Fresh Evidence of Complainant’s Dishonesty and Unsafe Rape Convictions: A Commentary on R v Stephenson [2025] EWCA Crim 1600

1. Introduction

This commentary analyses the decision of the Court of Appeal (Criminal Division) in R v Stephenson [2025] EWCA Crim 1600, a Criminal Cases Review Commission (CCRC) reference concerning a 2015 rape conviction. The appeal was heard by the Court of Appeal (Criminal Division), with Lord Justice Edis delivering the judgment.

At the core of the case lies a difficult and sensitive issue: how an appellate court should respond when, years after a conviction, the complainant in a sexual offence case admits to having lied about highly material aspects of the surrounding circumstances, whilst still maintaining that serious sexual offences were in fact committed against them.

The judgment is particularly important for:

its application of the principles governing fresh evidence on appeal, especially where a complainant has lied;
its treatment of credibility in consent-based sexual offences, where the complainant’s account is central;
its interaction with earlier decisions restricting evidence of sexual behaviour under s.41 Youth Justice and Criminal Evidence Act 1999;
its implications for CCRC practice, prosecutorial duties of disclosure, and the systemic handling of potential miscarriages of justice.

Although the Court did not claim to establish new doctrine, it provides a clear and concrete illustration of how existing principles—especially those in Dial and Dottin v State of Trinidad and Tobago [2005] UKPC 4—are to be applied when a complainant is later shown to have been deliberately untruthful about matters central to consent.

2. Factual and Procedural Background

2.1 The original conviction

In June 2015, at the Central Criminal Court, Jerome Stephenson (then in his 30s, now 43) was convicted by a majority jury verdict of one count of rape of a woman anonymised as “LW”, contrary to s.1 Sexual Offences Act 2003. In July 2015 he was sentenced to 12 years’ imprisonment.

Under the early release provisions then applicable, he was released from custody on or about 17 June 2021, but would have remained on licence until 17 June 2027 had his conviction stood. The Court notes that the quashing of the conviction brings that licence to an end.

The complainant LW is protected by lifetime anonymity under the Sexual Offences (Amendment) Act 1992. The Court restates that no information may be published which is likely to identify her as the alleged victim.

2.2 The alleged incident in 2013

The events concern a night in 2013 when LW was in a property with:

the appellant, Jerome Stephenson;
Byron Smith; and
a third male.

LW alleged sexual offences against all three men. According to her trial account:

from the outset she was in a threatening environment;
she was plied with alcohol and drugs which made her sick;
she was forced to watch pornography on a laptop;
the men used some violence, and then raped her on multiple occasions;
she was hauled upstairs by all three, prevented from leaving, and raped repeatedly over a couple of hours;
she made it clear throughout that she did not consent but was overborne by violence and the effect of drugs;
after Byron Smith and the third man left, the appellant (Stephenson) raped her.

Stephenson’s defence at trial was that the vaginal intercourse found by the jury to be rape had taken place with LW’s consent. Thus, the jury’s assessment of LW’s credibility was critical.

2.3 The earlier appeal in 2017

Stephenson had previously appealed against conviction. That appeal was dismissed on 9 February 2017: see R v Stephenson [2017] EWCA Crim 338.

The 2017 appeal:

challenged the exclusion of certain material under s.41 Youth Justice and Criminal Evidence Act 1999;
argued that material relevant to LW’s credibility had been wrongly kept from the jury;
sought to rely on fresh evidence to show that LW was a fantasist, but the relevant witness refused to attend court.

The Court of Appeal in 2017:

held that s.41 had been correctly applied at trial; and
found that the proposed fresh evidence would have carried little weight, given the unwillingness of the witness to attend.

Crucially, that court “had no inkling” of the fresh evidence later considered in 2025. That material did not exist until six years after the 2017 appeal.

2.4 CCRC involvement before 2025

After his 2017 appeal failed, Stephenson applied to the CCRC in 2017, again focusing on LW’s credibility. That application was ultimately rejected by the CCRC in a final decision dated 25 February 2021: the case was not referred back to the Court at that time, because the new evidence now in issue had not yet emerged.

2.5 The Byron Smith prosecution and LW’s new account (2023)

A separate prosecution was subsequently brought against Byron Smith in relation to the same night in 2013. Smith was, much later, charged with:

four counts of raping LW; and
one count of assault by penetration.

Smith’s trial was due to begin in September 2023. In preparation for that trial:

LW was shown a recording of her original Achieving Best Evidence (ABE) interview;
on 14 September 2023 she made a new witness statement admitting that she had lied in important respects in her evidence at Stephenson’s trial;
given the gravity of these admissions, a further video-recorded interview took place on 27 September 2023, in which LW explained her falsehoods and omissions and her reasons for them.

