R v Weldekidan, 2019 MBCA 109

Citation: R v Weldekidan, 2019 MBCA 109 Date: 20191024 Docket: AR18-30-09080 IN THE COURT OF APPEAL OF MANITOBA Coram: Madam Justice Barbara M. Hamilton Madam Justice Jennifer A. Pfuetzner Madam Justice Karen I. Simonsen BETWEEN: ) J. M. Mann and ) A. Y. Kotlerher majesty the queen ) for the Appellant ) Appellant ) E. J. Roitenberg and ) L. C.

Robinson- and - ) for the Respondent ) haben abrham weldekidan ) Appeal heard: ) May 27, 2019 (Accused) Respondent ) ) Judgment delivered: ) October 24, 2019 pfuetzner JA [1] The accused was tried on four counts of attempted murder, seven firearms offences and a breach of recognizancearising out of a multiple shooting at a house party in the City of Winnipeg. [2] Three victims of the shooting each provided a warned and affirmed video-recorded statement (collectively, thestatements) to the police while in the hospital recovering from their injuries.

In the statements, each victim gave a detailed account of theevents leading up to the shootings and the shootings themselves. All three victims identified the accused as the shooter. [3] Because of the trial testimony of the three victims, the Crown sought to introduce the statements into evidence.

Aftera voir dire, the trial judge ruled that the statements did not have sufficient threshold reliability to justify their admission under theprincipled exception to the rule against hearsay. [4] As a result of the trial judge’s ruling, the Crown called no further evidence and the trial judge dismissed the chargesagainst the accused. [5] The Crown appeals the acquittal and raises several issues. It argues that, in her voir dire ruling, the trial judgemisapplied the law regarding the principled exception to the hearsay rule with respect to both procedural and substantive reliability.

Second, the Crown submits that the trial judge misapprehended or failed to consider the evidence and, finally, that she erred byconsidering the ultimate reliability of the statements rather than their threshold reliability. [6] For the reasons that follow, I would allow the appeal and order a new trial. In my view, the trial judge made afundamental error in principle when she considered the ultimate reliability of the statements at the threshold admissibility stage.

Inaddition, the trial judge erred in her application of the law regarding procedural reliability, including by taking too narrow a view ofwhether the accused would have “a full opportunity to cross-examine” the victims at trial (R v B (KG), (SCC), [1993] 1SCR 740 (KGB) at 796). Background [7] Prior to attending the house party, the three victims, Miracle Tamana (Tamana), Henok Goitom (Goitom) andMerhawi Mehari (Mehari), were at a bar called the “Red Sea”. The accused, who had known the victims for several years, got into anargument at the bar with Goitom.

After the argument, the victims left the bar and drove to the house party where they socialized,consumed alcohol and listened to music.

[ 8 ] According to his statement, Goitom was in the basement at the house party when the accused came up to him, tapped him on the hand and told him to come outside. Goitom assumed that the accused wanted to engage in a one-on-one fight. Once outside in the back lane, the accused pulled out a gun and shot Goitom in the abdomen. The accused then went to the backyard of the house. [ 9 ] In the meantime, Tamana had gone outside onto the back deck of the house to clear his head and get fresh air.

In his statement, Tamana said that, as he sat on the stairs, he looked up and saw the accused standing no more than a couple of steps away. The accused began shooting at him as Tamana raised his legs and tried to twist backward in an attempt to get away. He was shot seven times. The accused then walked toward the parking lot behind the house after which Tamana heard more shots. A friend came to assist Tamana and he was taken to the hospital. [ 10 ] In his statement, Mehari said that he and Adonay Gizaw (Gizaw) were in the parking lot helping Goitom into a car when the accused returned from the backyard.

The accused shot Mehari in the legs and he fell to the ground between two cars. The accused stood over him and said, “Of everyone, I want you dead” and shot Mehari in the stomach. Gizaw started to drive away with Goitom. The accused turned and shot at the car. Mehari hid his head under a parked car. The accused turned back and attempted to shoot him again, but the gun did not fire. The accused then got into a vehicle and drove away.

