State v. Watson

Full title: State of New Jersey, Plaintiff-Respondent, v. Quintin D. Watson…

Court: Supreme Court of New Jersey

Date published: Aug 2, 2023

254 N.J. 558 (N.J. 2023)

298 A.3d 1049

Opinion

A-23 September Term 2022 087251

08-02-2023

STATE of New Jersey, Plaintiff-Respondent, v. Quintin D. WATSON, Defendant-Appellant.

Lauren S. Michaels, Assistant Deputy Public Defender, argued the cause for appellant (Joseph E. Krakora, Public Defender, attorney; Lauren S. Michaels and Ashley T. Brooks, Assistant Deputy Public Defender, of counsel and on the briefs). Nancy A. Hulett, Assistant Prosecutor, argued the cause for respondent (Yolanda Ciccone, Middlesex County Prosecutor, attorney; Nancy A. Hulett, of counsel and on the briefs). Alexander Shalom argued the cause for amicus curiae American Civil Liberties Union of New Jersey (American Civil Liberties Union of New Jersey Foundation, attorneys; Alexander Shalom, Newark, and Jeanne LoCicero, on the brief). Joseph A. Hayden, Jr., Hackensack, argued the cause for amicus curiae Association of Criminal Defense Lawyers of New Jersey (Pashman Stein Walder Hayden, attorneys; Joseph A. Hayden, Jr., on the brief). Ian S. Marx argued the cause for amicus curiae The Innocence Project (Greenberg Traurig and The Innocence Project, Inc., attorneys; Ian S. Marx, Florham Park, Caroline J. Heller (Greenberg Traurig) a member of the New York bar, admitted pro hac vice, M. Chris Fabricant (The Innocence Project, Inc.) a member of the New York bar, admitted pro hac vice, and Anton Robinson (The Innocence Project, Inc.) a member of the New York and Florida bars, admitted pro hac vice, on the brief). Amanda G. Schwartz, Deputy Attorney General, argued the cause for amicus curiae Attorney General of New Jersey (Matthew J. Platkin, Attorney General, attorney; Amanda G. Schwartz, of counsel and on the brief).

CHIEF JUSTICE RABNER delivered the opinion of the Court.

Lauren S. Michaels, Assistant Deputy Public Defender, argued the cause for appellant (Joseph E. Krakora, Public Defender, attorney; Lauren S. Michaels and Ashley T. Brooks, Assistant Deputy Public Defender, of counsel and on the briefs).

Nancy A. Hulett, Assistant Prosecutor, argued the cause for respondent (Yolanda Ciccone, Middlesex County Prosecutor, attorney; Nancy A. Hulett, of counsel and on the briefs).

Alexander Shalom argued the cause for amicus curiae American Civil Liberties Union of New Jersey (American Civil Liberties Union of New Jersey Foundation, attorneys; Alexander Shalom, Newark, and Jeanne LoCicero, on the brief).

Joseph A. Hayden, Jr., Hackensack, argued the cause for amicus curiae Association of Criminal Defense Lawyers of New Jersey (Pashman Stein Walder Hayden, attorneys; Joseph A. Hayden, Jr., on the brief).

Ian S. Marx argued the cause for amicus curiae The Innocence Project (Greenberg Traurig and The Innocence Project, Inc., attorneys; Ian S. Marx, Florham Park, Caroline J. Heller (Greenberg Traurig) a member of the New York bar, admitted pro hac vice, M. Chris Fabricant (The Innocence Project, Inc.) a member of the New York bar, admitted pro hac vice, and Anton Robinson (The Innocence Project, Inc.) a member of the New York and Florida bars, admitted pro hac vice, on the brief).

Amanda G. Schwartz, Deputy Attorney General, argued the cause for amicus curiae Attorney General of New Jersey (Matthew J. Platkin, Attorney General, attorney; Amanda G. Schwartz, of counsel and on the brief).

CHIEF JUSTICE RABNER delivered the opinion of the Court.

Defendant Quintin Watson was convicted of bank robbery after a jury trial. The evidence against him included testimony from a teller, who identified defendant for the first time in court, and from the lead detective, who narrated a bank surveillance video for the jury. This appeal raises two principal issues: (1) the propriety of first-time in-court identifications, and (2) the extent to which investigators may narrate video recordings.

In the case of first-time in-court identifications, witnesses are asked if they can identify their culprit long after the crime took place. They see a single individual seated at the defense table beside a lawyer. And it is evident the prosecution team believes that person is the culprit. The inherently suggestive nature of the procedure, conducted in front of a jury, evades well-settled protections that the law provides. It therefore risks depriving defendants of their due process rights. To avoid unduly suggestive identifications in court that can trigger serious due process concerns, we hold that first-time in-court identifications may only be conducted when there is good reason for them. In addition, before attempting to conduct a first-time in-court identification, the State must give fair notice to the defense. As to narration evidence, an investigator who has carefully reviewed a video recording can satisfy the "perception" prongs under N.J.R.E. 701 and 602 and the "helpfulness" prong under N.J.R.E. 701 and offer lay witness testimony. By drawing attention to key details that a jury might otherwise overlook -- as is the case, for example, with a potentially confusing, complex, or unclear recording -- such evidence can be particularly helpful.

Although an investigator's specific comments can assist a jury in determining facts in issue, the rules of evidence do not allow for continuous, running commentary on video evidence by someone who has merely studied a recording. We therefore identify certain safeguards to underscore the limited use of narration evidence: investigators should give focused responses to specific questions; they can provide objective, factual comments but not subjective interpretations; they may not comment on facts that are reasonably in dispute, which should be left for the jury to decide; and they should not offer testimony based on inferences drawn from other evidence. In addition, a party that intends to present narration evidence should provide opposing counsel a written summary of the proposed testimony before trial. The parties can then ask the court to address any disputed areas at a Rule 104 hearing.

We also briefly address a question related to the Confrontation Clause.

The evidence before the jury in this case ran afoul in all three areas. Based on the identification evidence alone, defendant's conviction cannot stand. We therefore reverse the judgment of the Appellate Division, which upheld the conviction, and remand for a new trial.

I.

A.

Shortly before noon on January 14, 2017, an individual robbed a bank in North Brunswick. A single teller and the branch manager were on duty when the robber entered the bank at about 11:53 a.m. He wore a baseball cap and gloves. The robber opened the bank's exterior and interior doors with his gloved right hand, removed his gloves, and approached the teller window. He then put a note on the counter that read, "everything now."

The teller first emptied the top drawer at his station. At the robber's direction, he removed cash from the bottom drawer as well. The robber took the cash -- which totaled $5,772 -- and left the bank. He placed his bare left hand on both interior door handles and then pushed open the exterior door with his left elbow.

Bank surveillance footage captured the entire robbery. It lasted all of 57 seconds. Neither the manager nor any of the few customers in the bank at the time realized what had happened.

The North Brunswick Police Department responded soon after the robbery. The teller told the police the suspect was Black, muscular, around six-foot-two or six-foot-three, and had an accent. Sergeant Frank Vitelli, Jr., the lead detective, checked the area for possible DNA evidence but found none. He lifted seven fingerprints from the door handles and counter, but none matched any fingerprints in the State Police database. Sergeant Vitelli also sent a bulletin known as a TRAKs message to other law enforcement agencies. It contained information about the robbery and a still photo of the suspect from the bank surveillance video.

Nine months later, in October 2017, J.H. -- whom we refer to as Joan -- read a newspaper article and saw an accompanying photo of a man wanted in connection with a different bank robbery in Princeton. She recognized the person in the photo as defendant Quintin Watson, her former boyfriend. Joan had dated and lived with defendant for four years. They broke up in 2012 when defendant told Joan he was leaving her for another woman who was pregnant. Defendant later married the other person.

Joan contacted the police and ended up speaking with officers about two other bank robberies in Somerset County. Based on Joan's identification and other evidence, defendant was charged in three robberies in November 2017: the one in Princeton and two in Somerset County.

According to the State, Somerset County officials contacted the North Brunswick Police Department that same month and suggested they investigate defendant for the January 2017 robbery in North Brunswick -- the subject of this appeal. After viewing surveillance footage from the Somerset robberies, the North Brunswick Police Department charged defendant in November 2017 with the bank robbery in North Brunswick. A grand jury in Middlesex County later indicted him for the robbery, contrary to N.J.S.A. 2C:15-1.

On September 25, 2018, 20 months after the robbery, a detective with the Middlesex County Prosecutor's Office showed the teller six photos, one at a time, and asked if he could identify the person who robbed the bank. The teller picked a photo of someone other than defendant Quintin Watson and said at trial that he was 75 to 90 percent sure of the identification. On cross-examination, he said he had been 85 percent sure.

The next month, October 2018, the Prosecutor's Office showed Joan a single still photo taken from the bank surveillance video. She said she was 100 percent positive the photo depicted defendant.

The State filed a pretrial motion to admit evidence of the other three robberies at trial. The trial court denied the motion and provided guidance about what Joan could say about her initial contact with the police. In addition, the trial court explained it would allow limited, sanitized testimony from a North Brunswick police officer about prior contacts he had with another law enforcement agency.

A two-day trial was conducted in November 2018 -- 22 months after the robbery. We focus on three points relevant to this appeal. B.

1.

First, the prosecutor asked the teller if he could identify the robber in court. The teller identified defendant, who was seated in between his lawyers at counsel table. The teller said he was "maybe like ... 80 percent" sure. The prosecution did not provide advance notice of the in-court identification, and defense counsel did not object to it.

