March 30, 2017
IN THE SUPREME COURT OF THE STATE OF OREGON
STATE OF OREGON,
Respondent on Review,
SHAWN EDWIN HAUGEN,
Petitioner on Review.
(CC 10CR0636; CA A151535; SC S063754)
On review from the Court of Appeals.* Argued and submitted September 22, 2016.
Neil F. Byl, Deputy Public Defender, Salem, argued the cause and filed the brief for the petitioner on review. Also on the briefs was Ernest G. Lannet, Chief Defender, Office of Public Defense Services.
Andrew M. Lavin, Assistant Attorney General, Salem, argued the cause and filed the brief for the respondent on review. Also on the brief were Ellen F. Rosenblum, Attorney General, and Benjamin Gutman, Solicitor General.
The decision of the Court of Appeals is reversed, and the case is remanded to the circuit court for further proceedings consistent with this decision.
* Appeal from Josephine County Circuit Court, Pat Wolke, Judge. 274 Or App 127, 360 P3d 560 (2015).
Case Summary: The victim was assaulted late in the evening in the parking lot of a bar. Immediately after the attack, he could not describe to a police officer exactly what had happened or who had assaulted him, but, a few days later, the victim identified defendant as one of the perpetrators. Before his trial on an assault charge, defendant moved to exclude the eyewitness identification, and, applying the then-applicable test for the admissibility of eyewitness testimony, the trial court ruled that the eyewitness identification was admissible. Defendant was convicted of the charged offense. Defendant appealed his conviction, and while the appeal was pending, the Oregon Supreme Court announced its decision in State v. Lawson/James, 352 Or 724, 291 P3d 673 (2012), in which it adopted a new test for the admissibility of eyewitness identification evidence. The Court of Appeals applied the new test and affirmed defendant’s conviction. Held: The identification procedures used in the case raised serious questions about the reliability of the victim’s identifications of defendant under Lawson/James, and, therefore, remand is necessary to give the trial court the opportunity to consider the admissibility of the identifications under the correct standard.
The decision of the Court of Appeals is reversed, and the case is remanded to the trial court for further proceedings consistent with this decision.
In this criminal case, defendant was convicted of one count of third-degree assault, based mainly on the victim’s eyewitness identification of him. Before his trial, defendant moved to exclude the eyewitness identification. Applying the test for admissibility of eyewitness testimony set out in State v. Classen, 285 Or 221, 590 P2d 1198 (1979), the trial court ruled that the victim’s eyewitness identification was admissible. While the case was pending on appeal, this court announced its decision in State v. Lawson/James, 352 Or 724, 291 P3d 673 (2012), in which the court substantially revised the Classen test for determining the admissibility of eyewitness testimony. In the Court of Appeals, defendant argued that the identification procedures used in this case raised serious questions about the reliability of the identification under Lawson/James, and, therefore, that the Court of Appeals should remand the case to the trial court for a new hearing and trial, with the trial court utilizing the Lawson/James test. The Court of Appeals disagreed, concluding that, even under Lawson/James, the trial court correctly denied defendant’s motion to suppress. State v. Haugen, 274 Or App 127, 360 P3d 560 (2015). We allowed review and, for the reasons explained below, we reverse the decision of the Court of Appeals and remand the case to the trial court for further proceedings consistent with this decision.
The following facts are undisputed. In September 2010, the victim met two friends at a bar in Grants Pass, Oregon. Already present were several people wearing clothing indicating that they were members of the Vagos motorcycle gang. The victim did not know any of the Vagos members present at the bar, but one of the victim’s companions knew and had a conversation with one of the gang members, a man later identified as Rives. During the evening, Rives approached the table where the victim was sitting and asked the victim if he knew a man named Moore, a former member of the Vagos who, several years earlier, had been a witness for the prosecution against other Vagos members. The victim answered, “Well, yes I do. I understand, you know, he used to be a Vago.” Rives then became irate and began ranting about Moore. The victim merely nodded in response. The encounter lasted two to three minutes.
The victim was at the bar for about two hours. During that time, he overheard a conversation in which several gang members were teasing defendant, who also appeared to be a member of the Vagos, about wearing a red shirt and tie to sell cars on a car lot. When the victim was leaving the bar around 12:30 a.m., defendant, who was then standing in a hall near the exit door, looked directly at the victim and said, “Have a good fucking night.” As the victim stepped outside the door to the bar, he encountered Rives. The victim thought that he had heard something behind him, turned to look, and saw someone whom he could not clearly see holding the door shut. Rives then said to the victim, “Are you here to kick us out?” The victim responded, “No, I’m going home. I just want to walk to my car.” At that moment, the victim saw that defendant was standing to his right. Defendant touched the victim’s shoulder and said, “Well, walk to your car.” As the victim began walking to his car, someone punched him on the side of the head. He fell to the ground and someone kicked him in the chest or shoulder. As the victim tried to get up, someone else struck him in the head with what he later thought was a small metal hammer. The assailants then left. The victim was nearly knocked unconscious but was able to get up and walk to his truck and drive home.
Once home, the victim called 9-1-1 and reported that he had been assaulted by several members of the Vagos motorcycle gang and that he had been punched and “blindsided.” A police officer, Nicklason, arrived later that night to interview the victim. The victim appeared to be in pain and to have been significantly injured. The victim told Nicklason that four to six Vagos members approached him as he was leaving the bar and asked him if he was trying to run them off. The victim stated that he was later struck on the right side by someone he did not see. He thought he had been punched but was not sure. He did not mention being hit with a hammer. The victim told Nicklason that there were no witnesses to the assault, and, although he recognized his assailants as part of a group that had been in the bar, “he couldn’t recognize them specifically individually.” That is, the victim did not describe to Nicklason the race, height, weight, or any identifying feature of the individuals who assaulted him, other than that they were male and members of the Vagos gang. The victim attributed his inability to provide a description of his assailants to the facts that it was dark and it was a brief encounter.