A friend came to assist Mehari until he could be taken to the hospital. [ 11 ] Tamana, Goitom and Mehari each gave individual statements to the police in the days after the shooting while recovering in the hospital.

The police followed the same procedure in obtaining the statements. [ 12 ] All three statements were video-recorded and the victims were told on camera that their statements would be video- recorded, and that they could be required to appear as a witness at trial. [ 13 ] The victims were warned on camera that it was an offence to make a false statement, and that if they changed their statement at trial, they could be charged with fabricating evidence. The victims were told on camera that they were not obligated to give a statement, and should disregard anyone’s attempt to persuade them to do so.

They acknowledged on camera that they understood each of these statements and agreed that they were choosing to give the statement and signed documents to that effect. They also affirmed on camera that the statement that they were about to give was true, and signed documents to that effect. [ 14 ] The statements show how they were obtained and they capture the demeanour of the victims during their statements and during the pre-statement procedures. [ 15 ] In the statements, each victim appears to be calm, cooperative, alert and coherent.

They respond to mainly open-ended questions, give significant details of the events, respond articulately to the questions asked, correct the officer if the officer misunderstands something, and they are willing to say “I don’t know” in response to some questions.

Proceedings in the Trial Court [ 16 ] At trial, Mehari and Goitom denied any memory of the shootings or of giving their statements to the police, essentially agreeing to defence suggestions that this was due to the medication they were on while in the hospital, and their intoxication on the night of the shooting. [ 17 ] Tamana provided detailed testimony about the events of the night and his shooting, but denied any memory about who shot him, stating that he was “black-out drunk.” [ 18 ] Gizaw testified that he was assisting Goitom into a car when the shooter appeared and shot at the car as he drove away.

A bullet shattered the windshield and injured Gizaw. However, he did not require medical attention. Gizaw was unable to identify the shooter, having only seen a figure in a black jacket. [ 19 ] Evidence admitted on consent at trial indicated that all three victims had been shot with the same gun. Medical records of each victim were admitted, but only Tamana’s included detailed consultation reports. A neurology consultation report indicates that he was “alert, oriented . . . pleasant . . . [illegible] cooperative” on the same day of his statement.

A clinical nurse specialist consultation report made two days prior to his statement indicates that he expressed concerns respecting his injuries, his fellow victims, his income, and his terror of being shot by his “friend”. [ 20 ] After unsuccessful cross-examinations pursuant to section 9(2) of the Canada Evidence Act , RSC 1985, c C-5 , the Crown applied to have the statements entered into evidence pursuant to KGB .

As part of the evidence on the KGB voir dire , the statements were shown and proved, and the officer who took the statements testified that he had asked permission from the nurses on duty to speak to the victims and that the victims seemed alert and cooperative at the time of their statements. [ 21 ] During the voir dire , Tamana agreed with defence counsel’s suggestion that the police told him who the shooter was prior to the statement, but the officer who took the statement specifically denied this.

Tamana also testified that there had been talk about the incident by hospital visitors, but he paid no attention to it and couldn’t recall any detailed conversation with visitors about what happened. He also testified that he had not spoken to Mehari or Goitom in the hospital. Voir Dire Decision [ 22 ] Following the KGB voir dire , the trial judge declined to admit any of the statements into evidence. Necessity of the statements was not an issue.

However, the trial judge stated that she had significant concerns about procedural reliability because, although the victims all appeared alert on the statements, there was evidence that they were medicated at the time the statements were taken.

The trial judge was troubled by the fact that there was “[n]o detailed evidence . . . provided . . . regarding . . . the potential effect of the medications upon their perception, memory or narration.” She also indicated that, “No evidence was called relative to how an individual on medication . . . would present to an observer in terms of their alertness, coherence and demeanour”. She also noted that she had no evidence on the issue of whether the victims “could appreciate the seriousness of [the] oath or affirmation,” or appreciate the

consequences of lying, given their medical condition and medications. [23] The trial judge found that there was no evidence that the statements were not given voluntarily. That is, there was noevidence of coercion, threats, promises or other misconduct on the part of the police (see KGB at pp 801-2; R v U (FJ), (SCC), [1995] 3 SCR 764 (UFJ) at para 51; and R v Khelawon, 2006 SCC 57 at paras 81, 87). [24] The trial judge specifically found that “each of the declarants can be cross-examined at trial”.

