickly through the door into the room. When the employee returned to work a few minutes later she noticed the man was in the store.

Returning from her break the employee replaced a coworker at a cash register. A short time later the same man walked up and placed some items on the counter. He pulled a gun, demanded all the store's money, and threatened to kill the employee. As she collected money from the safe and cash registers in the front of the store she continued watching the robber.

During this time, a customer had been in the store with her two little girls. As they approached the counter the man threatened to kill the employee and the two children. The customer did not hear the threat and had not realized there was a robbery in progress. However, she looked at the man as he turned to face her, looked at the store clerk, realized there was trouble, and immediately followed the man's order to take the children to the back of the store.

The robber told the employee he was going to shoot her. She watched him slowly cock the gun, point it at her, and shoot. In that moment she believed she had been shot. The bullet missed. The man left with the money. The entire episode transpired in approximately five minutes.

Shortly after the robbery the employee described the man as approximately five feet eight inches tall, with dark eyes and a face that was "pale," "pockmarked," "rough," and "ruddy." That night, in the course of doing a composite picture of the robber at the police station, the employee told the police artist the robber had "heavy active acne," scars, and sores. She saw the composite on two subsequent occasions, once in a local gas station and once in a newspaper article.

Approximately one month later, the employee identified the defendant in a photographic lineup. At trial she testified she was so emotional when she made the identification that she started crying because the picture "looked just like the man in the store" and she "couldn't believe they really had a picture of him." She testified that when she was viewing the lineup she was not relating it to her composite drawing, but rather to what the perpetrator looked like on the day of the robbery.

The employee also identified defendant at a pre-trial hearing and again at trial. She testified at trial that since a few months after the robbery she had seen the defendant's face in reoccurring nightmares.

In relating the details of the robbery at trial the employee described the skin of the robber as "real pale." However, after viewing the defendant from a very close distance she testified that, although he had the same general complexion as the perpetrator, the defendant's skin was tanner. Defendant's rough and ruddy complexion also did not appear as prominent to the witness as the robber's. Further, at the time of the photo identification the employee thought the face in the picture was heavier than that of the robber.

The customer also made a composite the same night as the robbery and positively identified the defendant from photographic lineups on two separate occasions; one approximately one month after the robbery and the other immediately prior to trial. She also identified the defendant in a pre-trial hearing and at trial, and testified she had seen defendant's face in nightmares before she had been shown any photos of him.

At trial, the customer also described the robber as having a white, scarred complexion. Like the employee, she testified that the defendant had the same general complexion as the perpetrator, but that the robber was paler and possessed facial scarring that was a little more pronounced than the defendant's. Both the employee and the customer immediately after the robbery had described the perpetrator as approximately three to four inches shorter than the defendant.

Both eyewitnesses nevertheless testified they were very positive that defendant was the robber. The employee was so confident in her identification that she would "bet her life on it." The record contains no direct or circumstantial evidence linking defendant to the robbery other than the testimony of the two eyewitnesses.

The defendant presented three witnesses at trial, a realty agent, a prospective tenant, and a meteorologist, to establish that on the day and at the time of the offense he was with the realty agent a substantial distance from the crime scene looking at a rental home. In addition, a doctor who examined the defendant approximately one month after the crime and a witness who through work knew defendant at the time of the robbery both testified that the defendant's face had not changed from then until the time of trial.

After receiving the case, the jury deliberated the balance of that day, a Friday, and returned on Monday to continue. At one point the jury had a question because it was deadlocked. The court instructed the jury to continue deliberating. It eventually found defendant guilty on all counts.

In the companion case, defendant was charged with a similar grocery store robbery committed a week later. As here, the sole evidence linking defendant to the robbery was the testimony of two eyewitnesses.

At trial both witnesses were confident in their identifications. Both eyewitnesses nevertheless testified about discrepancies in the complexion of the perpetrator and that of the defendant. One eyewitness thought the other's composite of the perpetrator was inaccurate. She also had substantial difficulties in making an identification prior to trial. In fact, she identified somebody other than the defendant with a ninety-percent degree of certainty as being the perpetrator and hesitantly selected a photograph of someone other than the defendant from a photographic lineup.

II.

Our supreme court has concluded that the analysis in United States v. Downing, 753 F.2d 1224 (3rd Cir. 1984), which uses a test derived from the federal rules of evidence identical to CRE 702 and 403, is "a sound method for ensuring that the admission of expert testimony on the accuracy of eyewitness identification will enhance the truth-seeking function of trial." Campbell v. People, 814 P.2d at 7. The test requires the trial court, after an evidentiary hearing, to consider three factors in determining whether the proffered expert testimony will satisfy the requirement of CRE 702 that the evidence assist the trier of fact: (1) the reliability of the scientific principles upon which the expert testimony rests; (2) the likelihood that introduction of the testimony may in some way overwhelm or mislead the jury; and (3) the "fit"; that is, the degree to which the proffered scientific evidence ties to the particular features of the eyewitness identification at issue. The court then retains jurisdiction to exclude under CRE 403 relevant evidence if its probative value is substantially outweighed by considerations of waste of time or confusion of the issues. United States v. Downing, supra.