During cross-examination, the teller revealed that he had met with the prosecutor prior to trial. This exchange followed:

DEFENSE COUNSEL: And [the prosecutor] informed you what was going to happen today, right?

TELLER: Yes.

DEFENSE COUNSEL: And he informed you that the individual who was accused of committing this robbery is in court seated at the defense table, right?

TELLER: Yes, he did.

The State asked no follow-up questions on that point.

Joan also testified at trial. She was shown two still photos from the bank surveillance video and testified she was 100 percent positive that each depicted defendant. She later confirmed that the top 20 to 25 percent of the robber's face was not visible in the photo she had viewed before and at trial. Because the suspect wore a hat pulled down over his eyes, it was not possible to see the person's eye color or whether he had hair.

2.

Second, Sergeant Vitelli testified. At the outset, he described his training and experience and then testified about the investigation. He explained that he interviewed witnesses, checked for DNA evidence, dusted for fingerprints, took photos, and watched surveillance footage taken from the bank and a convenience store nearby. He testified that he watched the bank's surveillance videos "numerous times." Over the objection of defense counsel, Sergeant Vitelli narrated the bank surveillance video. The prosecutor asked a series of questions while the video was played for the jury. They ranged from general inquiries -- "What do you see?" -- to specific ones -- "With what [did he open the door]?" The more open-ended questions invited and led to more open-ended narrative responses.

Among other things, the officer testified about how and where the suspect entered the bank, that he wore gloves when he entered and then removed the glove on his right hand, that he placed the demand note on the teller's counter and held it there with two fingers, that he used his left elbow to open the door to leave the bank, and that he appeared to run when he left the bank. At one point, Sergeant Vitelli offered his opinion that "the suspect was very careful" in the way he acted, "attempting [not] to leave any type of evidence behind." The trial court sustained defense counsel's objection to the comment but did not strike the answer or instruct the jury not to consider it.

Sergeant Vitelli also testified about several still photos from the bank surveillance footage and responded "yes" when asked whether the note was "a barrier between [the suspect's fingertips] and the" counter. In addition, the officer briefly narrated the convenience store video. He testified it showed someone walking toward the bank and, soon after, jogging back from that direction.

3.

Third, Sergeant Vitelli provided limited testimony about the history of the investigation and how the North Brunswick Police Department learned about defendant. The officer explained that, after the robbery, he sent a TRAKs message to other law enforcement agencies in the hope of getting relevant information in response. The following exchange then took place:

PROSECUTOR: In November 2017, were you contacted by another law enforcement agency regarding Quintin Watson?

SGT. VITELLI: Yes, I was.

PROSECUTOR: And at some point did you consult with that law enforcement agency and after which criminal complaints were signed against Mr. Watson?

SGT. VITELLI: Yes, they were.

C.

As part of its final instructions to the jury, the trial court read from the model jury charge on in-court and out-of-court identifications. See Model Jury Charges (Criminal), "Identification: In-Court and Out-of-Court Identifications" (rev. May 18, 2020). The jury found defendant guilty of the single charge of robbery. The court later sentenced him to an extended fifteen-year term of imprisonment and ordered defendant to pay $5,772 in restitution.

The Appellate Division affirmed defendant's conviction in a comprehensive opinion. State v. Watson, 472 N.J. Super. 381, 404, 277 A.3d 39 (App. Div. 2022). We summarize its findings on the three issues central to this appeal.

First, the appellate court concluded the trial court erred when it allowed Sergeant Vitelli "to testify that he had been contacted by and consulted with another police department just before filing" charges against defendant. Ibid. In the Appellate Division's view, the testimony violated defendant's Confrontation Clause rights under the Sixth Amendment because it created an "inescapable inference" that the other "department possessed incriminating evidence that was not presented to the jury." Id. at 404, 435, 277 A.3d 39. However, the appellate court found the error harmless beyond a reasonable doubt. Id. at 404-05, 445, 277 A.3d 39.

Second, the Appellate Division concluded the trial court did not abuse its discretion when it allowed Sergeant Vitelli to narrate the bank surveillance video. Id. at 405, 470, 277 A.3d 39. The appellate court declined to adopt a rule that would bar a witness "from pointing out an event or circumstance depicted in [a] video or from otherwise offering a description of or comment on any such event or circumstance." Id. at 459, 277 A.3d 39. Whether "to allow a witness to describe and highlight something on the screen that the jury [can] see for itself," the court explained, must be decided "on a case-by-case if not comment-by-comment basis." Ibid.

Focusing on N.J.R.E. 701, the appellate court found the rule's personal-knowledge prong was satisfied when a "police witness reviewed the surveillance video before trial"; the witness need not have seen the actual crime in real time. Id. at 463, 277 A.3d 39.

To help determine whether the rule's helpfulness prong is met in individual cases, the court outlined six factors: (1) whether the proposed comment provides background context about the video such as the location, angle, date, and time of the recording; (2) whether the comment explains "the length of the video" or "pertains to specific events or circumstances" using "neutral language"; (3) whether the comment "pertains to a fact in dispute"; (4) whether the "comment is based on an inference or deduction drawn from or supported by other facts in evidence, or from the officer's training and experience"; (5) "[t]he clarity and resolution of the video"; and (6) "the complexity of the video ... and the amount of visual information being displayed simultaneously." Id. at 464-69, 277 A.3d 39.

Applying a deferential standard of review, the Appellate Division upheld the decision to allow the narration testimony. Id. at 469-70, 277 A.3d 39. The court also directed that, in future cases, the State should move in advance to introduce video narration testimony at a Rule 104 hearing. Id. at 471-73, 277 A.3d 39. In addition, the court recommended that juries be told "they are free to reject any lay opinion testimony provided by the narrator." Id. at 474-75, 277 A.3d 39.

Third, the Appellate Division concluded it was not error to allow the bank teller to identify defendant for the first time in the courtroom. Id. at 476, 277 A.3d 39. The court declined to impose a new per se rule forbidding first-time in-court identifications. Id. at 475-76, 277 A.3d 39. Although it acknowledged the inherent suggestiveness and potential unreliability of those identifications, the appellate court stated that cross-examination and jury instructions could address those concerns. Id. at 501-02, 277 A.3d 39. We granted defendant's petition for certification limited to the three issues discussed above. 252 N.J. 598, 288 A.3d 442 (2022). We also granted leave to appear as friends of the court to the Attorney General, the American Civil Liberties Union of New Jersey (ACLU), the Association of Criminal Defense Lawyers of New Jersey (ACDL), and the Innocence Project.

II.

Defendant submits that the Appellate Division properly found a Confrontation Clause violation but incorrectly concluded the error was harmless. He also argues that Sergeant Vitelli's narration testimony was inadmissible under N.J.R.E. 701. He contends that viewing a video does not satisfy the rule's perception requirement and cannot assist the jury. In addition, defendant suggests that when lay opinion testimony by a narrator is proffered, courts should require notice and a pretrial hearing, and if the evidence is allowed, the jury should be instructed on how to consider it. Finally, defendant argues that first-time in-court identifications should be barred because they are inherently unreliable and unduly prejudicial.

Several amici present arguments that align with defendant's position. The ACLU and ACDL stress that lay opinion testimony must be based on firsthand perception, that the Appellate Division took an overly expansive view of N.J.R.E. 701, and that the Court should adopt clear procedural protections.

The Innocence Project submits that first-time in-court identifications implicate due process protections under the Federal and State Constitutions. The ACLU as well as the Innocence Project argue that first-time in-court identifications are extremely suggestive and unreliable, and should be barred.

The Middlesex County Prosecutor's Office, which prosecuted the case, and the Attorney General, as amicus, contend that Sergeant Vitelli's narration testimony was properly admitted. The Attorney General explains that police narration based on an officer's repeated viewing of video footage satisfies Rule 701 ’s perception requirement. In addition, the Attorney General submits that the Appellate Division's non-exhaustive list of six factors can guide courts as they decide whether proposed testimony will be helpful to the jury. Both the Prosecutor's Office and the Attorney General argue the officer's testimony in this case was proper. The Attorney General stresses that the testimony appropriately highlighted details that were not immediately apparent and did not invade the province of the jury. Both also endorse the use of Rule 104 hearings to determine the scope of narration testimony.

The Prosecutor's Office and the Attorney General also argue against a per se rule that would preclude first-time in-court identifications. They contend that concerns about suggestiveness and unreliability can be addressed on a case-by-case basis through cross-examination and appropriate jury instructions.

The Prosecutor's Office additionally argues that the Appellate Division improperly found a Confrontation Clause violation but correctly concluded the error was harmless.

III. First-Time In-Court Identifications

We begin with the in-court identification of defendant. In State v. Henderson, this Court addressed various issues related to the admissibility of eyewitness identification evidence. 208 N.J. 208, 27 A.3d 872 (2011). Based on an extensive record developed at an evidentiary hearing, the Court observed "that the possibility of mistaken identification is real" and "that eyewitness misidentification is the leading cause of wrongful convictions across the country." Id. at 218, 27 A.3d 872.

In light of the scientific evidence, the Court concluded "that memory is malleable, and that an array of variables can affect and dilute memory and lead to misidentifications." Ibid. The opinion detailed a series of variables and their possible impact on the reliability of identification evidence. Id. at 248-72, 27 A.3d 872. Among other factors, Henderson addressed "showups" -- "single-person lineups" in which "a single suspect is presented to a witness to make an identification," id. at 259, 27 A.3d 872 -- and reaffirmed that they are "inherently suggestive," id. at 261, 27 A.3d 872 (citing State v. Herrera, 187 N.J. 493, 504, 902 A.2d 177 (2006) ).

