COURT OF APPEAL FOR ONTARIO
Tulloch C.J.O., Roberts and Rahman JJ.A.
BETWEEN
His Majesty the King
Respondent
and
Everton Hornsby
Appellant
Jolene Hansell, for the appellant
Adrianna Mills, for the respondent
Heard: February 6, 2026
On appeal from the conviction entered by Justice Nancy S. Kastner of the Ontario Court of Justice, on September 20, 2021.
A. Overview
[1] The appellant appeals his convictions for one count of assault, one count of assault with a weapon (a laundry basket), and two counts of sexual assault, all involving his spouse at the time. The offences spanned the period from December 2018 to June 2019. On March 15, 2022, he received a global custodial sentence of 42 months.
[2] The appeal turns on the appellant’s challenge to the trial judge’s assessment of the evidence. The trial focused on the credibility and reliability of the complainant and the appellant – the principal witnesses at trial – whose testimonies about the incidents underlying the charges were diametrically opposed in material respects.
[3] For the following reasons, I accept the appellant’s overarching submission that the trial judge committed numerous analytical errors in her assessment of the evidence, including by subjecting the defence evidence to materially greater scrutiny than the Crown evidence. These errors resulted in a flawed credibility and reliability analysis of the complainant and the appellant, and rendered the trial unfair. I would quash the convictions on all counts and order a new trial.
B. Factual background
[4] The complainant and the appellant met in 2014 at a festival in Toronto. When they met, the appellant was a migrant worker visiting from Jamaica and had no immigration status in Canada. They married in 2017. The complainant sponsored the appellant and his two children to immigrate to Canada. The appellant and his children became permanent residents.
[5] Their marriage ran into difficulties in 2018. In late 2018, the complainant admitted to infidelity. Their relationship deteriorated, and they agreed that the marriage was over in May 2019. They continued to live together in the complainant’s house until June 29, 2019, when the complainant called the police because the appellant broke down the bedroom door. The appellant testified that he was seeking his passport. The police recommended that the complainant leave the house and arrested Mr. Hornsby for mischief. On July 1, 2019, she reported to the police the incidents that underlie the counts of which the appellant was found guilty.
C. Issues
[6] The appellant raises several grounds of appeal. It is sufficient to consider under the rubric of the trial judge’s uneven scrutiny of the evidence the specific assertions that the trial judge erred in the following ways: by failing to reconcile material inconsistencies and flaws in the complainant’s evidence and address how her credibility was affected; by applying stereotypical thinking to diminish the appellant’s credibility; by failing to address evidence supportive of the appellant’s case; and by applying different and harsher standards to the appellant’s and other evidence supportive of the defence.
D. Analysis
a. General principles
[7] The uneven scrutiny ground of appeal is made out where a trial judge errs in the consideration of the evidence by using a different degree of scrutiny in assessing the credibility or reliability of defence evidence than Crown evidence, which renders the trial unfair: R. v. Kiss, 2018 ONCA 184, at paras. 82-83; R. v. Howe (2005), 2005 CanLII 253 (ON CA), 192 C.C.C. (3d) 480 (Ont. C.A.), at paras. 58-59. Uneven scrutiny may be another way of expressing that the trial judge erred in principle in some manner in making an assessment of credibility: R. v. B.T.D., 2022 ONCA 732, 163 O.R. (3d) 561, at para. 55, citing R. v. Greif, 2021 BCCA 187, at para. 82, leave to appeal refused, [2021] S.C.C.A. No. 182. The argument of uneven scrutiny can serve as an overarching framework in which to consider overlapping arguments for why a trial judge’s credibility findings are problematic. Indeed, it is often “tacked on” to arguments including insufficiency of reasons, misapprehension of evidence, reversing the burden of proof, palpable and overriding error, or unreasonable verdict: R. v. G.F., 2021 SCC 20, [2021] 1 S.C.R. 801, at para. 100.
[8] I acknowledge that the appellant’s uneven scrutiny ground of appeal carries a high threshold. As noted by the Supreme Court, per Karakatsanis J. for the majority, in G.F., at para. 99, this argument commonly features in judge-alone trials where the result turns on the trial judge’s credibility assessments of the evidence given by the complainant and the accused. Appellate courts have noted it is a “notoriously difficult argument to prove” because “[c]redibility findings are the province of the trial judge and attract significant deference on appeal”: at para. 99. As this court observed in Kiss, at para. 83, “courts are justifiably skeptical of what may be veiled attempts to have an appellate court re-evaluate credibility”.