Nicklason forwarded the case to Detective Brown, who was responsible for dealing with outlaw motorcycle gangs in the area. Brown interviewed the victim at the police station five days later. The victim once again relayed the events of the evening when he was assaulted, this time in more detail. The victim told Brown that the people at the bar were definitely members of the Vagos gang—they were “flying their colors,” that is, wearing green bandanas and jackets with the Vagos insignia. The victim stated that the person who had struck him from the side was a “great big guy,” about 230 pounds, not fat, in his late 20s or early 30s, and had been teased at the bar earlier for dressing up for his job as a car salesman. The victim also told Brown that the person who had hit him with the hammer was a “little fat guy,” probably in his 40s, with a long ponytail. Brown asked the victim if the first assailant was “pretty buff.” When the victim answered in the affirmative, Brown said, “I think I know who you’re talking about.” After talking to the victim for a few more minutes, Brown said, “[W]hat we’ll do here in a minute, [victim], I’ll show you some photographs—and maybe that will help us—once we have some photos we can go through—and we’ll identify—who the little fat guy is.” Brown then gave the victim the following disclaimer:
“You are about to be shown some photographs. Just because the officer is showing you these photos, you are in no way obligated to identify anybody, okay? The person who committed the crime may or may not be in this group of photographs.”
Brown explained to the victim that he would ordinarily show him six photographs, “but because there’s so many people involved in this I’m just going to show you a group of photographs.” Brown then presented the victim with a binder of 23 photos. The photos were not chosen based on the subjects’ similarities to the descriptions that the victim provided to Brown of his assailants and the other people in their group; rather, all were Department of Motor Vehicle driver license photographs of known Vagos “Outlaw Motorcycle Gang” members and associates. The photos did not include any notes or names.
The victim identified one person he recognized as a “good man,” and Brown agreed and provided that individual’s name. Next, the victim identified a man he knew as “Ronnie.” Brown confirmed that man’s name. The victim thought that Ronnie had been present, but he was not sure whether Ronnie was involved in the assault. Then the victim pointed to a picture of defendant, stating, “That sure looks like the guy that hit me right there,” and “I’m pretty sure it was him.” Brown said, “We’ll just call him White Boy for now,” and set that picture aside. After identifying another person he thought was present at the bar, the victim asked Brown how old the pictures were. Brown responded that they were a couple of years old. Brown and the victim continued through the remaining pictures, with the victim pointing out people he thought might have been present in the bar on the night of the assault as well as people he knew, and with Brown providing feedback such as providing the real names and gang names of some of the people pictured and commenting on who was in prison, who had distinctive physical traits or habits, whose hair had or had not changed since the DMV photo was taken, and the like.
At one point, Brown directed the victim’s attention to a picture of Rives, saying, “This gentleman now has a little bit more gray. I’m talking about ‘Six Ball,’ or Steve Rives. He’s got more gray, but he’s got a ponytail that long on his back.” The victim responded,
“That sure looks like the little fat man right there. He looks different in this. His hair appears darker. But that sure looks like him because he did—his hair was lighter and the ponytail was probably about that long. * * * I believe he’s one of the ones involved. I wanted to come back to him, but * * * that’s why I asked you when I looked at his picture how old these pictures are.”
Brown set that photo aside as well.
Brown and the victim examined the remaining pictures, with Brown continuing his commentary and the victim pointing out individuals whom he thought might have been present at the bar on the night of the assault. When they had finished going through the photos, Brown returned the victim’s attention to defendant’s picture. Brown asked the victim how certain he was that defendant was the person who first struck him. The victim replied that he was 75 to 80 percent certain. He also told Brown that defendant was the person who, earlier in the night, had said to him, “Have a good fucking night.”
Brown then offered to present the victim with a series of photos of defendant and others that had been taken during an unrelated surveillance action a week earlier, saying that “maybe that will help a little bit.” Brown told the victim that he wanted to show him the more recent photos “because there are some significant differences in both the gentlemen that you have said that have assaulted you.” In the second set of photos, the pictured individuals were all wearing gang apparel. The victim identified Rives as the person who had hit him with the hammer, saying that he was “99.9 percent sure” it was Rives. The victim also identified defendant in one of the photos and again said that he was 80 percent certain that defendant was the person who first struck him. After being shown yet another recent surveillance photo of a group of individuals that included defendant, the victim raised his certainty level to 90 percent.
Defendant was charged with one count of third-
degree assault. Before the trial, defendant moved to suppress the victim’s identification of him during the interview with Brown and to suppress any future in-court identification. Defendant argued that the identification was inadmissible under Classen, because the lineup procedure Brown had used was suggestive, it needlessly departed from prescribed procedures for avoiding suggestiveness, and the circumstances surrounding the identification indicated that the victim’s identification was not made independent of the suggestive lineup. The trial court denied the motion. In a letter opinion, the trial court ruled that the lineup procedure that Brown had used was not unduly suggestive, because Brown did not violate guidelines for photographic identifications and because the court could find no other legal reason for concluding that the photo lineup was unduly suggestive. The trial court also found that the victim’s identification was reliable when considered under four of the factors that this court had set out for determining eyewitness identification reliability in Classen (the witness’s ability to observe the defendant; the timing and completeness of the witness’s description; the witness’s certainty in the identification; and the lapse of time between the event and the identification) and in light of the fact that various details that the victim remembered about his assailant were later corroborated by other witnesses at the bar. For those reasons, the trial court concluded that “the accuracy of the identification in this case is a matter for the trial jury; and not for this court.”