As a result:

the prosecution, advised by leading counsel Miss Knight KC, concluded that LW could not be relied on as a truthful witness in the proceedings against Smith;
the proceedings against Byron Smith were discontinued and not-guilty verdicts followed in March 2024.

At the same time, the prosecution disclosed LW’s 14 September statement and the 27 September interview to Stephenson’s legal representatives, so that a further CCRC review could be sought. The CCRC, on the basis of this fresh evidence, referred the case back to the Court of Appeal.

3. Summary of the 2025 Judgment

The Court of Appeal:

gave leave to adduce the fresh evidence; and
formally admitted LW’s 2023 statement and video interview as evidence.

The prosecution, having already abandoned the prosecution of Smith, did not oppose Stephenson’s appeal. They accepted that the jury in 2015 might well have reached a different verdict had they heard LW’s new account.

The Court:

emphasised that the question whether a conviction is “unsafe” is a matter for the court alone, even where the prosecution does not oppose the appeal;
applied the principles governing fresh evidence, as articulated in Dial and Dottin v The State of Trinidad and Tobago [2005] UKPC 4;
concluded that LW had told deliberate lies and made significant omissions at the original trial about matters central to the issue of consent;
found that the fresh evidence would have presented “a different picture” of the events, particularly the early “party atmosphere” and LW’s own actions in instigating or participating in sexualised activity;
held that the conviction was therefore unsafe and must be quashed;
noted that no retrial was sought, and declined to order one.

While LW has not retracted her allegations of serious sexual offending, the Court recognised that the new material fundamentally affects how a jury might assess the credibility of her account of non-consent. The appellant’s conviction is quashed, and the proceedings are concluded.

4. Legal Framework

4.1 Complainant anonymity in sexual cases

At the outset, the Court reiterates the effect of the Sexual Offences (Amendment) Act 1992:

where an allegation has been made that a specified sexual offence has been committed against a person,
no matter relating to that person shall be included in any publication during their lifetime,
if it is likely to lead members of the public to identify that person as the victim of the offence.

The anonymity applies to LW despite the conviction being quashed; it is tied to the allegation, not to a conviction. The Court adds that there are no other reporting restrictions in place, so the proceedings can be reported in full subject only to LW’s anonymity.

4.2 Appeals and fresh evidence

The Court is applying the usual criminal appeal test: whether the conviction is “unsafe”. Where fresh evidence is relied upon, appellate courts generally consider (under the Criminal Appeal Act 1968):

whether the evidence could with reasonable diligence have been adduced at trial;
whether it is credible; and
whether, taken with the evidence at trial, it might reasonably have led to a different verdict.

Here, the Court:

accepts that the evidence is genuinely “fresh” (it did not exist until 2023);
regards LW’s admissions of lying and her new narrative as significant and credible enough to be considered;
asks whether this new material renders the 2015 conviction unsafe.

4.3 Evidence of sexual behaviour and s.41 YJCEA (context)

Although s.41 of the Youth Justice and Criminal Evidence Act 1999 is not directly re-applied in this 2025 judgment, it provides important context. Section 41:

generally prohibits evidence or cross-examination about a complainant’s sexual behaviour,
subject to narrow exceptions where such evidence is necessary to ensure a fair trial,
and requires the court’s leave before such evidence may be adduced.

In 2017, the Court had held that s.41 was correctly applied in Stephenson’s trial, rejecting arguments that the defence had been improperly prevented from exploring material relating to LW’s credibility and sexual behaviour. The 2025 judgment, however, reveals that LW herself later admitted to having lied about:

her own instigation of sexualised behaviour at the time;
her role in putting on pornography;
her involvement with an escort agency and prostitution; and
her request that condoms be used.

Some of this would likely fall within the ambit of s.41, illustrating a tension between the protective policy of the provision and the need to present the jury with a full and truthful picture where issues of consent are finely balanced. The 2025 Court does not criticise the earlier s.41 rulings but shows that, had the truth been told by LW, the evidential landscape would have been markedly different.