However, she concludedthat, as Mehari and Goitom “now remember very little”, “they will not be cross-examined effectively” and “their statements cannot betested.” She determined that procedural reliability had therefore not been established with respect to their statements. [25] Regarding Tamana, the trial judge found that he was available to be cross-examined on his statement and that there was“some basis on which . . . the trier of fact could assess the content of his statements and inquire into his perception, memory andnarration at the time the statement was made”.

However, the trial judge again noted her “significant concerns” about the impact of themedication on him and his appreciation of the affirmation and warning, and concluded that, “my concerns relative to proceduralreliability override the fact that [Tamana] can be cross-examined.” The trial judge found that procedural reliability had not beenestablished with respect to Tamana’s statement. [26] The trial judge then considered whether substantive reliability had been established regarding the statements.

Relyingupon the four-step test set out in R v Bradshaw, 2017 SCC 35, she considered whether the statements were made “under circumstanceswhich substantially negate the possibility that the declarant was untruthful or mistaken” (at para 31) and whether there was anycorroborative evidence. [27] She found that the specific hearsay danger presented by the statements was the difficulty assessing “perception,memory and narration.” She found that alternative or speculative explanations for the statements included that hospital visitors or thepolice may have told the victims that the accused was responsible, or had been arrested, for the shootings, and it was this informationthat led the victims to identify the accused rather than their own observations and memory. [28] The trial judge determined that there was no corroborative evidence that ruled out the alternative and speculativeexplanations for the statements since the only potentially corroborative evidence led at the voir dire relative to each statement was thestatements of the other two victims, none of which were inherently trustworthy. [29] The trial judge also rejected the argument that the statements were strikingly similar and thus corroborative as wasfound in UFJ, concluding that: I cannot rule out the possibility that the similarities are coincidental, that the declarants heard each others’ statements and modelled theirown accordingly, or that any one of the statements was the result of collusion or outside influence for the reasons that I have alreadyidentified regarding conversations at the hospital. [30] Finally, the trial judge determined that, even considering both the procedural reliability factors and the substantivereliability factors together, she was not satisfied that the deficiencies of each could be overcome by the other.

Jurisdiction and Standard of Review [31] Section 676(1)(

a) of the Criminal Code limits the Crown’s right to appeal an acquittal to questions of law alone.

Asrecognised by Cromwell J in R v JMH, 2011 SCC 45, “This limited right of appeal engages the vexed question of what constitutes, forjurisdictional purposes, an error of law alone” (at para 24). [32] It is well established that the admissibility of evidence, including hearsay evidence, raises a question of law: see, forexample, Fanjoy v The Queen, (SCC), [1985] 2 SCR 233 at para 9. [33] This principle was confirmed in R v Youvarajah, 2013 SCC 41, where Karakatsanis J, for the majority, also set out thestandard of review for decisions on the admissibility of hearsay evidence (at para 31): The admissibility of hearsay evidence, such as the prior inconsistent statement in this case, is a question of law.

Of course, the factualfindings that go into that determination are entitled to deference and are not challenged in this case. As well, a trial judge is well placedto assess the hearsay dangers in a particular case and the effectiveness of any safeguards to assist in overcoming them. Thus, absent anerror in principle, the trial judge’s determination of threshold reliability is entitled to deference: R. v.

Couture, 2007 SCC 28, [2007] 2S.C.R. 517, at para. 81. [emphasis added] [34] In addition to establishing that there was an error in law at the trial, when seeking a new trial on the appeal of anacquittal, the Crown must show that this error “might reasonably be thought . . . to have had a material bearing on the acquittal” (R vGraveline, 2006 SCC 16 at para 14; and Youvarajah at para 32). The Law

[35] Hearsay evidence is presumptively inadmissible, as there are recognised dangers in triers of fact relying on out-of-courtstatements for their truth.