The case proceeded to trial in April 2012. During the trial, the victim identified defendant in court as the person who had punched and kicked him, as the person who had said, as the victim was leaving the bar, “Have a good fucking night,” and as the person who had been teased as a car salesman at the bar. The victim testified that he was 100 percent certain that defendant was the person who assaulted him. Ultimately, the jury convicted defendant of the charged offense.
After defendant filed the notice of appeal in this case, this court issued its opinion in Lawson/James, which we discuss in more detail below. Defendant argued to the Court of Appeals that Lawson/James dramatically changed the legal landscape for determining whether an eyewitness identification is admissible and that serious questions about the reliability of the victim’s identification of defendant under Lawson/James warranted reversal and remand for a new hearing in which the trial court could apply the Lawson/ James test. As noted, the Court of Appeals concluded that, even under Lawson/James, the trial court correctly denied defendant’s motion to suppress.
As discussed, the trial court applied the Classen test at the suppression hearing to determine the reliability, and thus the admissibility, of the eyewitness identification. It is helpful at this point, then, to describe the Classen methodology, which was intended to ensure that only sufficiently reliable identifications are admitted into evidence. Classen required a two-step inquiry: First, the court was to determine whether the process leading to the identification was suggestive or needlessly departed from procedures prescribed to avoid suggestiveness. 285 Or at 232. If so, then, second, the court was to determine whether the prosecution had established that there was some source for the identification other than the suggestive procedure or that other evidence substantially negated the risk that the identification was stimulated by the suggestive procedures. Id. Classen listed five nonexclusive factors for the trial court to consider in determining whether an identification had been made independent of the suggestive procedure: (1) the witness’s opportunity to clearly view the suspect; (2) the attention that the witness had paid to the suspect’s features; (3) the timing and completeness of the witness’s description of the suspect; (4) the witness’s certainty in his or her identification; and (5) the lapse of time between the original observation and the identification. Id. at 232-33. In this case, the trial court found that the identification process that Brown had used was not suggestive. That is, the court found that Brown did not needlessly depart from police guidelines for conducting identifications and that there was no other legal basis for concluding that the process was suggestive. Notwithstanding that initial conclusion, the trial court also reviewed the Classen factors and concluded that the victim’s identification of defendant was made independent of the identification procedure used.
While the case was pending on appeal, this court decided Lawson/James, which, as we have noted, significantly changed the framework for determining the admissibility of eyewitness identification testimony that is asserted to have been tainted by suggestive police practices. The impetus for revisiting the analysis that the court had used in Classen was new scientific research surrounding eyewitness identifications. Lawson/James, 352 Or at 73940. In particular, the court discussed two categories of factors that, according to the scientific community, affect the reliability of eyewitness identification: so-called “estimator variables” and “system variables.” Estimator variables are “characteristics of the witness, the alleged perpetrator, and the environmental conditions of the event that cannot be manipulated or adjusted by state actors.” Id. at 740. System variables relate “to the circumstances surrounding the identification procedure itself that are generally within the control of those administering the procedure.” Id. In light of the scientific research and those variables, the court in Lawson/ James concluded that the process outlined in Classen did “not accomplish its goal of ensuring that only sufficiently reliable identifications are admitted into evidence.” Id. at 746. In fact, the court stated, the reliability factors that the court listed in Classen were both incomplete and, at times, inconsistent with modern scientific findings, and the Classen inquiry itself was “somewhat at odds with its own goals and with current Oregon evidence law.” Id. Moreover, the Classen test prohibited courts from considering whether an identification was reliable until evidence of some suggestiveness was introduced. Id. at 746. And the Classen factors relied heavily on the witness’s self-reports to establish the existence or nonexistence of suggestibility factors, when current scientific research indicates that self-reported evidence of reliability can be inflated or exaggerated by the suggestive procedure itself. Id. at 748.
The court in Lawson/James therefore set out a new framework for determining the reliability of eyewitness identification testimony, based on generally applicable provisions of the Oregon Evidence Code, including the foundational principles that, under OEC 402, all relevant evidence is admissible; that, under OEC 401, evidence is relevant if it has any tendency to make the existence of a fact of consequence to the determination of the action more or less probable than it would be without the evidence; and that, under OEC 403, relevant evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice. Lawson/James, 352 Or at 752, 756. The court announced a two-step process for determining the admissibility of eyewitness identification testimony. First,
“[w]hen a witness’s perceptions are capable of supporting an inference of identification, but are nevertheless met with competing evidence of an impermissible basis for that inference—i.e., suggestive police procedures—an issue of fact arises as to whether the witness’s subsequent identification was derived from a permissible or impermissible basis.”
Id. at 755. At that point, the state, as the proponent of the evidence, has the burden to present sufficient evidence from which a juror could find that the witness’s identification was derived from a permissible basis. State v. Hickman, 355 Or 715, 729, 330 P3d 551 (2014). That is, to meet its preliminary burden to demonstrate that eyewitness testimony is relevant under OEC 402, the state must provide “proof under OEC 602 that the proffered eyewitness has personal knowledge of the matters to which the witness will testify, and proof under OEC 701 that any identification is both rationally based on the witness’s first-hand perceptions and helpful to the trier of fact.” Lawson/James, 352 Or at 761-62.
As the court stated,
“Although the initial admissibility requirements for eyewitness identification evidence establish a minimum baseline of reliability, the persuasive power of the evidence that meets that standard may nevertheless vary greatly, and many identifications possessing relatively low probative value may still pass that initial test.” Id. at 758.