[9] The persuasive burden for this argument was outlined by this court in Howe, at para. 59:
To succeed in this kind of argument, the appellant must point to something in the reasons of the trial judge or perhaps elsewhere in the record that make it clear that the trial judge had applied different standards in assessing the evidence of the appellant and the complainant.
[10] Or, stated another way: an appellant must persuade the appeal court that the trial judge relied on “a demonstrably flawed assessment methodology or reasoning process that affected the credibility determination": R. v. Roth, 2020 BCCA 240, 66 C.R. (7th) 107, at para. 50.
[11] The inquiry into whether there was uneven scrutiny is fact-specific: R. v. J.B., 2023 ONCA 264, at para. 26. While each case must be examined on its own facts, this ground of appeal appears to have succeeded in certain circumstances of analytical error in a trial judge’s assessment of the evidence.
[12] Relevant to the circumstances of the present case are the following instances of the trial judge’s uneven scrutiny of the evidence that, in Kiss, this court found amounted to an error of law and warranted a new trial:
The trial judge took “even the leanest opportunities to find reason to reject” the appellant’s evidence while being “uncommonly forgiving of similar and even more serious problems with [the complainant’s] testimony”: Kiss, at para. 84;
The trial judge “expected more of [the appellant] than he did of [the complainant]”, magnifying the importance of minor inconsistencies in the appellant’s evidence while treating as minor glaring inconsistencies in the complainant’s testimony on important issues: Kiss, at paras. 90, 96-103;
The trial judge gave “no adverse weight to the complications with [the complainant’s] evidence but instead treated her readiness to admit those difficulties as enhancing her credibility and reliability.” While the trial judge could note the candour of the complainant, he erred because he failed to weigh and consider “the impact of the difficulties she admitted” on her credibility and reliability: Kiss, at paras. 104-7.
[13] Again, relevant to the circumstances here, in other cases, appellate courts have found that similar examples constitute uneven scrutiny and reversible error:
A trial judge criticized the accused's evidence as “contrived” because of its detail but accepted the exact same level of detail in the complainant's version as a mark of credibility and reliability, without explaining why she drew this distinction: B.T.D., at paras. 58-61;
The trial judge rejected the accused's evidence for a reason that applied equally to the complainant and gave intense scrutiny to minor inconsistencies in the accused's evidence while significant inconsistencies in the complainant's evidence were not critically assessed: R. v. E.S. (2002), 2002 CanLII 49353 (ON CA), 158 O.A.C. 241 (C.A.), at paras. 8-11;
The trial judge failed to conduct a critical assessment of testimonial weaknesses that could undermine the Crown's evidentiary foundation on an essential element of the offence: Roth, at para. 142;
The trial judge made a negative credibility assessment of the accused’s testimony based on stereotypical assumptions and generalizations unsupported by the evidence: B.T.D., at para. 68; Roth, at para. 64-73;
The trial judge required the accused to explain himself, thus reversing the burden of proof: B.T.D., at para. 67;
The trial judge relied on minor or collateral inconsistencies in the accused’s evidence to reject his credibility while ignoring issues with the complainant’s evidence that go to the heart of her credibility and reliability: R. v. Willis, 2019 NSCA 64, 379 C.C.C. (3d) 30, at paras. 43, 61; and
The trial judge’s treatment of consistencies was inconsistent: despite relying on the consistency of the complainant’s testimony with the testimony of other witnesses, she failed to take into account other pieces of evidence that were consistent with the accused’s testimony, in particular that the accused’s testimony aligned very well with the testimony of others: R. v. M.P.H., 2022 BCCA 216, at paras. 47-56.
[14] In such cases, even if the evidence at trial was capable of supporting a conviction, where the trial judge has applied different standards to the assessment of prosecution and defence evidence, the appellant has not received a fair trial and thus has been the victim of a miscarriage of justice: B.T.D., at para. 56, citing R. v. Gravesande, 2015 ONCA 774, 128 O.R. (3d) 111, at para. 43.
b. Principles applied
[15] The appellant has met this high threshold. The trial judge’s flawed assessment of the evidence fatally tainted all four convictions. I start with some general comments about the trial judge’s approach, as they frame my analysis.