In the end, the Court revised the standard to assess whether eyewitness identification can be admitted in individual cases. Id. at 218-19, 27 A.3d 872. Henderson also called for the development of jury charges to help jurors better understand relevant factors related to reliability and weigh them properly. Id. at 219, 27 A.3d 872.

A.

Under the current legal framework for out-of-court identifications, defendants are entitled to a pretrial hearing if they can "show[ ] some evidence of suggestiveness that could lead to a mistaken identification." Id. at 288, 27 A.3d 872. "That evidence, in general, must be tied to a system ... variable" -- a "variable[ ] within the State's control." Id. at 248, 288-89, 27 A.3d 872. Next, "the State must ... offer proof to show that the proffered eyewitness identification is reliable[,] accounting for" variables within its control as well as those that are not. Id. at 289, 27 A.3d 872. The latter factors are known as "estimator variables." Id. at 261, 27 A.3d 872. Defendants have "the ultimate burden ... to prove a very substantial likelihood of irreparable misidentification." Id. at 289, 27 A.3d 872.

After weighing all the evidence, if the trial court finds the defendant has carried that burden, it "should suppress the identification evidence." Ibid. If the evidence is allowed before the jury, "the court should provide appropriate, tailored jury instructions" keyed to variables that are relevant to the case. Ibid.

B.

Henderson involved a suggestive pretrial out-of-court identification. Id. at 222-24, 27 A.3d 872. The opinion did not address in-court identifications. Certain factors discussed in the opinion, however, are directly relevant to first-time in-court identifications.¹

1 We use the phrase "first-time in-court identification" to refer to in-court identifications that are not preceded by a successful out-of-court identification. To be clear, if the discovery supplied to the defense reflects that a witness was unable to identify anyone or identified someone other than the defendant before trial, and the witness is then asked to identify the defendant in court, we consider the later attempt a "first-time in-court identification."

An in-court identification is essentially a live, single-person line-up in a courtroom. It is comparable to a showup but is conducted well after the crime has taken place. Settled case law, scientific studies, and common sense underscore the problems with first-time in-court identifications.

In 2006, this Court observed in Herrera that "showups are inherently suggestive ... because the victim can only choose from one person," who is generally in police custody. 187 N.J. at 504, 902 A.2d 177. But a showup, standing alone, is not considered impermissibly suggestive when it is conducted "on or near the scene ... ‘before memory has faded.’ " Ibid. (quoting State v. Wilkerson, 60 N.J. 452, 462, 291 A.2d 8 (1972) ).

Showups can serve a valuable purpose if conducted within hours of a crime. They may help the police confirm whether they "have captured the perpetrator or whether the perpetrator is still at large," while a witness's memory is still fresh. Commonwealth v. Crayton, 470 Mass. 228, 21 N.E.3d 157, 165 (2014). Some states tailor the admissibility of showup identifications to such situations. Henderson, 208 N.J. at 259-60, 27 A.3d 872. Massachusetts, for example, requires that there be "good reason for the use of a showup." Commonwealth v. Martin, 447 Mass. 274, 850 N.E.2d 555, 562-63 (2006).

Henderson reaffirmed the highly suggestive nature of showups and also found they "present a heightened risk of misidentification" if "conducted more than two hours after an event." 208 N.J. at 261, 27 A.3d 872. In reaching that conclusion, the Court relied on a number of scientific studies. See, e.g., A. Daniel Yarmey et al., Accuracy of Eyewitness Identifications in Showups and Lineups, 20 Law & Hum. Behav. 459, 464 (1996); Jennifer E. Dysart et al., Show-ups: The Critical Issue of Clothing Bias, 20 Applied Cognitive Psych. 1009, 1019 (2006).

The Court observed in Herrera "that only a little more is required in a showup to tip the scale toward impermissibly suggestive." 187 N.J. at 504, 902 A.2d 177. Identifications made in court for the first time present a good deal more to be concerned about. As the Court observed three decades ago in State v. Madison, in-court identifications are "extremely suggestive" when "the only person sitting at the defense table who reasonably could [be] the defendant" is the accused. 109 N.J. 223, 243, 536 A.2d 254 (1988).

Most anyone who has watched a trial on television knows where the defendant sits in a courtroom. Beyond that, compared to a showup, "the witness is given an even stronger impression that the authorities are already satisfied that they have the right man." Ibid. (quoting Wayne R. LaFave & Jerold H. Israel, Criminal Procedure, § 7.4 at 341-42 (1985)). So "[i]f a one-on-one confrontation at the police station is highly suggestive, then surely such a confrontation in court is the most suggestive situation of all." Ibid. (quoting LaFave & Israel, § 7.4 at 341). Plus in-court identifications are conducted in the presence of a judge, lending the court's imprimatur to the procedure.

A number of scholarly sources agree that first-time in-court identifications are highly suggestive and unreliable. The National Academy of Sciences in 2014 observed that in-court identifications "do not reliably test an eyewitness’ memory" because "[i]n the courtroom, the eyewitness can easily see where the defendant is sitting." Nat'l Rsch. Council of the Nat'l Acad. of Scis., Identifying the Culprit: Assessing Eyewitness Identification 36 n.28 (2014). The Academy also found that "in-court eyewitness identifications can influence juries in ways that cross-examination, expert testimony, or jury instructions are unable to counter effectively." Id. at 110.

For those and other reasons, the Academy concluded that "[a]n identification ... typically should not occur for the first time in the courtroom." Ibid. The Academy recommended that "[i]f no identification procedure was conducted during the investigation, a judge should consider ordering that an identification procedure be conducted before trial." Ibid.

In 2019, the Third Circuit Task Force on Eyewitness Identifications noted that "the extreme suggestivity of a defendant sitting at counsel table with defense counsel should, by itself, raise caution flags regarding the independent reliability of an in-court identification. Yet, jurors may not understand this point." 2019 Report of the Third Circuit Task Force on Eyewitness Identification 59 (2019); see also Aliza B. Kaplan & Janis C. Puracal, Who Could It Be Now? Challenging the Reliability of First Time In-Court Identifications After State v. Henderson and State v. Lawson, 105 J. Crim. L. & Criminology 947, 990 (2015) (concluding that first-time in-court identifications have "all the suggestiveness of ... show-up[s]" and should be banned).

Fading memories compound the problem. As we noted in Henderson, "[m]emory decay ‘is irreversible’; memories never improve" with time. 208 N.J. at 267, 27 A.3d 872. So "the more time that passes, the greater the possibility that a witness’ memory of a perpetrator will weaken." Ibid. (citing Carol Krafka & Steven Penrod, Reinstatement of Context in a Field Experiment on Eyewitness Identification, 49 J. Personality & Soc. Psych. 58, 65 (1985)); see also Kenneth A. Deffenbacher et al., Forgetting the Once-Seen Face: Estimating the Strength of an Eyewitness's Memory Representation, 14 J. Experimental Psych.: Applied 139, 142 (2008); Yarmey et al., 20 Law & Hum. Behav. at 464. In-court identifications at trial, of course, invariably occur months if not years after the crime was committed. C.

The Attorney General concedes that in-court identification procedures, guided by prosecutors who are state actors, should be analyzed through the lens of due process. As this Court has recognized, "suggestive police procedures may ‘so irreparably "taint[ ]" ... out-of-court and in-court identifications’ that a defendant is denied due process." Henderson, 208 N.J. at 285, 27 A.3d 872 (alteration in original) (quoting Madison, 109 N.J. at 239, 536 A.2d 254 ); see also Neil v. Biggers, 409 U.S. 188, 198, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972) ("It is the likelihood of misidentification which violates a defendant's right to due process ...."); State v. Emanuele, 51 F.3d 1123, 1128 (3d Cir. 1995) ("A government identification procedure violates due process when it is ‘unnecessarily suggestive’ and creates a ‘substantial risk of misidentification.’ " (quoting Gov't of Virgin Islands v. Riley, 973 F.2d 224, 228 (3d Cir. 1992) )).

Other courts have also found that in-court identification procedures implicate the due process rights of defendants. See State v. Dickson, 322 Conn. 410, 141 A.3d 810, 824, 828 (2016) ("[F]irst time in-court identifications, like in-court identifications that are tainted by an unduly suggestive out-of-court identification, implicate due process protections."); see also United States v. Greene, 704 F.3d 298, 305-07 (4th Cir. 2013) ; United States v. Rogers, 126 F.3d 655, 657-58 (5th Cir. 1997) ; United States v. Morgan, 248 F. Supp. 3d 208, 212-13 (D.D.C. 2017) ; City of Billings v. Nolan, 385 Mont. 190, 383 P.3d 219, 224-25 (2016).

Suggestive behavior by private actors, as opposed to suggestive procedures arranged by the police, "does not implicate due process concerns." State v. Chen, 208 N.J. 307, 317-18, 27 A.3d 930 (2011). But prosecutors who investigate and then attempt to prove criminal allegations in court are not private actors. "[A] prosecutor's conduct in court constitutes state action." Dickson, 141 A.3d at 828 ; see also Morgan, 248 F.Supp.3d at 213. Perry v. New Hampshire, 565 U.S. 228, 132 S.Ct. 716, 181 L.Ed.2d 694 (2012), did not hold otherwise. It addressed suggestive pretrial circumstances that were not arranged by law enforcement officers and explained "that due process concerns arise only when law enforcement officers use an identification procedure that is both suggestive and unnecessary." Id. at 234, 238-39, 248, 132 S.Ct. 716. The question whether suggestive in-court identifications carried out by a prosecutor could pose due process concerns was not before the Court, Dickson, 141 A.3d at 828 ; see also Perry, 565 U.S. at 244, 247, 132 S.Ct. 716, and we disagree with courts that have interpreted Perry differently, see, e.g., United States v. Whatley, 719 F.3d 1206, 1216 (11th Cir. 2013) ; Garner v. People, 436 P.3d 1107, 1119 (Colo. 2019).