If the state introduces evidence sufficient to support a finding that the eyewitness testimony is not barred by OEC 402, then the second step comes into play. At that point, the defendant assumes the burden of proving that OEC 403 nevertheless requires its exclusion. Id. at 757. The court stated,
“Thus, even after finding that the evidence meets the minimum requirements of OEC 602 and 701, trial courts must still conduct a thorough examination of all the pertinent factors in order to determine the probative value of the evidence under OEC 403.” Id. at 758.
Probative value is determined under OEC 403 by
considering the relative reliability of the eyewitness testimony, given the system variables and the estimator variables. The court stated, “The more factors—the presence of system variables alone or in combination with estimator variables—that weigh against reliability of the identification, the less persuasive the identification evidence will be to prove the fact of identification, and correspondingly, the less probative value that identification will have.” Id. at 757.
Further, the court stated,
“As a discrete evidentiary class, eyewitness identifications subjected to suggestive police procedures are particularly susceptible to concerns of unfair prejudice. Consequently, in cases in which an eyewitness has been exposed to suggestive police procedures, trial courts have a heightened role as an evidentiary gatekeeper because ‘traditional’ methods of testing reliability—like cross-examination— can be ineffective at discrediting unreliable or inaccurate eyewitness identification evidence.” Id. at 758.
Moreover, the court stated, expert testimony is particularly useful to this inquiry for two reasons. First, “traditional methods of informing factfinders of the pitfalls of eyewitness identification—cross-examination, closing argument, and generalized jury instructions—frequently are not adequate to inform factfinders of the factors affecting the reliability of such identifications.” Id. at 759. And second, system and estimator variables may be unknown to factfinders or contrary to common assumptions. Id. at 761.
Ultimately, though, as the court stated, the decision whether to admit or exclude eyewitness identification evidence is committed to the discretion of the trial court, and, as the court observed, “it is doubtful that issues concerning one or more of the estimator variables that we have identified will, without more, be enough to support an inference of unreliability sufficient to justify the exclusion of the eyewitness identification.” Id. at 762; see also Hickman, 355 Or at 726 (same).
The court then applied its new framework to the facts in Lawson and James. In Lawson, the court concluded that the record raised serious concerns about the reliability of the identification evidence proffered below. In that case, the victim and her husband had driven in their trailer to a campsite where the victim’s husband earlier had pitched a tent. When they arrived, the defendant’s truck was parked in their parking space and the defendant had moved into their tent. The husband told the defendant that it was their tent. The defendant apologized, moved his belongings out of the tent and into his truck, and moved to a vacant campsite nearby. The defendant was in the victim’s view for about 40 minutes. Later that same evening, the victim was standing by the window of the trailer when she was shot in the chest with a large caliber hunting rifle. Her husband called 9-1-1, but he too was shot while speaking with the operator, and he died shortly thereafter. The dispatcher called back and the victim told the dispatcher that she and her husband had been shot and that she did not know who had shot them.
When first responders arrived, they found the vic-
tim lying in the trailer, seriously injured but conscious. She was transported to a hospital. An ambulance attendant testified that the victim was rambling and hysterical, and other medical personnel reported that the victim referred to the shooter as “they,” mentioned several different people as the shooter, and stated that she had not seen the perpetrator and did not know who it was. The victim was near death when she arrived at the hospital and was immediately taken into surgery.
When a police detective interviewed the victim the next day, she was heavily medicated and sedated, and she could not speak because of a breathing tube in her throat. The defendant had come to the attention of police officers when he had volunteered to them that he had encountered the victim and her husband at the campsite in the morning on the day that they were shot, and the detective later showed the victim a photo lineup that included a picture of the defendant. The victim was unable to identify the defendant in that photo lineup. After the detective posed some leading questions, the victim agreed that she had seen the person who had shot her earlier in the day at the campsite, several hours before she was shot. Two weeks later, when interviewed again, the victim remembered that the perpetrator had covered her face with a pillow after entering the trailer so as to obscure her view of him, and she said that she could not identify the man because of the pillow and because it was dark. Two weeks after that, the victim told the detective that, notwithstanding the pillow over her face, she had briefly seen the perpetrator in the trailer, but she would not be able to pick the perpetrator’s picture out of a photo lineup. Between then and the trial two years later, however, on multiple separate occasions, the victim was exposed to the defendant’s likeness in contexts where the defendant was clearly the suspect who had been arrested for the crime. Ultimately, at the defendant’s trial, the victim identified the defendant as the man who had shot her and her husband.
The trial court denied the defendant’s motion to strike the identification, and the defendant was later convicted of aggravated murder, attempted aggravated murder, and other crimes. The Court of Appeals affirmed the defendant’s convictions, finding that, under Classen, the identification procedures prior to the trial were suggestive, but concluding, after weighing the Classen factors, that the victim’s identification of the defendant had been made independent of the suggestive procedures and, therefore, that the trial court had not erred in declining to strike the identification. State v. Lawson, 239 Or App 363, 244 P3d 860 (2010).
On review, this court considered its new framework for determining the admissibility of eyewitness testimony and observed that several estimator variables and system variables were at play. First, factors falling into the category of estimator variables, which negatively affected the reliability of the identification, included the facts that the eyewitness—the victim—was under tremendous stress from the shooting of her husband and in poor physical and mental condition from being shot herself when she first observed the perpetrator entering the trailer; the environmental conditions were poor, as it was dark and the victim was on the floor when she observed the perpetrator; the perpetrator had covered her head with a pillow; the victim stated repeatedly that she never got a good look at the perpetrator; and the in-court identification was over two years after the incident. Lawson/James, 352 Or at 763-64.