[16] The incidents underlying the four counts were discrete. Although it was open to the trial judge to make general assessments of the credibility of the complainant and the appellant, those assessments could not replace the analysis required in relation to each count: the trial judge was still required to determine whether the Crown had proved each count beyond a reasonable doubt. The trial judge’s reasons do not reveal that analysis. The difficulty is not that the trial judge made general credibility assessments or relied on them in considering the individual counts. Rather, as I explain below, those general assessments were themselves infected by uneven scrutiny, and that uneven scrutiny carried through into her assessment of the evidence relating to each count. Those assessments became the rigid analytical lens through which the trial judge then determined the appellant’s guilt on each count.
[17] The trial judge found the complainant to be “a very candid and forthcoming witness” and “very credible”. The trial judge acknowledged, without analyzing, “the flaws in the complainant’s evidence”, but she found “that none are so severe as to lead the court to entirely reject [the complainant’s] version of events” and, also without analysis, that the complainant “was relatively unshaken in her core narrative of the various events.” Among other reasons for finding the complainant credible, the trial judge stated that:
She did not gratuitously slam the defendant’s character. She acknowledged facts which did not further her version of events, such as unlocking her bedroom door from time to time and permitting [the appellant] to return to the master bedroom to cuddle. She admitted to incidents where she picked up objects so that [the appellant] would leave her alone including spilling shoes down on him. She did not attempt to minimize her own shortcomings.
[18] The above passage illustrates the trial judge’s uneven treatment of the evidence. The trial judge treated the complainant’s admissions against interest as enhancing her credibility, yet treated comparable admissions by the appellant differently, without explaining the difference.
[19] The trial judge found that the appellant’s evidence was not worthy of belief and did not raise a reasonable doubt. As I set out in more detail below, she “utterly” rejected his evidence with respect to the incident underlying the first count of sexual assault, describing his testimony as “incredible”. She characterized as “Nonsense!” his evidence that he was not angry with the complainant on the day he claims she slapped and kicked him in front of his children, on the day he claims she bloodied his nose with a plastic laundry basket, when he says she attacked him with a mirror, when she admitted to cheating on him nor when she kicked him out of their vehicle and he had to walk home. Further, the trial judge noted that he “constantly added that his wife smoked marijuana, even when it was unrelated to the question posed”, attempting to “besmirch her character by so doing.” Importantly, the Crown did not ask the trial judge to consider the appellant’s references to the complainant’s marijuana use as attempts to besmirch her character, nor did she raise this issue with counsel. The trial judge found that “[h]e portrayed himself as a peacemaker with Christian values and minimized his own emotions of anger and betrayal”, specifically noting that, “in cross-examination when asked about being angry enough to break the door on June 29, he said it was soft wood!”
[20] The difficulty with the trial judge’s generalized reasoning is that it led her, without examining the appellant’s evidence concerning the individual counts, to summarily reject his evidence whenever it conflicted with the complainant’s evidence, simply because it conflicted. She cursorily dismissed aspects of his evidence as “[n]onsense”, “concocted” or “bizarre”, without meaningfully engaging with his explanations, including his evidence that he regarded himself as a peacemaker, and then used that conclusion as part of generally adverse assessment of his credibility. This approach contributed to the trial judge’s rejection of the appellant’s evidence except where it served to corroborate the complainant’s account and to ignore where the complainant’s account served to the appellant’s evidence.
[21] The trial judge’s approach was flawed and represents reversible error. As I earlier stated, while the trial judge could make general credibility assessments, those assessments could not replace the independent analysis she was required to carry out in relation to each count. Nor could they become a fixed lens through which the evidence relating to each count was assessment. By inflexibly assessing the complainant’s credibility as unfailingly positive and the appellant’s as unrelentingly negative, the trial judge subjected their evidence to uneven scrutiny. She failed to grapple with material difficulties in the complainant’s evidence while relying on perceived difficulties in the appellant’s evidence to reject his testimony. As I explain below, that uneven approach affected her assessment of each count.
i. Count Three – sexual assault conviction
[22] I address first the trial judge’s erroneous treatment of the evidence related to the conviction on count three – the sexual assault on the porch of the complainant’s house. While out of chronological order as to when the underlying incidents are said to have occurred – this incident was said to have occurred in May or June of 2019 – it is important to start with this count because the trial judge’s analytical errors here fatally infected her credibility assessments of the complainant and the appellant. In particular, the trial judge’s treatment of this incident was central to her overall rejection of the appellant’s testimony as unworthy of belief.