In this case, the trial court had already properly conducted an initial inquiry by hearing the proffered expert testimony out of the presence of the jury. The court relied on that offer of proof in both cases on remand and that procedure is not challenged on appeal by either party.

The court on remand concluded that, in this case, the testimony presented in the offer of proof "would not assist the trier of fact to understand the evidence or to determine a fact in issue," and thus, excluded the expert's testimony in the "form of an opinion or otherwise." It excluded the evidence under both CRE 702 and 403.

The trial court did not state its analysis other than to observe that here, unlike the companion case, the eyewitnesses positively and unequivocally identified defendant and there was no uncertainty in their identification. We conclude that under the Downing analysis the evidence should have been admitted.

A.

The first prong of the Downing test is whether the scientific principles upon which the expert testimony rests are reliable. The prosecution cross-examined the expert during the preliminary hearing, but submitted no independent evidence to challenge the competence of the expert or the reliability of the studies on which she relied.

On appeal, the People first argue that the studies do not involve actual crimes. It would of course be difficult to design studies which replicate real-life violent crimes. There was no evidence presented to indicate that the studies for this reason were unreliable as a basis for expert testimony.

The Downing court concurred in the observation of the California Supreme Court in People v. McDonald, 37 Cal.3d 351, 371, 208 Cal.Rptr. 236, 245, 690 P.2d 709, 718 (1984): "The consistency of the results of these studies is impressive, and the courts can no longer remain oblivious to their implications for the administration of justice." We reach the same conclusion as the Third Circuit: "From the facts available in the record and otherwise, it would appear that the scientific basis for the expert evidence in question is sufficiently reliable to satisfy Rule 702." United States v. Downing, 753 F.2d at 1241.

The People next argue that the results of the studies on the correlation between eyewitness confidence and accuracy of identification are inconsistent. This merely demonstrates "a fairly weak relationship" between confidence and accuracy, which is defendant's point. See United States v. Stevens, 935 F.2d 1380 (3d Cir. 1991). The trial court's admission of the expert's testimony in the companion case indicates a proper determination that this evidence satisfied the requirement of reliability.

B.

The second prong of the Downing test arises from a concern with the jury being overwhelmed by expert testimony if, for example, a technique had assumed "a posture of mythic infallibility." United States v. Downing, 753 F.2d at 1239. The Third Circuit considered the danger of misleading the jury the greatest if the jury is not presented with the data on which the expert relies, but instead must accept the expert's assertions as to the accuracy of his conclusions.

Here, the data on which the expert relied was to be presented to the jury and there was no indication the proffered expert testimony was likely to overwhelm or mislead the jury. The trial court's admission of the testimony in the companion case again indicates it correctly determined that this portion of the Downing test was satisfied.

C.

We now turn our analysis to the proffered "fit" of the expert's testimony to the evidence. It was apparently because of a difference in the evidence pertaining to the identifications that the trial court concluded the testimony would be excluded in this case but admitted in the companion case.

The defendant must make a detailed offer of proof on the record, including an explanation of precisely how the expert's testimony is relevant to the eyewitness identifications under consideration. For example, the offer of proof should establish the presence of factors, such as stress, which have been found to impair the accuracy of eyewitness identifications, or the absence of an expected correlation, such as expression of confidence in the identification and actual accuracy. See United States v. Downing, supra.

Here, defendant's expert testified in limine regarding studies showing a weak correlation between witness confidence and reliability of identification. She also testified that, contrary to average juror expectations, stress actually decreases rather than increases accuracy of perception, with subsequent distortion of recall; that the presence of a gun or other weapon does not consciously assist but rather tends to distract a witness such that it could impair his or her memory, again contrary to the average juror expectations; that there is a phenomenon called "photo bias identification" if the witness has created a composite picture, and a mental process called "blending" which can reinforce the composite if there have been subsequent photo identifications, resulting in the identification of someone who looks more like the composite than the actual perpetrator.

The expression of confidence by the eyewitnesses in their identifications "fit" the proffered expert testimony regarding the correlation between witness confidence and accuracy of identification. See United States v. Stevens, supra; People v. McDonald, supra. Likewise, the presence of the gun and the witness stress "fit" the proffered testimony on the detrimental impact these elements have on eyewitness identification. People v. McDonald, supra. Finally, the presence of identifications through composite drawing and photo lineups "fit" the proffered testimony regarding "photo bias" and "blending." See State v. Chapple, 660 P.2d 1208 (Ariz. 1983).