Second, factors falling into the category of system variables included the fact that the detective first interviewed the victim when she was heavily medicated and sedated, which, the court stated, together with her impaired view of the defendant, would have made her especially vulnerable to memory contamination from suggestive questioning. Id. at 764. Additionally, the police then questioned her using leading questions that implicitly communicated their belief that the defendant was the shooter and implanted in her mind that the person she had seen earlier in the day at the campsite was the perpetrator. Id. According to the court, those factors could have affected every subsequent attempt that the victim made to recall the perpetrator, and, from that point forward, also would have made it extremely difficult for the victim to mentally separate her brief glimpse of the man in the trailer from the person she had seen at the campsite for 40 minutes in broad daylight earlier in the day. Id. Finally, the victim was unable to identify the defendant until after she had seen his likeness numerous times in suggestive circumstances, but, notwithstanding that fact, the victim identified the defendant at the trial and testified that she had absolutely no doubt as to her identification, stating, “I’ll never forget his face as long as I live. * * * I always knew it was him.” This court stated that the alteration in the victim’s statements over time were indicative of a memory altered by suggestion and confirming feedback. Id. at 765.
After discussing the foregoing estimator variables and system variables, the court concluded:
“In light of current scientific knowledge regarding the effects of suggestion and confirming feedback, the preceding circumstances raise serious questions concerning the reliability of the identification evidence admitted at defendant’s trial. In Lawson, because the Court of Appeals and trial court relied on the procedures set out in Classen— procedures that we have revised in this opinion—we reverse and remand the case to the trial court for a new trial. Due to the novelty and complexity of the procedures we have articulated today, the parties must be permitted on retrial to (1) supplement the record with any additional evidence that may bear on the reliability of the eyewitness identifications at issue here, and (2) present arguments regarding the appropriate application of the new procedures set out in this opinion.”
Id. Notably, the court did not conduct its own inquiry into the preliminary question whether the identification was barred under OEC 402 because the witness’s perceptions were incapable of supporting an inference of identification or any balancing of probative value against prejudicial effect under OEC 403.
Turning next to James, the court reached a differ-
ent conclusion; the court held that application of the revised framework “could not have resulted in the exclusion of the eyewitness identification evidence.” Id. James involved a midday theft from a grocery store. A store clerk had seen two Native American men—one small and one large— stuffing 40-ounce bottles of Steel Reserve 211 beer into a backpack. He alerted the assistant store manager and pointed out to him the two perpetrators as they were leaving the store. The employees were unable to stop the men, and they drove away in a gray van.
The employees called the police and, when the officer arrived, they gave him a detailed description of the perpetrators, including their race, approximate height and weight, hair styles, and clothing. In that interview, both employees were adamant that they would be able to identify the perpetrators if they saw them again. A few hours later, the officer saw two men who matched the description of the perpetrators of the grocery store theft “exactly”—the defendant (the larger man) and another man—and questioned them. The suspects denied any involvement in the theft but consented to a search of a backpack in their possession, which contained an unopened 40-ounce bottle of Steel Reserve 211 beer. The officer asked the men if they would be willing to go to the grocery store with him to “clear the matter up.” The men consented. The officer handcuffed the men and drove them to the grocery store, where he presented them to the employees for identification. The employees positively identified both suspects as the men who had stolen the beer.
The defendant was charged with multiple offenses. Before trial, he moved to suppress both the out-of-court identification and any in-court identification by the employees. He contended that the identification procedure in the parking lot was unduly suggestive, and, therefore, the identification was inadmissible under Classen. Although the trial court agreed that the identification procedure was suggestive, it denied the motion to suppress on the ground that the identification was made independent of the suggestive procedures.
On review, this court considered, as it had in
Lawson, whether application of the revised test for admissibility of eyewitness identification testimony could result in a different outcome; in James the court concluded that it could not. The court observed that, within minutes of the crime, the employees had given a police officer a detailed and accurate description of the suspects, including information about their race, size, weight, clothing, and the backpack that they were using. Lawson/James, 352 Or at 765-66. The officer apprehended two men a few hours later who matched the employees’ descriptions of the perpetrators “exactly,” and the employees later confirmed that the men whom the officer had apprehended were the men they had seen in the store. Id. at 766.
The court then analyzed the admissibility of that identification under the framework that it had just adopted. First, the court concluded that the OEC 602 requirement of personal knowledge was met. The court stated that, although some facts falling into the category of estimator variables could have negatively affected the witnesses’ perceptions (stress from the assault, for example), on the whole, the estimator variables suggested reliability: both employees were face-to-face with the perpetrators and had a clear opportunity to observe them for a lengthy period of time, they saw the perpetrators in good environmental conditions (inside a lighted store and outside, in broad daylight), and the perpetrators had distinctive features. Id. From that, the court concluded that “no reasonable decisionmaker could find that the witnesses did not have the personal knowledge necessary to identify the perpetrators.” Id.
Because the identification occurred during a “show up” procedure that the trial court had found to be unduly suggestive, however, the state, as the proponent of the identifications, was required under OEC 701 to introduce sufficient evidence to support a finding that the witnesses’ observations were based on their original observations, untainted by the suggestive procedures. Id. at 767. This court noted that the trial court had been satisfied that the suggestive show-up confrontation did not contribute to the witnesses identification of the defendant, because the employees had had a very good look at the perpetrators during the crime and had described them with particularity, including providing details about their features, the clothing that they were wearing, and the unusual-sized bottles of an unusual brand of beer that they had stolen. Id. As this court stated, the witnesses’ accuracy in describing those details demonstrated their reliability, and, therefore, the court held, the trial court did not err in reaching its factual conclusion that the witnesses’ identifications of the defendant were based on their original observations. Id.