However, hearsay evidence may be admissible pursuant to traditional exceptions or pursuant to the principledexception to the hearsay rule if the party tendering it demonstrates that necessity and threshold reliability are met on a balance ofprobabilities (see Bradshaw at para 23; and Khelawon at para 47). [36] In Bradshaw, Karakatsanis J explained how threshold reliability is established, stating (at paras 26-27): Threshold reliability is established when the hearsay “is sufficiently reliable to overcome the dangers arising from the difficulty oftesting it” (Khelawon, at para. 49).

These dangers arise notably due to the absence of contemporaneous cross-examination of the hearsaydeclarant before the trier of fact (Khelawon, at paras. 35 and 48). In assessing threshold reliability, the trial judge must identify thespecific hearsay dangers presented by the statement and consider any means of overcoming them (Khelawon, at paras. 4 and 49; R. v.Hawkins, (SCC), [1996] 3 S.C.R. 1043, at para. 75).

The dangers relate to the difficulties of assessing the declarant’sperception, memory, narration, or sincerity and should be defined with precision to permit a realistic evaluation of whether they havebeen overcome.

The hearsay dangers can be overcome and threshold reliability can be established by showing that (1) there are adequate substitutes fortesting truth and accuracy (procedural reliability) or (2) there are sufficient circumstantial or evidentiary guarantees that the statement isinherently trustworthy (substantive reliability) (Khelawon, at paras. 61-63; Youvarajah, at para. 30). [emphasis added] [37] As indicated in Bradshaw, there is no question that threshold reliability can be established solely on the basis ofprocedural reliability (see para 27; see also R v Okemow, 2019 MBCA 37 at paras 101-2). [38] In KGB, Lamer CJC, writing for the majority, described the basic difference between ultimate reliability and thresholdreliability.

He stated (at p 787): The ultimate reliability of the statement and the weight to be attached to it remain, as with all evidence, determinations for the trier offact.

What the reliability component of the principled approach to hearsay exceptions addresses is a threshold of reliability, rather thanultimate or certain reliability. [39] Lamer CJC warned that trial judges at the admissibility stage should not decide whether the prior inconsistent statementis true or more reliable than the witness’s trial testimony, as this would involve assessing the credibility and ultimate reliability of thestatement, which is a matter for the trier of fact.

It is only if the statement is admitted that the trier of fact may decide whether to makesubstantive use of the statement (see KGB at pp 803-4; and UFJ at para 52). [40] Maintaining the distinction between threshold and ultimate reliability is just as important in a judge-alone trial as it is ina trial by judge and jury. In Khelawon, Charron J wrote (at para 50): It is important that the trier of fact’s domain not be encroached upon at the admissibility stage.

If the trial is before a judge and jury, it iscrucial that questions of ultimate reliability be left for the jury — in a criminal trial, it is constitutionally imperative. If the judge sitswithout a jury, it is equally important that he or she not prejudge the ultimate reliability of the evidence before having heard all of theevidence in the case. [emphasis added] [41] In KGB, Lamer CJC discussed the task of the trier of fact in assessing the ultimate reliability of prior inconsistentstatements.

He stated (at pp 786-87): The reliability concern is sharpened in the case of prior inconsistent statements because the trier of fact is asked to choose between twostatements from the same witness, as opposed to other forms of hearsay in which only one account from the declarant is tendered. Inother words, the focus of the inquiry in the case of prior inconsistent statements is on the comparative reliability of the prior statementand the testimony offered at trial.

[emphasis added] [ 42 ] This succinctly illustrates the unique role of the trier of fact in dealing with a prior inconsistent statement of a trial witness. The trier of fact is required to compare the reliability of two different statements from the same person. As a result, in the admissibility hearing, the trial judge’s obligation is to ensure that the trier of fact will have the tools required to make that comparison.

As Charron J wrote in Khelawon (at para 76 ): The admissibility inquiry into threshold reliability, therefore, is not so focussed on the question whether there is reason to believe the statement is true, as it is on the question whether the trier of fact will be in a position to rationally evaluate the evidence.

