The Youth Justice and Criminal Evidence Act 1999 (YJCEA 1999) s 41, with its narrow gateways and strict prohibition on using sexual behaviour evidence (SBE) as relevant to credibility, has induced migraines in counsel and judges from the outset. It has never been amended, apart from the early, crucial judicial interpolation of the ‘Article 6 gloss’ for the defendant’s (D’s) fair trial rights (R v A (No 2) [2001] UKHL 25).

On 3 December 2025, the Ministry of Justice announced its intention to ‘stamp out rape myths’ and ‘stand up for rape victims’ (sic) who will ‘receive a fairer trial’, by amending the Criminal Justice Act 2003 s 100 and YJCEA 1999 s 41. The statement is baffling. Section 41 will now require ‘a higher admissibility threshold’ for ‘genuinely valuable’ evidence. But what legal language could construct barriers higher in the stratosphere than the current cumulative tests that the evidence:

  • must not impugn the complainant’s (C’s) credibility (s 41(3) and (4));
  • relate to a ‘relevant issue’ which the prosecution or defence is required to prove (s 42(1)(a));
  • be so important that refusal of leave might render the jury's conclusion on the relevant issue ‘unsafe’ (s 41(2)(b)); and
  • refusal of leave would deprive D of the right to a fair trial?

These are probably the most stringent admissibility tests found in criminal evidence law. Section 41 is also impenetrable, often misses the relevance and probative value marks, and should be scrapped and redrafted (Hoyano, ‘Cross-Examination of Sexual Assault Complainants on Previous Sexual Behaviour: Views from the Barristers' Row’ [2019] Crim LR 77).

The Law Commission of England and Wales has grasped this nettle in Evidence in Sexual Offences Prosecutions: a Final Report (Law Com No 420, July 2025). While its intentions for demolition and reconstruction are laudable, as is the preference for a structured judicial discretion regime, its preferred model cannot work. It conflicts with the legal context in which it must be applied.

The starting point of the structured discretion model, echoing the Criminal Code of Canada (CCC) s 276, first enacted in 1992, is a prohibition on evidence which risks perpetuating the ‘twin myths’ identified in Seaboyer [1991] 2 SCR 577 and endorsed in A (No 2), viz, where the SBE is tendered to support an inference that, by reason of the sexual nature of that activity, C is more likely to have consented, or is less worthy of belief. They become the immediate, constant focus of the admissibility decision, against which the court works through a detailed analytical framework to identify legitimately relevant and probative evidence. Now, s 41 does not identify the risks of SBE.

Sound recommendations include that the protection be extended from sex offences to any case where SBE might be tendered, e.g., domestic abuse, and from Cs to other witnesses (and, perhaps problematically, to homicide victims).

The difficulties arise with the Law Commission’s preferred structure of judicial analysis, rejecting s 276’s balancing formula that the SBE has ‘significant probative value that is not substantially outweighed by the danger of prejudice to the proper administration of justice’. Weighing probative value against prejudicial effect is familiar to English jurists, being deployed throughout the law of evidence.

The Law Commission claims that the Canadian approach is fundamentally flawed, eschewing any balancing of the interests of C, D and the administration of justice. Instead, there should be two separate, disconnected analytical steps (recommendation 14):

  • Stage 1: the evidence has substantial probative value in relation to a matter in issue in the proceedings, which is of substantial importance in the context of the case as a whole; and
  • Stage 2: its admission would not significantly prejudice the proper administration of justice.

Each stage lists familiar factors. They cannot be transported from one list to the other, but some seem connected. Stage 1(a) ‘evidence holding substantial probative value’ is fundamental to stage 2(b) ‘defendant’s right to a fair trial’. The risk of ‘myths and misconceptions’ appears on both lists, double-weighting against D’s interests.

The Law Commission argues that C’s Article 8 rights should not be weighed against the interests of justice, as it ‘risks reducing the complex issues to a binary approach, pitting the defendant’s right to a fair trial against the complainant’s right to privacy, in which the right to a fair trial will always prevail’ (para 3.48). They claim that a ‘further benefit’ of the two-stage test is that ‘it encourages consideration of the complainant’s article 8 right as a relevant factor in its own right, not only where it is in direct competition with the defendant’s right to a fair trial under article 6’ (para 3.49).

Both claims are perplexing. The Law Commission reiterates that the two-stage test will operate to ensure that whenever SBE is necessary for a fair trial, it will be admissible (paras 3.36, 3.43, 3.49, 3.65, 3.71). Yet the stated rationale for separating the tests is that D’s right to a fair trial should not always trump C’s privacy interest – confirmed by D’s fair trial right being designated merely a discretionary factor weighed at stage 2, not as the ultimate, determinative admissibility test, as now under A (No 2).

Moreover, the notion that Article 8 is a distinct right to be determined separately from D’s rights is belied by its definition. The scope of the Article 8(1) right is circumscribed by Article 8(2), that any interference must be ‘in accordance with the law and is necessary in a democratic society’, inter alia, ‘for the prevention of disorder or crime… or for the protection of the rights and freedoms of others.’ Hence it is a qualified right, delineated by the interests of criminal justice, and by D’s rights – including the absolute fair trial right under Article 6.