D.

Other jurisdictions have addressed the question of first-time in-court identifications. In Crayton, the Supreme Judicial Court of Massachusetts held that when "an eyewitness has not participated before trial in an identification procedure," an in-court identification by the witness should be treated "as an in-court showup" and allowed "only where there is ‘good reason’ for its admission." 21 N.E.3d at 169 (emphasis added).

The court explained that "good reason" in this setting can include several situations, such as when a witness is already "familiar with the defendant" from before the crime, or when "the witness is an arresting officer who" saw the crime and "the identification merely confirms that the defendant is the person who was arrested for the" offense. Id. at 170. In both circumstances, the court reasoned, the identifications are confirmatory. As a result, "there is little risk of misidentification" based on the suggestiveness of the in-court showup because the witness is not relying "solely on his or her memory of witnessing the defendant at the time of the crime." Ibid. The Crayton court added that other good reasons that would generally justify out-of-court showups -- "concerns for public safety"; the need for a prompt investigation right after the crime; and the benefits of a "prompt confirmation" -- would "never justify an in-court" identification. Ibid. Those justifications, the court explained, "depend[ ] on the short duration of time between the crime and the showup." Ibid.

In a companion case, Commonwealth v. Collins, the Supreme Judicial Court imposed tighter restrictions when a witness previously "failed to make a positive identification." 470 Mass. 255, 21 N.E.3d 528, 536 (2014). In those circumstances, the court held that an in-court identification would usually have to be justified by a showing that it "is more reliable than the witness's earlier failure." Ibid.

Crayton requires the prosecution to move in limine to admit in-court identifications. 21 N.E.3d at 171. The court encouraged prosecutors to file motions before trial so that there is time to conduct less suggestive out-of-court identifications in the event the in-court identification is suppressed. Ibid.

The Connecticut Supreme Court takes an even more restrictive approach to first-time in-court identifications. In Dickson, the court held that they may be conducted "only if ... there is no factual dispute as to the identity of the perpetrator, or the ability of the particular eyewitness to identify the defendant is not at issue." 141 A.3d at 835-36. First-time in-court identifications also require advance permission from the trial court. Id. at 835. And if an eyewitness cannot identify the defendant in a nonsuggestive pretrial procedure, "a one-on-one in-court identification should not be allowed." Id. at 836-37.

Other courts do not accord special treatment to first-time in-court identifications. See, e.g., State v. Doolin, 942 N.W.2d 500, 512 (Iowa 2020) ; Garner, 436 P.3d at 1119 ; United States v. Thomas, 849 F.3d 906, 910-11 (10th Cir. 2017) ; Whatley, 719 F.3d at 1216-17 ; Byrd v. State, 25 A.3d 761, 767 (Del. 2011) ; United States v. Recendiz, 557 F.3d 511, 525 (7th Cir. 2009) ; State v. King, 156 N.H. 371, 934 A.2d 556, 561 (2007).

E.

The approach outlined in Henderson was meant to "both protect the rights of defendants, by minimizing the risk of misidentification, and enable the State to introduce vital evidence." 208 N.J. at 219-20, 27 A.3d 872. We consider in-court identifications with the same concerns in mind. Asking witnesses long after a crime was committed if they can identify the culprit -- when the only person at counsel table who could reasonably be the defendant would be obvious to the witness, and when it is evident the prosecution team believes the person is the culprit -- presents an even greater risk of misidentification than an out-of-court showup.

The approach hardly resembles the appropriate pretrial standard -- having a person who does not know the identity of the suspect conduct the identification procedure after first giving neutral instructions to the witness. See id. at 248-53, 289-91, 27 A.3d 872. Plus, given the setting, confirmatory feedback is built into the prosecutor's question. See id. at 253-55, 27 A.3d 872. And the lengthy delay between the time of the offense and the identification inevitably allows for memory to decay. See id. at 267, 292, 27 A.3d 872.

The concerns outlined in Henderson therefore apply with even greater force to first-time in-court identifications than they do to showups. Yet first-time in-court identifications are not currently subject to advance scrutiny. They simply unfold in front of a jury.

Suppose the State were to conduct the following identification procedure months after a crime but before trial: the prosecution first informs a witness that it has located and identified the culprit; it next singles out or otherwise highlights a particular person for the witness to view; and it then asks if the witness can make an identification. To be sure, such a highly suggestive process could readily pose "a very substantial likelihood of irreparable misidentification." Id. at 289, 27 A.3d 872 (citing Manson v. Brathwaite, 432 U.S. 98, 116, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977) ); see also Crayton, 21 N.E.3d at 167. The process also mirrors a first-time in-court identification.

Courts should not sanction such highly suggestive procedures -- not in an individual case and not as a categorical approach to the introduction of certain evidence. Indeed, it is hard to see how the court system can justify overseeing the very type of identification procedure it would likely criticize law enforcement officers for conducting. Courts should instead take steps to guard against practices that pose serious due process concerns -- especially inside a court of law in front of a jury. See Dickson, 141 A.3d at 824 n.11 (imposing a prophylactic constitutional rule to prevent the risk of due process violations in the context of first-time in-court identifications).

There are existing protections against unconstitutional eyewitness identifications. Law enforcement can follow federal and state authority about how to conduct identification procedures ahead of trial. Nothing in today's ruling prevents the State from doing so. Defendants can then challenge an identification at a pretrial hearing and try to prevent the jury from learning about potentially tainted evidence. See Henderson, 208 N.J. at 288-89, 27 A.3d 872.

It is not appropriate to bypass those protections if, for whatever reason, the State has not gathered identification evidence before trial. Otherwise, witnesses who could not identify the defendant in a fair, non-suggestive pretrial proceeding would be allowed to make an identification in court because suggestive cues directed them to the person on trial. See Dickson, 141 A.3d at 824 n.11. By conducting a suggestive identification procedure in a courtroom, the State may implicate due process concerns and deprive defendants of their due process rights in a way that neither cross-examination nor jury instructions can adequately address.

F.

To avoid unduly suggestive identifications of defendants in court that may trigger serious due process concerns under the State Constitution, we draw on the standard in Crayton and hold as follows: first-time in-court identifications can be conducted only when there is "good reason" for them. See Crayton, 21 N.E.3d at 169.

As the Supreme Judicial Court of Massachusetts observed, traditional "good reasons" for out-of-court showups no longer apply at trial. Id. at 170. By the time a criminal trial begins, the original investigation is long over, a suspect has been apprehended, and prompt confirmation is no longer needed to help ensure public safety. Ibid.

But other good reasons may exist, such as when an "eyewitness was familiar with the defendant before" the crime. Ibid. Victims of domestic violence, for example, could properly be allowed to identify their assailant in court for the first time. Ibid. Friends or associates, among others, could identify someone they have known for some time. And officers could confirm that the person on trial was the same person they arrested. Ibid.

The better practice, however, is for the State to conduct appropriate identification procedures before trial. If it does not, there may well be no good reason to allow an in-court identification for the first time. And if a witness fails to make a positive identification at an earlier procedure, the State must show that the proposed, upcoming in-court identification would be more reliable and would "pose[ ] little risk of misidentification despite its suggestiveness." Collins, 21 N.E.3d at 536-37. To meet that burden, it is not enough to rely on the difference between still photos shown pretrial and the live presence of an individual in court.

To ensure fair and orderly proceedings, we also require the following practices going forward for proposed first-time in-court identifications. First, as with suggestive out-of-court identifications, defendants are entitled to advance notice and an opportunity to challenge in-court identification evidence before trial. With that in mind, the State must file a motion in limine if it intends to conduct a first-time in-court identification procedure. Although defendants are typically required to file motions to suppress, that approach does not make sense when only the prosecution knows whether it will ask a witness to make an identification in court. See Crayton, 21 N.E.3d at 170. At the hearing, the parties and the court should explore whether good reason exists to conduct a first-time in-court identification.

When a witness has identified the defendant at a pretrial identification procedure, and the State has provided notice consistent with Rules 3:11 and 3:13-3(b)(1)(J), no additional notice is required to conduct an in-court identification, and no hearing is necessarily called for.

Second, just as law enforcement officers are required to make a record of an out-of-court identification under Rule 3:11, prosecutors must disclose in writing anything discussed with a witness during trial preparation that relates to an upcoming in-court identification. For example, if a prosecutor or law enforcement officer tells a witness that the defendant will be in court, or describes where the defendant will be seated, that information must be revealed to the defense. Similarly, if witnesses during trial preparation are shown photos they had previously viewed at prior out-of-court identifications, that must be disclosed as well. Cf. Henderson, 208 N.J. at 255-56, 27 A.3d 872 (discussing the effect of multiple viewings of a suspect).

Third, if a hearing is needed to determine admissibility, it should be conducted and resolved before the start of trial.