The court then turned to the final analytical steps under the new framework: determining whether the witnesses’ identifications were helpful to the trier of fact and whether OEC 403 required their exclusion. Id. On those points, the court acknowledged that an argument could be made that the witnesses’ identifications did not provide the jury with information that was more helpful than their descriptions of the perpetrators, and that, therefore, the persuasive value of the identifications was limited and outweighed by the unfair prejudice introduced by the identifications. Id. But, the court held, the concern of unfair prejudice was negligible. Id. According to the court, because the employees’ descriptions of the perpetrators so closely matched the two men apprehended by the police, their later identification of the defendant could have prejudiced him little, if at all. Id. at 767-68.
To summarize, in Lawson, consideration of both estimator variables and system variables raised serious questions about the reliability of the witness’s identification of the defendant. Because, for that reason, the result could have been different under an application of the revised framework, the court reversed the defendant’s conviction and remanded the case to the trial court for a new suppression hearing and trial. In James, the court noted that facts falling into the category of system variables rendered the identification procedure suggestive, but because the witnesses had had the opportunity to view the perpetrators for a lengthy period of time during the commission of the crime and had been able to give the police officer a detailed and accurate description of the perpetrators immediately after the crimes, the court was able to determine that any error in applying the Classen framework was harmless. That is, even under the new framework, no reasonable decision-maker could have found that the witnesses did not have personal knowledge necessary to identify the perpetrators, the state would have been able to introduce sufficient evidence to support a finding that the witness identifications were based on their original observations, and admission of the identification evidence did not unduly prejudice the defendant.
APPLICATION OF LAWSON/JAMES
As we have stated, the trial court, in relying on Classen, applied what we now view as an incorrect legal standard. On review, we consider whether there was “little likelihood” that application of that incorrect legal standard affected the verdict. Hickman, 355 Or at 749 (applying standard); State v. Davis, 336 Or 19, 32, 77 P3d 1111 (2003) (“Oregon’s constitutional test for affirmance despite error consists of a single inquiry: Is there little likelihood that the particular error affected the verdict?”). Only if we are able to conclude that application of the Lawson/James test could not have resulted in the exclusion of the eyewitness identification evidence will we affirm the ruling of the trial court. See Lawson/James, 352 Or at 765.
The Court of Appeals did so conclude and held that the trial court correctly denied defendant’s motion to dismiss. Haugen, 274 Or App at 150. In so holding, the Court of Appeals first considered whether the record demonstrated that the state introduced sufficient evidence to permit a finding that the eyewitness testimony would not be barred under OEC 402. To that end, the court found, first, that the victim’s identification satisfied the low threshold of relevance under OEC 401. Id. at 146. Second, the Court of Appeals concluded, based on the facts that the victim was able to describe defendant in detail at the police station several days after the attack and that the victim remembered both defendant and Rives talking to him immediately before the assault, that there was “more than enough [evidence] to demonstrate, for purposes of OEC 602, that the victim’s identification was based on personal knowledge.” Id. at 146. Third, the Court of Appeals rejected defendant’s challenges to the identification procedures that Brown had used and concluded that the question whether there was sufficient evidence to permit a finding that the victim’s identification was based on his first-hand perceptions so as to satisfy OEC 701 could reasonably be resolved only in the state’s favor. Id. at 147. In light of those conclusions, the court held, “the victim’s identifications of defendant meet the threshold requirements for admissibility under Lawson/James.” Id. at 150. The Court of Appeals then stated that, therefore,
“defendant can prevail only by showing that the evidence is inadmissible under OEC 403 because its probative value is outweighed by the danger of unfair prejudice.”
Id. And, the court stated, because defendant had made no OEC 403 argument on appeal, OEC 403 provided no basis for excluding the evidence. In light of that holding, the Court of Appeals did not consider whether any of the estimator variables or system variables presented in the case raised serious questions about the reliability, and thus the probative value, of the victim’s identifications.
While we agree with the Court of Appeals that the state met its burden under the Lawson/James test to establish a “minimum baseline of reliability,” Lawson/James, 352 Or at 758, we disagree with the Court of Appeals’ conclusion that OEC 403 provided no basis for excluding the identifications. Although, in his argument to the Court of Appeals, defendant did not cite OEC 403 as the specific basis for remanding the case to the trial court for a new suppression hearing, we cannot agree that defendant failed to preserve an OEC 403 argument on appeal. In the Court of Appeals, defendant asserted that, in Lawson/James, this court had replaced the Classen test for determining the admissibility of eyewitness identification evidence with a new test that requires consideration of system and estimator variables under OEC 602, 701, and 403. He pointed to many system variables and estimator variables that he contended were present in the case, and he argued that those “system and estimator variables reveal that there are questions about the reliability of the identifications sufficient to warrant remand for a new hearing and trial, wherein the trial court can apply the new Lawson/James test.” In other words, defendant argued that the probative value of the evidence should be determined by considering the relative reliability of the eyewitness testimony, given the system variables and the estimator variables, as explained in Lawson/James. He argued that the result of the suppression hearing, and, therefore, the verdict, could have been different if the court had done so, because the identifications were prejudicial and facts falling into the categories of system and estimator variables cast doubt on their reliability and, therefore, their probative value. That, at its core, is an OEC 403 argument.
We turn, then, to consider whether the presence of facts falling into the categories of estimator and system variables raises serious questions about the reliability, and, thus, the probative value, of the eyewitness identifications. If so, then we must consider whether there is little likelihood that the error affected the verdict.