The court's exclusion of the evidence in this case and admission in the companion case appears to have been based on the fact that the identifications in this case were more "positive" and "unequivocal," and there was "no uncertainty, whatsoever, in either witnesses' identification of defendant." However, "[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 a witness' recollection of an event." See United States v. Downing, 753 F.2d at 1231 (fn.6).

The courts permitting expert eyewitness testimony on these matters do so in part because there is no other effective way to reveal any weaknesses in an eyewitness identification. It is also in part because the evidence is "counterintuitive" or "contrary to common wisdom." The evidence therefore assists the trier of fact in assessing the reliability of eyewitness testimony. See United States v. Stevens, supra; State v. Chapple, supra; People v. McDonald, supra.

We agree with the Third Circuit's observation in similar circumstances:

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

Downing, 753 F.2d 1242.

United States v. Stevens, 935 F.2d at 1400.

We conclude that the under CRE 702 the expert's testimony is admissible. The question remaining is whether the evidence was properly excluded under CRE 403.

III.

The trial court on remand found that "any possible probative value of that evidence would be substantially outweighed by the danger of confusion of the issues or misleading the jury." The court provided no explanation why this would be true in this case but not the companion case.

The Third Circuit in Downing concluded that concern with confusion is more significant when there is evidence of defendant's guilt other than the eyewitness testimony. The other evidence may minimize the probative value of the expert's testimony, and it may possess the propensity to make the testimony cumulative, confusing, and a waste of time. See State v. Chapple, supra. Here, there was no other evidence linking defendant to the robbery.

The Downing Court's analysis is appropriate in the circumstances of this case:

"The availability of other methods that would serve the purposes for which the appellant seeks to introduce expert testimony may also serve to justify exclusion under Rule 403, though we do not perceive the viability of such an alternative in this case. Moreover, as an abstract proposition, it would seem anomalous to hold that the probative value of expert opinion offered to show the unreliability of eyewitness testimony so wastes time or confuses the issue that it cannot be considered even when its putative effect is to vitiate the only (eyewitness) evidence offered by the government."

United States v. Downing, 753 F.2d at 1243.

The trial court is entitled to deference in the exercise of its discretion to admit or exclude expert testimony on eyewitness identification. However, that discretion must be exercised within the Downing framework. As in Chapple, the facts or inferences from them in this case are not in dispute and there are few, if any, conflicting factual, procedural, or equitable considerations. Hence, our conclusion contrary to the trial court is "not predicated upon conflicting factual contentions or equitable considerations." State v. Chapple, 660 P.2d at 1223-24.

A "decision excluding expert testimony offered by a criminal defendant is perhaps somewhat more susceptible of reversal because of the courts' sensitivity to the defendant's need and lack of access to the personnel available to the state." People v. Williams, 790 P.2d 796, 798 (Colo. 1990). In the case of expert testimony on eyewitness identifications, we must respond to the call for "the courts to face up to the reliability problems of eyewitness identification, to inform themselves of the results of scientific studies of those problems, and to allow juries access to that information in aid of their fact finding tasks." People v. McDonald, 37 Cal. 3d at ___, 208 Cal.Rptr. at 244, 690 P.2d at 717.

The United States Supreme Court years ago observed: "The vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification." United States v. Wade, 388 U.S. 218, 228, 87 S.Ct. 1926, 1933, 18 L.Ed.2d 1149, 1158 (1967). The Second Circuit has further elaborated: "Centuries of experience in the administration of criminal justice have shown that convictions based solely on testimony that identifies a defendant previously unknown to the witness is highly suspect. Of all the various kinds of evidence it is the least reliable, especially where unsupported by corroborating evidence." Jackson v. Fogg, 589 F.2d 108, 112 (2d Cir. 1978); see also People v. McDonald, supra; State v. Chapple, supra; Note, Did Your Eyes Deceive You? Expert Psychological Testimony on the Unreliability of Eyewitness Identification, 29 Stan. L. Rev. 969 (1977); G. Williams, The Proof of Guilt 106 (3d ed. 1963).

This is not to say that there is an expectation that expert eyewitness testimony should be admitted in every case involving an eyewitness identification. The guiding principle which we adopt has been well summarized:

"We expect that such evidence will not often be needed, and in the usual case the appellate court will continue to defer to the trial court's discretion in this matter. Yet deference is not abdication. When an eyewitness identification of the defendant is a key element of the prosecutor's case but is not substantially corroborated by evidence giving it independent reliability, and the defendant offers qualified expert testimony on specific psychological factors shown by the record that could have affected the accuracy of the identification but are not likely to be fully known to or understood by the jury, it will ordinarily be error to exclude that testimony."

People v. McDonald, 37 Cal.3d at ___, 208 Cal.Rptr. at 254, 690 P.2d at 727.

We conclude that the proffered expert testimony on the reliability of eyewitness identifications in this case should have been admitted.

The judgments of convictions are reversed, and the cause is remanded for a new trial.

JUDGE PIERCE and JUDGE RULAND concur.