We rely on the Court's supervisory authority under Article VI, Section 2, Paragraph 3 of the New Jersey Constitution to require the above practices. We also call on the Criminal Practice Committee to consider amendments to the court rules, including Rules 3:11 and 3:13(b)(1)(J).²

2 We do not address the Innocence Project's argument that eyewitnesses should only be allowed "to testify up to the degree of certainty they had when they made their out-of-court identification." The teller did not identify defendant prior to trial in this case, and no party has raised the issue. See State v. Mackroy-Davis, 251 N.J. 217, 238 n.7, 276 A.3d 1099 (2022).

We apply the above standard here and provide clearer rules going forward. Today's holding applies to this and future cases, and to State v. Roberson Burney, 255 N.J. 1, 298 A.3d 1080 (2023), filed today. See Henderson, 208 N.J. at 302, 27 A.3d 872.

G.

We turn now to the in-court identification in this appeal.

Twenty months after the robbery, the bank teller reviewed an array of six photos that included defendant's picture. The teller identified a different person and said he was 75 to 90 percent sure that person was the robber. The State properly disclosed the misidentification. It did not ask the teller to attempt another out-of-court identification, which might have raised concerns about multiple viewings of the same suspect. See Henderson, 208 N.J. at 255-56, 27 A.3d 872.

Defendant's trial began two months later. Before the teller testified, the prosecutor told him that the person "accused of committing [the] robbery" would be "in court seated at the defense table." At trial, the teller identified defendant, who was seated at that very spot.

There was no "good reason" for the State to conduct a first-time in-court identification here. The teller did not have a prior relationship with defendant or know him from before. The extremely suggestive in-court identification in this case therefore should not have taken place. No court rule at the time required the State to give advance notice or disclose the guidance it gave the teller during trial preparation, and we do not suggest bad faith on the part of the prosecution. With or without the benefit of today's ruling, however, the nature of the identification in this case raises concerns.

Not only was the procedure highly suggestive given the layout of the courtroom and the stage of the case, but it is also difficult to imagine a more suggestive comment than telling a witness ahead of time where the defendant will be sitting -- and then asking the witness to make an identification. Law enforcement officers should avoid similar comments before out-of-court identifications. Henderson, 208 N.J. at 250, 261, 27 A.3d 872. The same applies to the prosecution team before trial.

In this case, there was no physical or forensic evidence aside from video surveillance tapes. The State's case rested on identification evidence, which can be quite powerful. "[T]here is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says ‘That's the one!’ " Watkins v. Sowders, 449 U.S. 341, 352, 101 S.Ct. 654, 66 L.Ed.2d 549 (1981) (Brennan, J., dissenting) (quoting Elizabeth F. Loftus, Eyewitness Testimony 19 (1979)).

The jury saw a video of the robber wearing a baseball cap that obscured the top 20 to 25 percent of his face. The jury also heard two witnesses identify defendant -- the teller and defendant's former girlfriend, Joan. On cross-examination, defense counsel probed Joan's possible bias in light of the way defendant ended their years-long relationship. But rather than assess her testimony on its own, the jury heard inadmissible identification evidence from the teller that reinforced her pivotal identification.

Because defendant did not object at trial, we review the issue for plain error under Rule 2:10-2. We find that the error in this case was "clearly capable of producing an unjust result." Ibid. Because defendant's identification was central to the case, the error raises a reasonable doubt as to whether the jury reached a result it otherwise might not have. See State v. Singh, 245 N.J. 1, 13, 243 A.3d 662 (2021). We therefore reverse defendant's conviction.

At a retrial, the State may not ask the teller to identify defendant again. Twenty-two months after the robbery, and months after he identified another person as the robber, the prosecution in effect told the teller whom to identify in court as the defendant. Under the circumstances, any further identification by the teller presents a "very substantial likelihood of irreparable misidentification." See Henderson, 208 N.J. at 289, 27 A.3d 872.

IV. Video Narration Testimony

We next turn to Sergeant Vitelli's narration of the surveillance videos. Defendant objected to the testimony at trial and argued that it violated N.J.R.E. 701 and related evidentiary principles.

A.

We begin with Rule 701, which governs the use of lay opinion testimony and serves as a springboard for our analysis. The rule provides that "testimony in the form of opinions or inferences may be admitted if it: (a) is rationally based on the witness’ perception; and (b) will assist in understanding the witness’ testimony or determining a fact in issue." N.J.R.E. 701. The rule follows Fed. R. Evid. 701 in substance. See Sup. Ct. Comm. Cmt. on N.J.R.E. 701 (1991), reprinted in Biunno et al., Current N.J. Rules of Evidence 749 (2023).

Underlying Rule 701 is a basic principle set forth in N.J.R.E. 602 :

A witness may testify to a matter only if evidence is introduced sufficient to support a finding that the witness has personal knowledge of the matter. Evidence to prove personal knowledge may, but need not, consist of the witness’ own testimony. This rule does not apply to expert testimony under [N.J.R.E.] 703.

The rule also follows its federal counterpart, Fed. R. Evid. 602. See Sup. Ct. Comm. Cmt. on N.J.R.E. 602 (1991), in Biunno, at 749. In short, Rule 602 requires that a lay witness have "personal knowledge" to be allowed to testify. Expert testimony, by comparison, is addressed by N.J.R.E. 702 : "If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education may testify thereto in the form of an opinion or otherwise." The Court recently addressed Rule 701 and certain relevant case law in State v. Higgs, 253 N.J. 333, 290 A.3d 1235 (2023). As the Court explained, the rule has two parts. Drawing on language from State v. McLean, 205 N.J. 438, 457, 16 A.3d 332 (2011), the Court noted that the first prong -- the "perception" prong -- requires that a witness's "testimony ... be based on the witness's ‘perception,’ which ‘rests on the acquisition of knowledge through the use of one's sense of touch, taste, sight, smell or hearing.’ " Higgs, 253 N.J. at 363, 290 A.3d 1235 (omission in original) (citation omitted); see also N.J.R.E. 602. The second prong -- the "helpfulness" prong -- limits lay opinion testimony "to testimony that will assist the trier of fact either by helping to explain the witness's testimony or by shedding light on the determination of a disputed factual issue." Ibid. (citation omitted).

The personal knowledge requirement in Rule 602 has similarly been described as requiring "firsthand knowledge -- that which comes to the witness through his own senses, mostly sight and hearing." Christopher B. Mueller & Laird C. Kirkpatrick, Federal Evidence, § 6.5 (4th ed., July 2022 update); see also id. § 6.6 ("[P]ersonal knowledge means seeing directly the acts, events, or conditions in question"; "hearing them"; or getting sensory input from "the other senses of touch, smell, or taste."). The threshold showing "is low, requiring only enough ‘to support a finding’ by some rational juror of personal knowledge." Paul F. Rothstein, Federal Rules of Evidence, Rule 602 (2023).

B.

Recent case law has considered how Rules 701 and 602 apply to testimony by law enforcement officers. We review the cases for context and background even though they do not squarely address the issue presented in this appeal.

In State v. Lazo, the Court reviewed whether it was proper for an officer to testify that an arrest photo taken of the defendant in another matter "closely resembled a composite sketch of the assailant" in the case under investigation. 209 N.J. 9, 12, 14, 34 A.3d 1233 (2012). The sketch had been drawn based on the victim's description. Id. at 14, 34 A.3d 1233. The officer "had not witnessed the crime and did not know [the] defendant; [his] opinion stemmed entirely from the victim's description." Id. at 24, 34 A.3d 1233. We found that the testimony, which "was not based on prior knowledge," did not comply with Rule 701. Ibid.

In Singh, a detective testified that sneakers depicted in a surveillance video were similar to ones he saw the defendant wearing at the time of his arrest. 245 N.J. at 19, 243 A.3d 662. The sneakers and the video were in evidence. Id. at 4, 243 A.3d 662. Because the detective "had first-hand knowledge of what the sneakers looked like" from the arrest, we found the testimony "was rationally based on his perception." Id. at 19-20, 243 A.3d 662. We also found the testimony was "helpful to the jury" even though the jurors "may have been able" to compare the items. Id. at 20, 243 A.3d 662. We explained that Rule 701 does not require that the lay witness have "superior" ability or "offer something that the jury does not possess." Id. at 19, 243 A.3d 662.

In State v. Sanchez, the Court addressed identification testimony from a parole officer who had met with the defendant more than thirty times. 247 N.J. 450, 458, 255 A.3d 1118 (2021). As part of a homicide and robbery investigation, the officer identified the defendant in a still photo from a surveillance video. Ibid. The trial court found the proposed testimony failed both prongs of Rule 701. Id. at 462, 255 A.3d 1118. We reached the opposite conclusion. Because the officer had met with the defendant on many occasions, we found the identification satisfied Rule 701 ’s perception prong. Id. at 469, 255 A.3d 1118. As to the helpfulness prong, we noted that the officer's contacts with the defendant "enable[d] her to identify him ... more accurately than a jury could," and that the quality of the surveillance photo in question -- neither too blurry nor too clear -- also weighed in favor of admitting the testimony. Id. at 474-75, 255 A.3d 1118.

Most recently, in Higgs, we considered comments a detective made at trial about an image on a dashcam video. 253 N.J. at 365, 290 A.3d 1235. The officer testified the defendant had a gun in his back waistband, which contradicted a critical point in the defendant's testimony. Ibid. We found the admission of the testimony violated Rule 701. Id. at 366, 290 A.3d 1235. As the Court explained, "the jury was able to view the video" and was "as competent" as the officer to determine what it showed. Ibid. The Court concluded that testimony "invaded the province of the jury." Ibid.