The search is for adequate substitutes for the process that would have been available had the evidence been presented in the usual way, namely through the witness, under oath or affirmation, and subject to the scrutiny of contemporaneous cross-examination. [emphasis added] [ 43 ] Importantly, in KGB , Lamer CJC wrote that, in assessing ultimate reliability, the trier of fact should take into account all of the circumstances including (at p 804): [T]he demeanour of the witness at all relevant times (which could include when making the statement, when recanting at trial, and/or when presenting conflicting testimony at trial), the reasons offered by the witness for his or her recantation, any motivation and/or opportunity the witness had to fabricate his or her evidence when making the previous statement or when testifying at trial, the events leading up to the making of the first statement and the nature of the interview at which the statement was made (including the use of leading questions, and the existence of pre-statement interviews or coaching), corroboration of the facts in the statement by other evidence, and the extent to which the nature of the witness’s recantation limits the effectiveness of cross-examination on the previous statement . [emphasis added] [ 44 ] In KGB , Lamer CJC concluded that threshold procedural reliability of a prior inconsistent statement will be established and “there will be sufficient circumstantial guarantees of reliability to allow the jury to make substantive use of the [prior inconsistent] statement” (at p 796) where (at pp 795-96): (

i) the statement is made under oath or solemn affirmation following a warning as to the existence of sanctions and the significance of the oath or affirmation, (ii) the statement is videotaped in its entirety, and (iii) the opposing party, whether the Crown or the defence, has a full opportunity to cross-examine the witness respecting the statement. [ 45 ] There is limited jurisprudence that addresses the third requirement under the KGB test and what it means for the opposing party to have “a full opportunity to cross-examine the witness” (at p 796).

I will review some of the relevant cases from other provincial appellate courts. [ 46 ] There are cases where out-of-court statements have been held to be inadmissible, at least in part, on the basis that there was an insufficient opportunity for cross-examination where the declarant was available at trial, but testified that he or she had no memory of the events. Most of these cases are factually distinguishable from the present case where the statements are compelling in their detail and clarity and there is no question that they were voluntary.

See: R v Post , 2007 BCCA 123 (witness statement was not under oath and was the product of “coercion, suggestion and manipulation by the interviewing police officer” (at para 3)); R v TGN , 2007 BCCA 2 (witness was under arrest at the time of giving the statement and had a strong motivation to incriminate others (see para 20)); R v James , 2011 ONCA 839 (witness made an audio-recorded statement to police several months after the events and 12 years prior to the trial where the witness identified the accused from a photo array); R v Johnson , 2018 SKCA 28 ( KGB requirements were not met as the oath and warning were given after the statement was taken, the witness had possible motives to lie and there was an indication of police inducement (see para 50)); and R v Newsham , 2019 BCCA 126 (witness was the getaway driver in a robbery and had motive to minimise his involvement and to blame others). [ 47 ] On the other hand, there are provincial appellate cases where statements have been admitted, despite the declarant not being responsive at trial or claiming to have no memory.

See: R v PG , 2011 ONCA 50 (“A declarant’s lack of memory of the prior statement is merely one factor to take into account when determining the reliability of the statement” (at para 5)); R v Gardipy , 2012 SKCA 58 (the inability to cross-examine a witness who refuses to testify is “ just one consideration in the constellation of factors bearing on the issue of reliability” (at para 40)); R v Woodman , 2016 ONCA 63 ; and R v Zaba , 2016 ONCA 167 . [ 48 ] As previously noted, the Supreme Court of Canada in KGB clearly contemplated that a prior inconsistent statement could meet the test for procedural reliability even where “the nature of the witness’s recantation limits the effectiveness of cross- examination” on the prior inconsistent statement (at p 804). [ 49 ] In UFJ , Lamer CJC revisited KGB and provided some guidance on the third part of the test, indicating that (at para 34):

“I held in [ KGB ] that the gravest danger associated with hearsay evidence simply does not exist in the case of prior inconsistent statements because the witness is available for cross-examination.” Later, in discussing the proper procedure for the admissibility voir dire , he wrote (at para 46): The first factor contributing to reliability is the cross-examination of the witness. If the witness provides an explanation for changing his or her story, the trier of fact will be able to assess both versions of the story, as well as the explanation.