[23] There was no issue that the appellant ejaculated on the porch in front of the complainant; however, there was a serious conflict in the evidence as to how it occurred that the trial judge needed to critically examine. The complainant testified that the appellant left the home during an argument at around 2:00 a.m. before he returned ten minutes later and committed this act, ejaculating on her in close proximity without her consent. The appellant denied that he ejaculated on the complainant and maintained that the complainant taunted him to masturbate, which he did at some distance away, while she was smoking marijuana on the porch. The complainant denied that she was smoking marijuana on the porch on that occasion, although she admitted to smoking marijuana on other occasions.
[24] The trial judge accepted the complainant’s evidence, rejected the appellant’s account as “concocted”, and found he attempted to “besmirch” the complainant’s character by “constantly” referring to her smoking marijuana. The trial judge’s approach to the evidence was materially flawed in that she evaluated the appellant’s evidence through an overly critical lens and engaged in uneven scrutiny of his and the complainant’s evidence.
[25] First, the trial judge allowed her strong disapproval of the appellant’s conduct to influence her assessment of his evidence concerning how the incident occurred. She rejected the appellant’s evidence about how the event unfolded based on her censure of his masturbating on the porch: “The whole idea of conducting himself this way in public, even if late at night, on the open veranda of a small semidetached house in a residential area, is absurd. Yet on the evidence of both witnesses, it occurred.” Her clear disapproval of the appellant’s act closed her mind to considering whether the complainant had egged him on to do it, as he testified, and whether the complainant’s evidence at all corroborated what the appellant said.
[26] Instead, the trial judge’s strong disapproval of the act as a “dehumanizing episode” led her to enhance the complainant’s credibility and characterize the appellant’s evidence as to how it happened as “bizarre” and “concocted”. The trial judge made this finding of concoction without any explanation aside from her disapproval of the act and her positive assessment of the complainant. The appellant’s explanation was no more incredible than the complainant’s evidence about the incident, which the trial judge described as “the most bizarre of the allegations”. That the trial judge’s reasoning was unduly affected by her positive assessment of the complainant is also plain from her use of the appellant’s evidence to corroborate the complainant’s account. Even though she indicated she rejected the appellant’s testimony about the incident, the trial judge was nevertheless prepared to accept the appellant’s testimony where it corroborated the complainant’s evidence and served to enhance her credibility.
[27] Further, the trial judge erred by viewing the appellant’s evidence that the complainant was smoking marijuana as an attempt to “besmirch” her character. The complainant herself acknowledged using marijuana on other occasions. There was no basis to treat the appellant’s reference to her marijuana use, without more, as an attempt at character assassination. As earlier noted, the Crown did not ask the trial judge to treat the appellant’s reference to the complainant’s marijuana use as an attempt to denigrate the complainant, nor did the trial judge raise this concern with counsel.
[28] More importantly, the appellant’s evidence was potentially relevant to the competing accounts of what occurred: the appellant maintained that the complainant was smoking marijuana on the porch, which formed part of his explanation for her presence there at 2:00 a.m. Rather than considering whether that evidence supported his account, the trial judge used it as a reason to diminish the appellant’s credibility.
[29] Finally, the trial judge gave no credit to the appellant for admitting the ejaculation against his interest, while, as earlier indicated, in her general assessment of the complainant’s credibility, she used the complainant’s admissions against interest to enhance her credibility.
ii. Count Two – conviction for assault with a weapon
[30] This conviction resulted from a fight between the complainant and the appellant involving a laundry basket. This incident was said to have occurred at the end of May 2019.