The Court also noted the detective's testimony was based entirely on his watching the video but did "not rule out the possibility of allowing a law enforcement officer to testify about a sequence in a video that is complex or particularly difficult to perceive." Id. at 366-67, 290 A.3d 1235 (citing United States v. Begay, 42 F.3d 486, 502-03 (9th Cir. 1994) ; United States v. Torralba-Mendia, 784 F.3d 652, 659-60 (9th Cir. 2015) ).

C.

Courts in other jurisdictions have wrestled with the same questions presented in this appeal: under what circumstances can a lay witness testify about a video recording, and what limits apply to that type of testimony? We separately consider their analysis of Rule 701 ’s perception and helpfulness prongs and other pertinent evidence rules.

1.

Several courts have concluded that a lay witness's careful review of a video recording before trial satisfies the "perception" or "personal knowledge" requirements embodied in Rules 701 and 602.

In Begay, the Ninth Circuit addressed narration testimony by an officer who had watched a recording more than 100 times but did not witness the underlying event. 42 F.3d at 502. He also prepared an enhanced version of portions of the video and used it to explain his observations. Ibid. The appellate court found the officer had "sufficient personal knowledge" to narrate the enhanced video "based on his own perceptions" and "extensive review" of the full recording, even though he was not present at the incident depicted. Id. at 503 (citing Fed. R. Evid. 602 ).

Relying on Begay, the Ninth Circuit reached the same conclusion in Torralba-Mendia, 784 F.3d at 659. In that case, the court affirmed the admission of lay witness narration testimony from an officer who "had watched each video roughly fifty times." Ibid.

Other courts have similarly held that watching a recording of an event can satisfy the "perception" or "personal knowledge" requirements of Rules 701 and 602. See, e.g., McRae v. Commonwealth, 635 S.W.3d 60, 68, 70-71 (Ky. 2021) (citing Kentucky's version of Rules 701 and 602 and upholding a detective's narration of videos, including "events he was not personally familiar with"); State v. Small, 351 Wis.2d 46, 839 N.W.2d 160, 165 (Ct. App. 2013) (finding an officer could provide a lay opinion about the contents of a video he had watched 50 to 100 times "because it was ‘rationally based’ on his ‘perception’ "); People v. Fomby, 300 Mich.App. 46, 831 N.W.2d 887, 890-91 (2013) (finding that an officer's identification of individuals depicted in still photos taken from a surveillance video was "rationally based on his ... perception of the video" under the equivalent of Rule 701, even though he "did not observe firsthand the events depicted on the video"); State v. Walker, 135 N.E.3d 444, 460 (Ohio Ct. App. 2019) (finding that the lead investigator who reviewed surveillance footage he had seized from a home "had personal knowledge of the contents of the video" under the equivalent of Rule 602, even though he "did not personally make or appear in the ... video"), rev'd on other grounds sub nom., State v. Dent, 163 Ohio St.3d 390, 170 N.E.3d 816 (2020) ; see also Guay v. Sig Sauer, Inc., 615 F. Supp. 3d 66, 72 (D.N.H. 2022) (finding that an employee's deposition testimony about the contents of a video, based on "personal knowledge" from having watched the video, could be introduced under Fed. R. Evid. 602, without reference to Fed. R. Evid. 701 ); State v. Holley, 327 Conn. 576, 175 A.3d 514, 538-39 (2018) (noting in dicta that, although courts are divided, "there is significant authority under [R]ule 701 of the Federal Rules of Evidence to support the proposition that a lay witness narrating a video to a jury may state his or her impressions of what is depicted in the video, even if he or she did not observe those events firsthand," and concluding, without deciding whether there was error, that "any error" in the case was harmless).

Some courts, however, have barred lay witnesses from narrating a recording unless they were present at the event depicted. See Callaham v. United States, 268 A.3d 833, 848 (D.C. 2022) (rejecting the argument that detectives "obtained personal knowledge" of events "solely by watching recorded surveillance footage"); Boyd v. Commonwealth, 439 S.W.3d 126, 131-32 (Ky. 2014) (finding that lay witnesses could narrate segments of a recording depicting events they "perceived in real time," but not parts of the video "they did not perceive in real time"); Wells v. State, 604 So. 2d 271, 279 (Miss. 1992) (finding that a lay witness's narration testimony was improper because he did not have "first-hand knowledge of the events depicted on the tape"); Nadeau v. Hunter Lawn Care, LLC, 585 F. Supp. 3d 158, 161 (D. Mass. 2022) (barring a witness from testifying "as to what he saw on [a] surveillance video because he did not perceive those events as they happened").

2.

Other courts have evaluated Rule 701 ’s helpfulness prong and found that certain lay witness narration testimony can be "helpful" to the jury and pass muster under the equivalent of Rule 701(b). In Begay, the Ninth Circuit found that an officer's narration testimony "was likely to have been helpful to the jury." 42 F.3d at 503. In that case, the officer watched a video recording that depicted a demonstration involving more than 200 people. Id. at 502-03. The demonstration turned into "a violent confrontation with tribal police" in which officers were assaulted, police cars were vandalized, a building was ransacked, and two demonstrators were killed. Id. at 489. The officer also copied parts of the video in slow motion and added "color-coded circles and arrows" to help the jury trace the movements of the defendants. Id. at 502.

The jury watched the full video in its entirety. Id. at 503. The court nonetheless noted "it is reasonable to assume that one viewing a videotape of" the demonstration "would likely not see certain details, given the tremendous array of events all occurring simultaneously." Ibid. As a result, the officer's "testimony concerning which persons were engaged in what conduct at any given moment could help the jury discern correctly and efficiently the events depicted in the videotape." Ibid. The court added that to have the jury devote as much time as the officer to view the complex video "would be an extremely inefficient use of the jury's and the court's time." Ibid.

In Torralba-Mendia, the Ninth Circuit similarly found that an officer's narration testimony "helped the jury understand what they were seeing." 784 F.3d at 659-60. The prosecution involved a human smuggling operation in which individuals were smuggled across the border with Mexico, driven to a shuttle company in Tuscon, and then shuttled to safe houses where they were held "until family members paid for their release." Id. at 657. The officer narrated surveillance videos that showed "vehicles dropping off and picking up people from" the shuttle company. Ibid. He also testified that he "often watch[ed] the video feed live while it was being recorded." Ibid.

The Ninth Circuit noted it had "previously held that an officer who has extensively reviewed a video may offer a narration, pointing out particulars that a casual observer might not see." Id. at 659 (citing Begay, 42 F.3d at 502-03 ). In this case, the court explained, the officer's testimony

provided the length of time lapses between video clips[;] ... pointed out unique characteristics of the vehicles -- like their makes, models, and whether any bodywork had been done to them -- that helped the jury identify the same cars in subsequent videos[;] ... linked the different cars to specific conspirators[;] ... counted the number of passengers exiting or entering the vehicles (a difficult task because the video's angle obscured the view)[;] [a]nd ... pointed out the particular clothing of certain passengers, to show that a person dropped off in one video was the same person picked up in a later video.

[ Id. at 659-60.]

The officer also offered expert testimony in other areas, which is not relevant to this appeal. Id. at 657. When he narrated the surveillance videos, he testified as a lay witness. Ibid.

In People v. Son, a detective "pointed out various aspects" of a video, which the trial court found "helpful" under the equivalent of Rule 701(b). 56 Cal.App.5th 689, 270 Cal. Rptr. 3d 83, 87, 89-90 (2020). The detective testified she had "watched the video at least 50 times and that subsequent viewings revealed details she had not picked up on at first." Id. at 89. Although the court acknowledged "the jury could have watched the video repeatedly and picked those details up on their own," it emphasized that "the standard is not whether the testimony was essential. It's whether it was helpful." Id. at 90. The court added that the detective's testimony helped the jury "speed up the process of teasing out obscure details in the video." Ibid.

The court in People v. Brown similarly found that a detective's narration of the content of a surveillance recording was helpful to the jury. 415 Ill.Dec. 203, 82 N.E.3d 148, 167 (Ill. App. Ct. 2017) (relying on the state's version of Rules 701 and 602 ). The officer testified about interactions between certain individuals and their movements in a parking lot. Id. at 157. The court noted the testimony "helped the jury focus on the relevant action because numerous people entered and exited the parking lot area, their faces were too distant from the camera to be discernible, their outdoor winter clothing somewhat concealed their identities, and the color of their clothing was muted." Id. at 167. The court found it was error only for the officer to identify the defendant in the video. Id. at 167-68.

D.

1.

Our reading of the evidence rules and the case law suggests that no single rule is a perfect fit for narration evidence by a witness who did not observe events depicted in a video in real time. To resolve the issue of admissibility, we borrow from key aspects of Rules 701, 602, and 403.

A fact witness who participated in an event can of course testify about it and recount relevant evidence without resort to Rule 701. Rule 701, in its purest form, allows witnesses to offer lay opinion testimony based on their firsthand observation of events in real time. Beyond that, they can offer lay opinion testimony about parts of a recording that depict what they perceived in real time. In those instances, they plainly have sufficient personal knowledge within the meaning of Rules 602 and 701.

By way of example, bank employees can testify about the portion of a recording that depicts their encounter with a robber. Investigators can likewise testify about parts of a recording that reflect their investigative work at a crime scene or their familiarity with an area depicted in a video based on prior experience. See, e.g., McRae, 635 S.W.3d at 69.

