The principled approach to hearsay evidence focuses on whether there is some fact or circumstance that "compensates for, or stands in the stead of" the safeguards of reliability that exist when a witness gives first-hand evidence in the court room, namely, that the witness: (1) testifies under oath; (2) is present before the trier of fact to facilitate assessment of credibility; and (3) is subject to being tested by cross-examination: see R. v. B.(K.G.),  1 S.C.R. 740, at p. 787ff.
 In Khelawon, at paras. 62-63, the Supreme Court of Canada pointed to two ways the reliability requirement can be satisfied. The first is by showing "that there is no real concern about whether the statement is true or not because of the circumstances in which it came about". Second, it can be satisfied by showing that although the statement is in the form of hearsay, "its truth and accuracy can nonetheless be sufficiently tested."
 In my view, the trial judge failed to pay sufficient attention to the fact that these statements did not benefit from any of the hallmarks of reliability that have been identified in the case law as important to compensate for the absence of or stand in the stead of the oath, the presence of the witness before the trier of fact and the availability of cross-examination.
 First, none of the statements were made under oath and the complainant could not provide the police with a coherent explanation to demonstrate that she understood the difference between telling the truth and telling a lie. Although the trial judge found that the complainant did not appear to understand the difference between telling the truth and telling a lie, he gave that no weight in his reliability assessment. Moreover, he failed to deal with obvious misstatements she made as to her age and school grade or her confusion as to the number of times her father had "peed" on her. The trial judge found that reliability could be attributed to the fact that a young child would not have knowledge of the sexual acts described, yet he failed to deal with RP's evidence that he thought that she might have acquired that knowledge by watching pornography.
 A related point arises from the trial judge's finding that "[t]he statements were provided relatively close in time to the events described" and the use of that finding as an indicia of reliability. This reasoning reveals both a misapprehension of the evidence and a misreading of the leading case of Khan. It is clear from the record that the statements were not provided until at least several months after the events described. This stands in stark contrast to the situation in Khan where the infant complainant reported an act of sexual abuse within minutes of its occurrence. The delay in disclosure here was a factor that detracted from its reliability yet the trial judge treated the timing as a factor favouring reliability.
 Second, the complainant's statements to RP and LP were not recorded and the jury had only their recollection of what the complainant said to them. The statements to the police were recorded but they varied considerably from the statements made to RP and LP. Moreover, to the extent that the reliability assessment had to be made on the basis of the recollections of RP and LP, there were features of those statements that give rise to serious concerns as to reliability. The trial judge failed to advert to or explain those features. In particular, the suggestion that the complainant's mother said she looked "pretty" with her father's semen smeared on her face is highly implausible as was the suggestion that the appellant penetrated her from the front while she was on his lap, an allegation the trial judge expressly rejected in his reasons for sentence. Likewise, the trial judge failed to provide an adequate reason to alleviate the concern that significant aspects of the initial disclosure to RP were given in response to leading questions.
 Third, the appellant had no opportunity to cross-examine the complainant on her statements even though this could easily have been dealt with by acceding to the defence request that the complainant be made available for cross-examination. Cross-examination of the declarant is an important safeguard that will often render hearsay evidence sufficiently reliable to warrant admission: R. v. U.(F.J.),  3 S.C.R. 764, at para. 32; R. v. B.(K.G.), at p. 794.
 Clearly, in cases involving sexual offences against young children, cross-examination will often not be possible or appropriate. For example, had the necessity element rested on the fact that the complainant would be traumatized by testifying before the jury, cross-examination would be out of the question. However, here the complainant had testified without incident or protest before the mistrial and again on the voir dire when the trial resumed. In these circumstances, the trial judge should not have admitted the out-of-court statements unless the appellant had the opportunity to cross-examine her.
 In my opinion, admitting the out-of-court statements without affording the appellant the opportunity to cross-examine the complainant before the jury produced an unfair trial. The jury was left with an incomplete and potentially misleading picture. The jury did not know that the complainant had no present recollection of the alleged abuse or of having complained of such abuse. Even if the complainant's hearsay statements were otherwise admissible, fairness required that the appellant have the opportunity to point out to the jury that the complainant claimed to have no present recollection of her allegations of abuse and, if he chose to do so, to explore the reason or explanation for her lack of memory.