[31] The complainant testified that, in the midst of a loud argument with the appellant, she threw a laundry basket at him, and he then hit her in the arm with the laundry basket. The appellant testified that the complainant hit him in the face with a laundry basket, bloodying his nose, and that he then used the laundry basket to protect himself from the complainant who was trying to hit him with a mirror. The complainant testified that the basket had scraped her arm and elbow and produced at trial a photograph that she had taken of it. In her statement to police, the complainant did not tell the police that her finger was allegedly broken. That evidence did not emerge until trial. Unlike the scraped arm, there was no photograph of her allegedly broken finger nor was there supporting medical documentation, although she testified that she told her doctor about it.
[32] There is no analysis of the appellant’s evidence with respect to this incident other than the trial judge’s cursory, generalized rejection of the appellant’s evidence because she thought his testimony regarding his lack of anger with respect to the laundry basket incident, as with the others, was “Nonsense!”. She seems to have accepted the complainant’s evidence about this incident solely because of the photograph she produced of her scraped arm and elbow. This was in keeping with the trial judge’s general approach that “[w]here the evidence of [the complainant] is either corroborated, or there is consistent evidence, it may augment the weight to be given to it. Some examples are the photographs of injuries consistent with her testimony.” There is no other analysis. The trial judge erred in failing to do so.
[33] Importantly, with respect to this incident, the trial judge failed to reconcile important issues with the complainant’s evidence. Most significantly, the trial judge did not address the complainant’s evidence that her finger was broken, an allegation that was not mentioned to police and emerged only at trial. She was required to consider whether this late-emerging evidence represented an inconsistency or exaggeration and, if so, what effect it had on the complainant’s credibility and reliability.
iii. Count One – sexual assault conviction
[34] The sexual assault conviction on count one related to sexual relations that occurred between the complainant and the appellant in her bedroom in late June 2019. There was no dispute that the sexual relations had occurred. The real question was whether they were consensual.
[35] The complainant testified that the appellant forced sexual intercourse with her while she attempted to resist him. She additionally testified, at first, that at this point in the relationship, the appellant was sleeping in the basement because she was sleeping with the bedroom door locked. However, in cross-examination, she conceded that it could have been that the appellant was already sleeping and cuddling with her in her bedroom that evening. She explained that it could have been one of the times she consented to sleeping in her bed with him, but she denied consenting to sexual intercourse. The appellant testified that their sexual relations were entirely consensual.
[36] The trial judge erred by failing to analyze the inconsistencies in the complainant’s evidence about how the appellant came to be in her bedroom that evening. Specifically, she did not analyze whether and how the complainant’s initial, more inculpatory account of the encounter, which she corrected on cross-examination, affected her credibility about the incident and whether it corroborated in any way the appellant’s evidence.
[37] The trial judge accepted the complainant’s evidence for two problematic reasons. First, she erred by accepting the complainant’s general statement that she did not want to have sexual relations with the appellant during their separation, rather than the specific evidence related to the count. Second, she incorrectly imported into count one her problematic findings on count three: “His total disregard of her feelings and her dignity a short time prior when he ejaculated on her escalated to this act when he was yet again denied what he felt was his spousal right. I believe [the complainant] and accept that is what happened.”
[38] Finally, she erred by failing to address the appellant’s evidence that the sexual relations were consensual or explain why she rejected it, aside from her reference to her findings on count three.
iv. Count Four – assault conviction
[39] The conviction on count four arose out of another dispute between the complainant and the appellant. Chronologically, it was first in time of the four counts, occurring at the end of December 2018, after the complainant confessed to infidelity. She testified that the appellant was upset and slapped her thigh, leaving a handprint. She took a photograph of her thigh and produced it at trial. The appellant testified that the complainant first slapped his face in front of his children and then kicked him; he slapped her leg to stop her from kicking him.
[40] There was no meaningful analysis of the evidence, especially the appellant’s evidence. The trial judge appears to have rejected it simply because she saw his lack of anger as: “Nonsense!” The trial judge’s stated reason for accepting the complainant’s evidence was because of the photograph that she produced at trial. There is no other analysis. The trial judge erred in failing to do so.
E. Disposition
[41] For these reasons, I would allow the appeal, quash the convictions and order a new trial.
Released: August 27, 2026 “M.T.”
“L.B. Roberts J.A.”
“I agree. M. Tulloch C.J.O.”
“I agree. M. Rahman J.A.”
1This appeal is subject to a publication ban pursuant to s. 486.4 of the Criminal Code, R.S.C. 1985, c. C-46.