As discussed above, courts differ as to whether Rule 701(a) authorizes narration testimony from individuals who did not witness the actual event captured in a video as it happened. Their testimony would be based on observations of an objective record of the events in question. Their testimony would also differ from typical lay opinion witnesses in another important way. As we discuss later, narration evidence by a witness who did not observe events depicted in a video in real time may not include opinions about a video's content and may not comment on facts the parties reasonably dispute. Analyzing whether lay opinion testimony is helpful under Rule 701(b) is more straightforward and depends on the facts of the individual case.

If narration testimony is allowed under Rule 701, other concerns arise, like the proper scope of what a witness may say. Plus the discovery rules do not require advance disclosure of typical lay opinion testimony under Rule 701, unlike requirements that apply to expert reports linked to Rule 702. See Rule 3:13-3(b)(I).

Narration testimony is not a classic form of expert testimony under N.J.R.E. 702 either. If an investigator simply watches a video multiple times, and even slows it down to better grasp what it reveals, it is not clear the person has acquired "scientific, technical, or other specialized knowledge," as the rule requires. See N.J.R.E. 702. Put another way, what type of expert would investigators who have viewed a complex video with care be? What specialized knowledge do they possess? To what field of expertise do they belong? Designating someone as an expert also poses another challenge: Does the designation enhance the witness's status and the weight of their testimony in the eyes of the jury? Does it aggravate concerns that an "expert" is invading the jury's province?

At the same time, it is not reasonable to expect that jurors can perceive and understand all parts of a complex incident, or even a fleeting gesture in a simple scene, that is depicted in a video. Counsel, of course, can pinpoint particular spots in a video during closing argument. But that is not a satisfactory solution in all cases. Not only would it be impractical with a confusing or complex video, but it might also invite objections that counsel is essentially testifying without being subject to cross-examination.

2.

We believe that Rules 701, 602, and 403 in tandem provide the proper framework to assess video narration evidence by a witness who did not observe events in real time. The individual is not an eyewitness to a crime, which would require firsthand knowledge of the event. The witness is instead commenting on an independent source of evidence -- an objective recording of the event. And the testimony offered is based on the witness's direct perception of the video. An investigator who has carefully reviewed a video a sufficient number of times prior to trial can therefore satisfy the rules’ "perception" and "personal knowledge" requirements as to what the video depicts. See N.J.R.E. 602, 701(a) ; see also Begay, 42 F.3d at 502-03 ; Torralba-Mendia, 784 F.3d at 659.

But Rule 701 also contains a critical limiting requirement that is important in the context of narration evidence: a lay witness may only present testimony that will be helpful to the jury. N.J.R.E. 701(b). Whether narration testimony is "helpful" depends heavily on the nature of the recording and the proposed comments. Even though it is for the jury to determine what a recording depicts, there are times when narration testimony can help jurors better understand video evidence and aid them in "determining a fact in issue." Ibid.

Scenes like the video of hundreds of demonstrators and violent incidents in Begay are a good example. See 42 F.3d at 502. Other types of chaotic or confusing recorded events that are not easy to follow raise similar concerns. It is unrealistic to think jurors can absorb that type of content when a video is played at trial, or that they will repeatedly play a video during deliberations and trace individual movements from frame to frame. But whether a narrator is needed is not the critical question. Rule 701(b) asks whether evidence is helpful, not whether it is necessary. See Son, 270 Cal. Rptr. 3d at 90.

Lengthy videos present similar problems. Even with short videos, jurors might miss a small or nuanced detail. An investigator who has carefully analyzed a video and can draw a jury's attention to particular spots can be quite helpful to the finder of fact. The jury can then "make its own evaluation." Watson, 472 N.J. Super. at 465, 277 A.3d 39. The clarity of a video can also be a relevant consideration. Id. at 468, 277 A.3d 39. If footage is unclear or grainy, but not reasonably in dispute, testimony from investigators might help the jury. The opposite is true if a video is so unclear that it is difficult to decipher and the parties dispute its contents. See, e.g., Higgs, 253 N.J. at 366-67, 290 A.3d 1235. Also, narration testimony may not be helpful if a video is clear and not otherwise hard to follow or grasp. See Sanchez, 247 N.J. at 473, 255 A.3d 1118.

Screenshots, stills, composite videos, and other demonstrative aids can help focus on details that casual observers might miss. Any testimony about them, however, is likewise limited by the rules of evidence considered here.

In sum, whether narration evidence is helpful turns on the facts of each case. Rule 701 ’s helpfulness prong can be satisfied when an investigator draws attention to key details that might be missed, or helps jurors follow potentially confusing, complex, or unclear videos that may otherwise be difficult to grasp. Counsel may offer other reasons to allow limited narration testimony, which courts should evaluate with care.

Narration testimony must also comply with N.J.R.E. 403. The rule guards against the risk of "[u]ndue prejudice, confusion of issues, ... misleading the jury, ... [and] needless presentation of cumulative evidence." Placing appropriate limits on narration testimony can help avoid those problems.

In holding that Rules 701, 602, and 403 provide a framework for the admission of narration evidence, we note again that such testimony must accord with specific limits we discuss later -- specifically, that witnesses cannot offer opinions on the content of a recording or comment on reasonably disputed facts.

A trial court's evidentiary ruling on the admissibility and scope of narration testimony is subject to review for abuse of discretion. Singh, 245 N.J. at 20, 243 A.3d 662. E.

We provide additional guidance both to emphasize the limited nature of narration testimony and to ensure that the testimony does not improperly intrude on the jury's domain. Although much of the case law addresses testimony by law enforcement officials, defense investigators can also present appropriate narration testimony. The following principles apply to both.

First, neither the rules of evidence nor the case law contemplates continuous commentary during a video by an investigator whose knowledge is based only on viewing the recording. To avoid running commentary, counsel must ask focused questions designed to elicit specific, helpful responses. "What do you see?" as an introductory question misses the mark.

Second, investigators can describe what appears on a recording but may not offer opinions about the content. In other words, they can present objective, factual comments, but not subjective interpretations. See Boyd, 439 S.W.3d at 131 ; Begay, 42 F.3d at 503 ; see also N.J.R.E. 403. The "individual opened the door with his elbow" can be allowed if not reasonably disputed; he did so "to avoid leaving fingerprints" cannot. See Watson, 472 N.J. Super. at 462, 277 A.3d 39. The former comment appropriately sets the stage for the factfinder to reach its own conclusion.

Third, investigators may not offer their views on factual issues that are reasonably disputed. See id. at 467-68, 277 A.3d 39. Those issues are for the jury to decide. See Higgs, 253 N.J. at 366-67, 290 A.3d 1235 (cautioning against testimony by a law enforcement officer that "usurp[ed] the jury's assessment of" disputed facts). So a witness cannot testify that a video shows a certain act when the opposing party reasonably contends that it does not. We include a reasonableness requirement to prevent a party from disputing all facts in a recording in a manner that does not reflect good faith. Fourth, although lay witnesses generally may offer opinion testimony under Rule 701 based on inferences, investigators should not comment on what is depicted in a video based on inferences or deductions, including any drawn from other evidence. That type of comment is appropriate only for closing argument. We therefore do not adopt factor four -- "Inferences and Deductions" -- in the Appellate Division's opinion. See Watson, 472 N.J. Super. at 468, 277 A.3d 39.

Consistent with those principles, an investigator who carefully reviewed a video in advance could draw attention to a distinctive shirt or a particular style of car that appear in different frames, which a jury might otherwise overlook.

In a drunk-driving case, an investigator could testify that a bartender served a beverage to a customer with a thick beard at 11:15 p.m., at 11:45 p.m., and at 12:20 a.m. -- while focused on each individual frame. If the person's identity is not in dispute, the investigator could testify that the bartender served the defendant.

But the investigator could not say the defendant had been drinking heavily for an hour and a half. Nor could an investigator say, "that's the same blue car" or "that's the defendant," if those facts were disputed. See Singh, 245 N.J. at 18, 243 A.3d 662. It is for the jury to draw those conclusions.

State v. Dante Allen provides additional examples of what narration evidence may and may not include. 254 N.J. 530, 298 A.3d 1033 (2023). The case involved a shooting incident with sharply disputed facts. The State maintained the "defendant deliberately shot at [a] police officer"; defendant claimed his "handgun fired accidentally" and "was not aimed at the officer." Id. at 534, 298 A.3d 1033. Both the officer and the defendant testified. The lead detective, who helped process the crime scene but did not observe the shooting incident, testified as well. Ibid. Among other things, he testified about the contents of a surveillance video that captured the shooting incident. Ibid. The Court found that certain parts of the detective's testimony were improper. In particular, his testimony that the video showed the defendant turn and fire his weapon at the officer -- facts that were reasonably disputed -- was improper. Id. at 547–49, 298 A.3d 1033. The Court, however, found no error in the detective's narration testimony about how parts of the video helped him "locate areas of interest for his processing of the crime scene." Id. at 539, 548–49, 298 A.3d 1033.

The above examples are illustrative; they are not a comprehensive list. To avoid missteps before the jury, prosecutors must provide a written summary of proposed narration testimony to defense counsel, and vice versa, before trial. Counsel should then confer among themselves to try to narrow areas of disagreement. For items that remain in dispute, the proponent of the evidence should file a motion in limine to introduce the narration testimony. See Watson, 472 N.J. Super. at 473, 277 A.3d 39. The trial court, in its discretion, may conduct a Rule 104 hearing to resolve any outstanding issues. Ibid. We call on the Criminal Practice Committee to consider including those steps in the court rules.

The limiting principles and notice requirement outlined above address legitimate practical concerns about the admission of narration testimony.