However, where a witness does not recall making an earlier statement, or refuses to answer questions, the trial judge should take into account that this may impede the jury’s ability to assess the ultimate reliability of the statement . [emphasis added] [ 50 ] In R v Devine , 2008 SCC 36 , Charron J stated that “the availability of the declarant to be cross-examined will not necessarily tip the scales in favour of admissibility.

In order for this factor to weigh in favour of admission, there must be ‘a full opportunity to cross-examine the witness’ at trial ( KGB , at p. 796)” (at para 26). [ 51 ] In Devine , the witness gave a video-recorded statement to the police under oath and following a warning about the consequences of lying. In the statement, she identified the accused as the perpetrator of an assault and robbery. At trial, the witness testified that she could not remember many details of the incident, including the identity of the perpetrator, due to her extensive drug use.

The trial judge viewed the video-recorded statement and noted that the witness was extremely coherent, detailed in her memory and confident in her description of the events (see R v Devine , 2005 ABPC 162 at para 20 ). Given his observations, and despite the drug issue, the trial judge admitted the statement into evidence on the basis that it was procedurally reliable. [ 52 ] On appeal to the Supreme Court of Canada, Charron J agreed that the statement met threshold reliability on the basis that all the KGB requirements had been met.

Charron J noted that, although the witness recanted her identification of the accused, “there was a meaningful opportunity to test her evidence through cross-examination” (emphasis added) (at para 28) and that the trier of fact was able to assess the witness’s demeanour in court and “gave a detailed account of her evasiveness and reluctance to identify [the accused] in the courtroom” ( ibid ).

In other words, cross-examination of the witness allowed the trier of fact to assess the veracity of the witness’s claim that she had no memory. [ 53 ] While both KGB and Devine refer to the importance of “a full opportunity to cross-examine the witness”, it is clear from both decisions that there are reasonable limitations on this aspect of the test for procedural reliability.

In other words, even though the nature of the witness’s recantation may limit the effectiveness of the opposing party’s cross-examination of the witness, this does not mean that the prior inconsistent statement cannot, or should not, be put before the trier of fact. Positions of the Parties [ 54 ] The Crown argues that the concerns identified by the trial judge, including the potential impact of the victims’ medications, were questions of ultimate reliability appropriate for trial and had no place in the procedural reliability analysis.

The Crown maintains that the statements can be admitted based solely on procedural reliability in the circumstances of this case. [ 55 ] The accused argues that the trial judge was entitled to be concerned about the effect of the medications on the victims and she did not err in finding that the evidence called by the Crown did not provide an adequate substitute for contemporaneous cross- examination. The accused points out that many of the factors that the trial judge was required to consider in assessing threshold reliability are also relevant to ultimate reliability.

The accused submits that the trial judge’s decision was fact-driven and there is no error in law. Analysis [ 56 ] As I described earlier, in Devine , a recanting witness’s statement was found to have met threshold reliability, despite her claim at trial that she could not remember much due to her drug use.

Accordingly, once it was determined that the KGB indicia of procedural reliability were present, such that the statement could be assessed by the trier of fact for its ultimate truth and accuracy, there was no need to consider the potential impact of the witness’s drug use at the time she made her statement, as those were issues relating to the ultimate reliability of the statement.

The parallels to the present case are clear. [ 57 ] As previously indicated, on several occasions the trial judge noted that all three victims were taking medications, but that there was no evidence as to the potential effects of the medications on them.

These statements reflect the trial judge’s ongoing concerns about the state of mind of the victims at the time their statements were given, and her attempts to determine whether the statements could be considered truthful and accurate given the medicated state of the victims. [ 58 ] In my view, the trial judge erred in principle in applying the law relating to procedural reliability by moving beyond a consideration of whether there were indicia of reliability upon which the trier of fact could rely to assess the truthfulness of the statements and, instead, determined that the statements were untrustworthy.