Critics of the consultation paper argued that it was impossible to conceive of a case where SBE could pass stage 1 (substantial probative value) but fail stage 2 (significantly prejudice the administration of justice), so it was impossible to evaluate them separately. To rebut this objection in their final report, the Law Commission devised the following scenario (reproduced as drafted, emphasis added):

The complainant has alleged the defendant violently raped her. The defence’s case is that it was consensual rough sex, that the complainant enjoys rough sex but is now lying because other people found out and she is embarrassed by that. The complainant made a statement that she does not enjoy engaging in rough sex and would not consent to it. The defence wish to adduce two pieces of SBE to dispute this. The first is a series of text messages between the complainant and the defendant where they mutually discuss engaging in consensual rough sex highly similar to the acts in the alleged rape. The second is a video recorded by a previous partner. In the video the complainant is seen engaging with the previous partner in highly similar rough sex. The complainant’s case is that in neither situation was she truly consenting although accepts they both appear to show she does. Both pieces of SBE could have substantial probative value in relation to whether the prosecution’s case that the complainant does not enjoy, and therefore would not consent to, rough sex, is credible. However, the video of her with the previous partner has a much higher risk of causing unjustified distress and humiliation to the complainant. Further, as the complainant denies she was consenting in that video, it risks distracting the jury with a satellite issue, as well as necessitating disproportionate court time in establishing consent in an event involving third parties that is unconnected to the matter being tried. The judge could conclude that it would prejudice the administration of justice to admit the video. Under the test we recommend, the court could therefore exclude the video, and admit the text messages. This would not infringe the defendant’s right to a fair trial. 

The multiple problems with the Law Commission’s analysis will be evident to any RASSO [rape and serious sexual offences] barrister. Some facts lack plausibility. In what is incorrectly described as ‘the complainant’s (sic) case’, she claims that her text exchanges with D were not voluntary – but it would have been easy to ignore them, blocking D. Setting that aside, the Law Commission’s conclusion that the texts have substantial probative value is supported by R v Reimer 2024 ONCA 519. The Ontario Court of Appeal held that messages expressing wishes to engage in specific sexual acts, while not admissible to establish advance consent, may yield a ‘logical, permissible inference’ that C’s stated intention to engage in consensual sexual activity with D logically tends to increase the likelihood that she did consent on the occasion charged; this inference would ‘not trade on sexism, and it is not built on myths about female sexuality’.

Most significantly, the scenario contains three errors of law regarding the video.

First, the Law Commission overlooks the essential element of the rape offence which the Crown must prove, lack of reasonable belief in consent (Sexual Offences Act 2003 s 1). The video is self-evidently highly probative, since C concedes that it appears to show her consenting to ‘highly similar’ rough sex, which could corroborate D’s account of his reading of her reaction. Yet the Law Commission asserts, without analysis, that this would cause C ‘unjustified’ distress and humiliation. What level is justified? If the evidence holds great probative value regarding a fact in issue central to D’s defence, then the level of distress surely is irrelevant – if, per McLachlin J in Seaboyer (at 621), the shield’s purpose is to find the truth, and the correct verdict (also Goldfinch 2019 SCC 38 at [81]).

Second, the Law Commission asserts that exclusion is warranted because it would necessitate ‘disproportionate court time in establishing consent’ to a third-party event, requiring ‘satellite litigation’. This ignores the basic collateral issue rule, devised to forestall such extended diversions from events alleged on the indictment, if C denies it in cross-examination. The Law Commission concludes that the court could determine that this ‘substantially probative’ video evidence would prejudice the administration of justice and exclude it as not prejudicing the defendant’s right to a fair trial. So, this justifies the bifurcated approach. Neither line of reasoning is sustainable in law.

Finally, we are not informed of the crucial question whether the video was filmed consensually or surreptitiously, which would constitute the offence of voyeurism, nor how the defence came to possess it. If obtained in a criminal breach of C’s rights, there would be a powerful argument under the Canadian balancing exercise that its probative value is outweighed by damage to the administration of justice by condoning an intrusive criminal offence through admitting its fruits into evidence, thereby incentivising illicit recording. The Law Commission fails to identify this question, notwithstanding it is the first a trial advocate would investigate.

Therefore, subject to settling the illegality point, the video would have to be admitted in the interests of justice, as relevant to the credibility of D’s position that he reasonably believed from C’s conduct that she consented.

Thus, the Law Commission’s posited example proves the opposite of its contention that a compartmentalised two-stage framework is necessary for proper adjudication of the competing rights of D and C.

A further difficulty with the Law Commission’s compartmentalised analysis is that it allocates the court’s consideration of the statutory factors into two separate channels. This risks committing the category errors of s 41, where Parliament in 1999 hubristically predicted every circumstance in which SBE could legitimately be relevant. Unlike CCC s 276(3)(h), the Law Commission’s lists of factors are closed, not permitting consideration of any additional factor which might be raised by the case.

The Law Commission’s decision to depart from the Canadian model it claims to emulate is unjustified on its own terms. If Parliament prefers the Law Commission’s model, questions may arise whether Canada’s large corpus of reasoned jurisprudence, determined with a clear vision of the human rights of C and D (as the Law Commission desires), can be used in training judges and advocates, or to inform judgments as English jurisprudence starts to develop. It would be regrettable to exile this source of invaluable guidance from the Judicial College, the Inns of Court which train barristers, and our courtrooms.