Like the Appellate Division, we ask the Model Criminal Jury Charge Committee to develop appropriate jury instructions for narration testimony. See id. at 474-75, 277 A.3d 39. Among other things, jurors should be told that they, and not a narration witness, are to decide what a video shows, and that they may reject any narration testimony. Ibid.; see, e.g., Callaham, 268 A.3d at 848-49 ("[W]e endorse the trial court's efforts to ensure that the jurors in this case understood that they were the finders of fact and it was for them to decide what the video footage showed.").

F.

The above discussion does not alter other settled issues. Any party may hire an expert to enhance the quality of an electronic or video recording. The witness, if qualified as an expert, can testify about the modification, consistent with N.J.R.E. 702.

To present evidence of more elaborate forensic techniques that track an individual across a video, like "pixel tracking," a party is typically required to call an expert witness pursuant to Rule 702. See State v. Bruny, 342 Conn. 169, 269 A.3d 38, 51-55 (2022) (upholding expert testimony from an FBI forensic video analyst who relied on "his specialized experience in tracking analyses and the adaptation of computer technology to enhance ... video surveillance footage" and create a 52-minute enhanced video; the expert "compiled footage from multiple cameras, lightened images, added alphanumeric labels" to track "individuals throughout the enhanced video, inserted ‘halos’ over images ... and simultaneously displayed coverage from two cameras in a slow motion, split screen image"); Commonwealth v. Williams, 255 A.3d 565, 572-76 (Pa. Super. Ct. 2021) (upholding the expert testimony of a certified forensic video analyst who used "specialized tools" to track the movement of individuals and "validate that tracking").

Specialized knowledge would not ordinarily be required for other types of adjustments, like adjusting the speed of a video or creating a straightforward composite video, a screenshot, or an enlarged photo from a video. A lay witness can testify about those basic techniques, and parties can, of course, stipulate to admissibility in this and other areas. Once again, though, more elaborate forensic techniques should be presented by a qualified expert consistent with N.J.R.E. 702. As video technology advances, such expertise may become more prevalent.

A lay witness with the requisite knowledge can also testify about basic foundation issues: how long a recording is; when it was recorded; what device recorded it; the angle of the recording device; and similar points. See Watson, 472 N.J. Super. at 466, 277 A.3d 39. G.

Because we reverse defendant's conviction on the basis of the in-court identification evidence, it is not necessary to analyze each aspect of Sergeant Vitelli's narration testimony. We instead offer a few observations about places where the testimony went awry.

As noted earlier, the testimony began with an open-ended question -- "What do you see?" -- rather than focused questions about particular factual, reasonably undisputed details that might help the jury. The prompt, and questions that followed, invited "play-by-play" commentary about the suspect's movements throughout the 57-second surveillance video: how he entered the bank, that he removed a glove, placed a demand note on the teller's counter, held the note with two fingers, pushed open the door with his elbow, and appeared to run across the parking lot. The officer also described what appeared in various stills of the surveillance footage, without an objection from defense counsel.

The video was not confusing, chaotic, or lengthy; the jury, in general, could follow it on its own. Continuous commentary and speculation about each step the robber took inside the bank did not satisfy the requirements of the rules of evidence.

We make note of certain additional comments as well. At various points, the officer used qualifying language that sounded as though he was offering an opinion, rather than stating a fact. He testified, for example, that "it appears the suspect removes the glove"; "it appears that either his left hand is on the door or in the area of the door"; and "it looks like he's using his elbow." The officer may have used softer language in presenting the facts as a matter of style. On remand, the officer can clarify the nature of his proposed testimony before trial begins. He may not, however, testify at a retrial about what he suspected or believed had happened based on his observation of the video. See id. at 450, 277 A.3d 39. Nor may he offer his opinion about what the video or still photos depict. For example, it was not proper to tell the jury, "In my opinion, from my observations, it looks like the suspect has two fingers on the note, holding the note as it's on the counter." Similarly, it was not appropriate to offer an opinion that "the paper" between the robber's fingertips and the surface was "a barrier."

Sergeant Vitelli also testified that "the suspect was very careful in ... not attempting to leave any type of evidence behind." The judge properly sustained defense counsel's objection to the officer's subjective opinion. The court should have also stricken the testimony and directed the jury not to consider it.

Despite the brevity of the recording, it was appropriate to draw attention to the split second when the suspect used his elbow to open the outer door -- a point that might otherwise have been missed. On cross-examination, defense counsel could likewise have asked about the times defendant touched surfaces in the bank with his bare hands. Consistent with the above principles, defendants can also affirmatively call investigators, who have carefully reviewed a video, to highlight certain points.

We note as well that the officer appropriately refrained from referring to the robber as "the defendant" in his testimony. See Singh, 245 N.J. at 5, 18, 243 A.3d 662.

V. Confrontation Clause

Finally, we briefly address defendant's Confrontation Clause argument.

A.

The trial court denied the State's pretrial motion to admit evidence of "other crimes" -- the robberies in Princeton and Somerset County -- under N.J.R.E. 404(b). The trial court allowed only limited sanitized testimony to explain how defendant had become a suspect. At trial, Sergeant Vitelli answered two leading questions in order to provide context. He answered "yes" to "In November 2017, were you contacted by another law enforcement agency regarding Quintin Watson?" And he answered "yes, they were" to the compound question, "[A]t some point did you consult with that law enforcement agency ... after which criminal complaints were signed against Mr. Watson?" (emphasis added).

The Appellate Division found the testimony violated the Confrontation Clause but was harmless error.³ Because the State prevailed on the issue on appeal, it was not required to file a cross-petition as to the court's finding of a violation. See State v. Lefante, 14 N.J. 584, 589-90, 103 A.2d 585 (1954) ("[A] respondent who is merely seeking to maintain his judgment may brief and argue on the appeal any points that will sustain his judgment and if he does not brief and argue such points he will be taken to have waived them. Only when the respondent in certification is not relying on his judgment below but is seeking affirmatively to overrule or modify it must he cross-petition for certification."); Chimes v. Oritani Motor Hotel, Inc., 195 N.J. Super. 435, 443, 480 A.2d 218 (App. Div. 1984) ("[A]ppeals are taken from judgments, not opinions, and, without having filed a cross-appeal, a respondent can argue any point on the appeal to sustain the trial court's judgment.").

3 The parties debate the import of footnote 22 in the Appellate Division decision. See Watson, 472 N.J. Super. at 444 n.22, 277 A.3d 39. Parties should attempt to include all the reasons why they believe an error was -- or was not -- harmless in their briefs. But they do not waive a strand of a harmless error argument if they neglect to present it in writing and instead raise it at oral argument.

B.

A defendant's right "to be confronted with the witnesses against him" is guaranteed by the Sixth Amendment to the Federal Constitution and Article I, Paragraph 10 of the State Constitution. When an officer conveys information from someone who does not testify, either "directly or by inference," and the information incriminates the defendant, both the Confrontation Clause and the hearsay rule are implicated. State v. Branch, 182 N.J. 338, 350, 865 A.2d 673 (2005) ; see also State v. Bankston, 63 N.J. 263, 268-69, 307 A.2d 65 (1973).

"The common thread" throughout our case law "is that a police officer may not imply to the jury that he possesses superior knowledge, outside the record, that incriminates the defendant." Branch, 182 N.J. at 351, 865 A.2d 673 ; see also Bankston, 63 N.J. at 271, 307 A.2d 65 ; State v. Irving, 114 N.J. 427, 445-46, 555 A.2d 575 (1989) ; State v. Medina, 242 N.J. 397, 414-16, 231 A.3d 689 (2020).

Bankston elaborated on the problem. In that case, a police officer testified that he "had been talking to an informer and that based on information received," he approached the defendant at a tavern. 63 N.J. at 266, 307 A.2d 65. Although the officer "never specifically repeated what the informer had told" him, the Court found that "the inescapable inference from [the] testimony was that the informer had given information that" implicated the defendant in a narcotics offense. Id. at 271, 307 A.2d 65. The informer never testified and was not subjected to cross-examination. Ibid. The Court concluded the testimony should not have been admitted and amounted to reversible error. Id. at 271-73, 307 A.2d 65.

Confrontation Clause challenges are fact-specific. In this case, the Appellate Division focused primarily on the prosecutor's second question: Did the officer "consult with" another law enforcement agency "after which complaints were signed against" defendant? (emphasis added). The court concluded that, by answering affirmatively, the officer's "testimony did indeed imply that the other agency possessed incriminating evidence that was shared with Officer Vitelli but not revealed to the jury." Watson, 472 N.J. Super. at 429, 277 A.3d 39. The court explained the use of the word "consult" implied "an exchange of information" that "influenced" the decision to charge defendant with robbery in this case. Id. at 430-31, 277 A.3d 39. It is not entirely clear, however, whether the officer agreed to one or both parts of the compound question posed.

At a retrial, the prosecution should not refer, directly or indirectly, to the fact that it "consulted" with another law enforcement agency.⁴ Nor should the prosecution link any interaction with another police department with the filing of a criminal complaint, to avoid suggesting that incriminating information was conveyed. VI.

4 We do not opine on what may be presented if the defense suggests the police acted arbitrarily or with an improper motive in conducting the investigation. See Branch, 182 N.J. at 352, 865 A.2d 673. That did not happen at the first trial.

For the reasons outlined above, we reverse the judgment of the Appellate Division and remand for a new trial.

JUSTICES PATTERSON, SOLOMON, PIERRE-LOUIS, WAINER APTER, and FASCIALE and JUDGE SABATINO (temporarily assigned) join in CHIEF JUSTICE RABNER's opinion.