This was a matter for the trier of fact. Effectively, she pre- determined the weight that would be given to the trial evidence of the victims compared to their statements and found that the trial evidence would be more persuasive (see Khelawon at para 50 ).

The error is clearly illustrated by the trial judge’s statement that, “my concerns relative to procedural reliability override the fact that [Tamana] can be cross-examined.” While she refers to “procedural reliability”, in reality she was concerned about the ultimate reliability of Tamana’s statement. [ 59 ] In addition, the trial judge’s error of considering the ultimate reliability of the statements was compounded when she

assigned an overly narrow meaning to the third part of the KGB test—“a full opportunity to cross-examine the witness” (at p 796).

This part of the test must be interpreted in light of the purpose of the procedural reliability inquiry, the focus of which in the case of prior inconsistent statements is on ensuring that the trier of fact will be able to assess the comparative reliability of the prior statement and the testimony offered at trial (see KGB at pp 786-87). [ 60 ] The third part of the KGB test does not require that the witness be able to give substantively responsive answers to questions posed to him or her.

Indeed, the manner in which a witness responds to questions on cross-examination can speak volumes about the comparative reliability of his or her trial testimony and the prior statement, even if the substance of the witness’s answers are not responsive—for example, the witness says that he or she can’t remember the events or that he or she lied in the prior statement (see Devine (SCC) at para 28). [ 61 ] As I have explained, in KGB , Lamer CJC set out some of the factors that must be considered by the trier of fact in assessing the comparative reliability of two statements.

These include the witness’s demeanour at all relevant times, any reason offered for the change in evidence, any opportunity or motivation for fabrication, the events leading up to the making of the first statement and any corroborative evidence (see p 804). The trial judge was right to consider the extent to which the opportunity to cross-examine may be affected by the nature of the testimony offered at trial (see UFJ at para 46 ).

However, when the trial judge’s reasons are read as a whole, it is clear that she did not turn her mind to the fact that cross-examination of the victims could provide the trier of fact with important evidence relevant to the above factors. [ 62 ] Accordingly, it was a clear error for the trial judge to have concluded that procedural reliability had not been established with respect to Tamana’s statement after she found that “there is some basis on which [he] could be cross-examined and the trier of fact could assess the content of his statements and inquire into his perception, memory and narration at the time the statement was made with a view to assessing the ultimate reliability of the statement.” [ 63 ] The trial judge was appropriately alive to the fact that Tamana did provide more substantive answers to questions on cross-examination than Goitom and Mehari.

However, because she failed to consider the factors set out in KGB at p 804, the trial judge also erred in principle when she concluded, at the threshold admissibility voir dire , that Goitom and Mehari “will not be cross-examined effectively” and “their statements cannot be tested.” [ 64 ] While not necessary for the disposition of this appeal, in my view, there is an alternate basis for admission of the statements of Goitom and Mehari. As I have explained, there is no question that Tamana’s statement is admissible solely on the basis of procedural reliability.

Once that statement is admitted, it provides corroboration of material aspects of the statements of Goitom and Mehari, in particular, the identity of the shooter. Accordingly, the Crown could have used the statement of Tamana, along with other corroborative evidence such as the ballistics report, to support the admission of Goitom and Mehari’s statements on the basis of substantive reliability (see Bradshaw ; and Okemow at para 103 ). Conclusion [ 65 ] In conclusion, the trial judge made two errors in principle.

Fundamentally, she erred by stepping beyond an assessment of procedural reliability and considered ultimate reliability. Second, she erred in her application of the test for procedural reliability, including by giving an overly restrictive meaning to “a full opportunity to cross-examine the witness” ( KGB at p 796). [ 66 ] I am not persuaded that if the trial judge had applied the legal principles properly, the outcome of the case would have been the same.

If she had not strayed into considering the ultimate reliability of the statements and had properly applied the legal principles relating to procedural reliability, she would have concluded that the statements were procedurally reliable and admitted them as substantive evidence. The failure to admit the statements might reasonably be thought to have had a material bearing on the acquittal as each statement identified the accused as the shooter. [ 67 ] I would allow the appeal and order a new trial. JA I agree: JA I agree: JA