Viewing the facts from defendant’s perspective, a trial court applying Lawson/James could find that a number of estimator variables were at play that negatively affected the victim’s perceptions. To reiterate, at 12:30 a.m., the victim was punched in the side of the head and then kicked in the chest or shoulder by one individual, and, as he tried to get up, a second person struck him in the head. The victim was nearly knocked unconscious and the assailants then fled. A trial court could find the following facts that would support the existence of estimator variables that would raise questions about the reliability of the eyewitness identification. First, the environmental conditions were poor, insofar as it was late at night and dark in the parking lot, and the victim’s encounter with the assailants was very brief. Additionally, the victim told a police officer a few hours after the assault that he had been “blindsided.” That night, he also stated to the officer that he could not recognize his attackers “specifically, individually,” and he could not describe them. On the night of the attack, the victim was not sure about having been punched, and he did not remember being hit in the head or a hammer being used to assault him. Those facts could permit a court to find that the victim was under stress because of the surprise nature of the attack and because he was badly injured.
The trial court in this case expressly discounted those estimator variables in its letter opinion, but, as this court explained in Lawson/James, there are recent scientific studies that show that poor environmental viewing conditions, shorter durations of exposure, witness inattention, and high levels of stress all can impair a witness’s ability to encode memory and undermine the reliability of eyewitness identification. Lawson/James, 352 Or at 744-46.
Similarly, the following facts, viewed in a light favorable to defendant, could support a trial court finding that several system variables were present and raise concerns. Five days after the assault, the victim told Detective Brown that he was attacked by two Vagos members whom he had seen earlier in the bar. He told Brown that he was punched and kicked by a “great big guy,” about 230 pounds, not fat, in his late 20s or early 30s, and that he was hit with a hammer by a “little fat guy,” probably in his 40s, with a long ponytail. Brown informed the victim that he would show him some photos, and he recited a statement to the effect that the perpetrators may or may not be among them. As this court explained in Lawson/James, a photo lineup should consist of pictures of individuals matching the descriptions of the assailants and include among them pictures of subjects known to the officer to be innocent. However, instead of showing the victim pictures of people resembling the “great big guy” or the “little fat guy,” Brown showed the victim 23 photographs of known Vagos motorcycle gang members whom Brown suspected either witnessed or participated in the assault.
A trial court also could find that Brown provided continuous suggestive and confirming feedback during the photo lineup. For example, Detective Brown asked if the bigger man was “pretty buff”; when the victim agreed, Brown told the victim that he thought he knew who he was talking about. Although Brown later testified that he actually had had someone other than defendant in mind, a trial court could find that Brown’s comment was nonetheless suggestive. From the victim’s perspective, Brown had implied that defendant’s image would be among the photos shown to him and had encouraged the victim to identify someone from that collection. Similarly, Brown implied that the “fat little guy” would be among the photos when he told the victim, “I’ll show you some photographs—and maybe that will help us—once we have some photos we can go through—and we’ll identify—who the little fat guy is.” Brown also confirmed the identities and gang names of individuals whom the victim identified and provided additional information about other suspects in the photos, including how their appearances had changed since the photos were taken, sometimes before the victim himself had commented on the photo. As the court stated in Lawson/James, “[w]itness memory can become contaminated by external information or assumptions embedded in questions or otherwise communicated to the witness.” 352 Or at 743. Further, the court stated, “confirming feedback tends to falsely inflate witnesses’ confidence in the accuracy of their identifications, as well as their recollections concerning the quality of their opportunity to view a perpetrator and an event.” Id. at 744.
Finally, Brown subjected the victim to multiple viewings of images of defendant and Rives. After the victim completed his review of the initial 23 photos and had identified defendant and Rives, Brown told the victim that he was going to show him some “more recent photographs because there are some significant differences in both the gentlemen that you have said that have assaulted you.” Based on that conversation, a trial court could find that, in other words, Brown told the victim that he was going to show him some recent photographs of the individuals whom the victim had already identified. Brown then showed the victim a series of surveillance photographs that included images of defendant, Rives, and others whom the victim had identified as having been present during the assault, and asked the victim if he could identify anyone in those photographs. The victim identified defendant, Rives, and others. As the court stated in Lawson/James, “[v]iewing a suspect multiple times throughout an investigation can adversely affect the reliability of any identification that follows that viewing.” 352 Or at 743. That is, even if a witness does not select the proper suspect in an initial identification procedure, the procedure increases the witness’s familiarity with that suspect’s face. When the witness later sees the same suspect or his or her image again, the suspect tends to stand out or appear familiar to the witness. Id. In this case, the victim became more and more certain that defendant was the person who assaulted him as he was shown additional pictures of him. But the trial court could find that, in light of the scientific research that the court discussed in Lawson/ James, the victim’s identification actually became less reliable through multiple viewings. 
In addition, in Lawson/James, the court identified the related problem of “source confusion,” in which a witness may be confused or unable to discern the source of his or her recognition of the suspect. 352 Or at 743. In Lawson, the issue arose because the police suspected that the man whom the victim had seen earlier in the day at the campsite was the perpetrator, and, once that idea was implanted in the victim’s mind, the court stated, “it would have been extremely difficult for [the victim] to mentally separate the task of identifying the perpetrator from her brief glimpse of his profile in the dark from the task of identifying the man she saw earlier in her campsite for about 40 minutes in broad daylight.” Id. at 764. For similar reasons, a trial court could find that that system variable also may be at play in this case. As we have stated, immediately after he was attacked, the victim was unable to identify his attackers and, in fact, stated that he had been blind-sided. A few days later, however, the victim identified defendant and Rives as his attackers, and he was able to describe defendant and Rives in detail and pick out their photos from those that Brown showed him. In the circumstance, however, the fact that the victim was able to identify defendant and Rives in the photos shown to him by Detective Brown does not necessarily increase the reliability of his identification of defendant and Rives as his attackers. That is so, because, like the victim in Lawson, it would have been difficult for the victim in this case to mentally separate the task of identifying the perpetrators of the attack, who had blindsided him during a brief encounter in a dark parking lot, from the task of identifying the people whom he had had ample opportunity to observe over the course of the two hours he had spent in the bar with his friends, and who had spoken to him as he was leaving the bar.
Relatedly, the trial court (and the Court of Appeals) found it compelling that various witnesses “corroborated” the victim’s testimony that his assailant was a member of the Vagos gang, that the assailant was wearing Vagos attire, that the assailant was tall and solidly built, and that the assailant was a car salesman. But it is undisputed that there were no witnesses to the attack. Thus the “corroboration” that the trial court found important to the reliability of the identification was not corroboration of the fact that defendant was the victim’s assailant; rather, the witnesses on whom the trial court relied merely corroborated the victim’s memory that defendant had been at the bar.
In evaluating the admissibility of the eyewitness identification evidence in this case, the trial court relied on Classen, which this court later held was an incorrect legal standard. Defendant, therefore, did not have an opportunity to present evidence or expert testimony at the suppression hearing explaining any of the estimator variables and the system variables that we have just described. And because it applied Classen at the suppression hearing, the trial court did not have an opportunity to consider them.
As the court stated in Lawson/James, even when the state has met its burden to establish a minimum baseline of reliability, “trial courts still must conduct a thorough examination of all the pertinent factors in order to determine the probative value of the evidence under OEC 403.” 352 Or at 758. Where an eyewitness has been exposed to suggestive police procedures, the trial court has a “heightened role as an evidentiary gatekeeper because ‘traditional’ methods of testing reliability—like cross-examination—can be ineffective at discrediting unreliable or inaccurate eyewitness identification evidence.” Id. In this case, however, the court did not have an opportunity to properly exercise its gatekeeping role. As we have stated, facts falling into the categories of estimator variables and system variables in this case raised serious questions about the reliability of the identification evidence admitted at defendant’s trial. It follows that, as in Lawson, the trial court, not having assessed those estimator and system variables, could not have made an informed decision about the admissibility of the eyewitness identification evidence under OEC 403. We conclude, then, that application of the Lawson/James framework could have resulted in the exclusion of the eyewitness identification evidence and, therefore, that the error was not harmless.
We conclude that the identification procedures used in this case raise serious questions about the reliability of the victim’s identifications of defendant under Lawson/James. Remand, therefore, is necessary to give the trial court the opportunity to consider the admissibility of the identifications under the correct standard. On remand, at the evidentiary hearing, the parties must be permitted to supplement the record with any additional evidence that may bear on the reliability of the eyewitness identification at issue and present arguments regarding the appropriate application of the procedures set out in Lawson/James. Id. at 765.
The decision of the Court of Appeals is reversed, and the case is remanded to the circuit court for further proceedings consistent with this decision.
 As the victim later testified, he was “really dazed and * * * really out of it,” and he “was in intense pain” when he was talking to Nicklason.
 As the victim confirmed at trial, on the night of the assault, he did not recall being hit with a hammer. When asked at trial whether, that night, he remembered a ball-peen hammer being used, the victim testified, “No, I didn’t. I was really out of it.”
 Estimator variables include, among other things, whether the witness was under stress at the time of the incident; where the witness’s attention was focused during the incident; the amount of time that the witness spent looking at the perpetrator; the environmental viewing conditions; factors affecting the witness’s visual, physical, and mental acuity; distinctive characteristics of the perpetrator; and the time elapsed between the incident and the identification. Id. at 744-46.
 System variables include, among other things, whether the identification was conducted by a “blind” administrator (a person who does not know the identity of the suspect); whether the witness was instructed prior to the procedure that the suspect may or may not be among those shown; whether the lineup was made up of subjects based on their physical similarity to the witness’s description of the perpetrator; whether the individuals in the lineup or photo array are shown sequentially or simultaneously; whether the witness was shown the suspect multiple times throughout the course of the investigation; whether there was evidence of suggestive questioning or other sources of post-event memory contamination; and whether the administrator provided suggestive or confirming feedback. Id. at 741-44.
 To take one example, research shows that a witness’s confidence or level of certainty (one of the Classen factors) may not be a good indicator of identification accuracy, but it nonetheless has a substantial potential to influence jurors. Id. at 745.
 In Lawson/James, the court described the state’s burden at this point as establishing by a “preponderance of the evidence” that the identification was based on a permissible basis. 352 Or at 755. In Hickman, the court clarified that, under OEC 602, a preponderance of the evidence need only present evidence from which a reasonable juror could find that the eyewitness observed facts necessary to make an identification. 355 Or at 728-30.
 That is, we agree that the victim’s identifications were logically relevant under OEC 401, and that the state adduced sufficient evidence to support a finding that the victim had personal knowledge under OEC 602 and that the identification was based on the perception of the victim under OEC 701.
 We note that the trial court found the victim’s identification reliable in part because the victim became more and more certain that defendant was one of his assailants as he was shown additional pictures of him. The court in Lawson/
James, however, explained that,
“[u]nder most circumstances, witness confidence or certainty is not a good indicator of identification accuracy. Retrospective self-reports of certainty are highly susceptible to suggestive procedures and confirming feedback, a factor that further limits the utility of the certainty variable. Witness certainty, although a poor indicator of identification accuracy in most cases, nevertheless has substantial potential to influence jurors.” 352 Or at 745.