5. Analysis of the Court’s Reasoning

5.1 The central precedent: Dial and Dottin v State of Trinidad and Tobago

The Court cites Dial and Dottin v The State of Trinidad and Tobago [2005] UKPC 4 as the key authority. Lord Brown there stated (at [42]):

“Wherever fresh evidence establishes that a material prosecution witness has told a lie, the question arising for the Appeal Court's determination is whether that realistically places the appellant's guilt in reasonable doubt – whether, in other words, the verdict is now to be regarded as unsafe. That necessarily must depend upon all the evidence in the case. However barefaced the lie and however central to the prosecution case the witness who told it, the Court of Appeal is bound in law to address that question. Even in a case of capital murder it cannot be right to allow an appeal, without more, simply on the basis that the State's main witness has later been shown to have told an outright lie.”

The key points drawn from this authority are:

Evidence that a prosecution witness has lied, even outright, does not automatically render the conviction unsafe;
The appellate court must consider the entire evidential picture and ask whether the fresh evidence realistically introduces reasonable doubt;
The decision is not mechanical or driven merely by the existence of the lie; it is a contextual assessment.

The Court of Appeal in Stephenson explicitly adopts and applies this approach. It does not treat LW’s admissions of lying as automatically fatal to the conviction, but evaluates the materiality and effect of those lies on the core issue of consent.

5.2 The nature and scope of LW’s lies and omissions

LW’s 2023 evidence significantly altered the narrative she had given at trial. The Court notes several major changes, many of which LW accepts were deliberate lies:

(a) “Party atmosphere” versus coercive setting

At trial, LW described an atmosphere of:

fear, oppression, dominance and coercion;
forced consumption of drugs and alcohol that incapacitated her;
being compelled to watch pornography.

In 2023, LW gave a strikingly different picture of the early part of the evening, describing:

knowingly and willingly taking drugs and alcohol;
a “party atmosphere” where she was “high and getting attention” from the men;
behaviour which was “flirtatious” and which she “was enjoying” at that stage;
possibly dancing seductively, and consciously behaving in a sexually provocative manner.

She admitted that these aspects of her behaviour had been deliberately “edited out” of her earlier account because she feared the jury would think she had “led them on”.

(b) Instigation of pornography

At trial, it was put to LW that she had put pornography on the computer; she denied this. In her 2023 statement she admitted:

“I did actually put pornography on, on the computer. I was asked to put the porn on, and I did. They asked me what porn I liked and I said that I liked Latino girls.”

This is a direct contradiction of her trial evidence and supports the defence narrative that she was an active participant in initiating sexualised activity at the early stage.

(c) Physical intimacy: sitting on Stephenson’s lap

The appellant’s case at trial was that LW had voluntarily come to sit on his lap while pornography was playing and that she was grinding on him. LW denied this at trial.

In 2023, she accepted that:

she had indeed sat on a man’s lap, probably Stephenson’s;
she had been “grinding” on him; and
everyone was becoming aroused.

Again, this admission directly supports the defence’s description of a mutually sexualised interaction at an early stage, contradicting LW’s earlier portrayal of herself as purely passive and coerced.

(d) Use of condoms

At trial, LW denied that she had asked the men to use condoms, which could be taken as indicating that the activity was entirely non-consensual. In 2023 she admitted that this denial was a lie:

she accepted that she had asked for condoms to be used;
she explained that she lied at trial because she feared the jury would take this as a sign of consent.

The Court treats this as significant. Requests for protection are often highly probative in consent disputes: they can be interpreted as an indication of some level of participation or agency, depending on the wider context.

(e) Escort agency and prostitution

There was trial evidence about potential contact with an escort agency. The appellant had said:

they were searching for local escorts; and
LW said she knew somebody who could come and made a call.

At trial LW insisted that she had been told by the appellant to order an escort, denying that it had been her idea.

In the 2023 video interview, LW stated:

that she told Stephenson she was a prostitute; and
that she may have suggested calling the agency she herself worked for to procure another woman.

This is again directly supportive of defence assertions made at trial and again undermines her earlier denials under cross-examination.

(f) Overall assessment of these changes

The Court concludes that these changes are “significant” and that LW accepts that at least some of them were deliberate lies, not mere misremembering. Importantly, all of the changes relate to:

the early phase of events,
the level of voluntariness and mutuality in sexualised behaviour, and
her own agency and participation.

They therefore go to the heart of the issue at Stephenson’s trial: whether LW consented to sexual activity with him, at least initially, and whether his account that sexual contact was consensual was credible.

LW, however, has not retracted her allegation that she was subjected to serious sexual offending. She continues to say that, at some point, the situation turned coercive and that she did not consent to the subsequent sexual activity. The Court is careful to record this, emphasising that to acknowledge her earlier dishonesty is not to declare that her ultimate allegation of rape was necessarily false.

5.3 Why the lies made the conviction unsafe

The Court’s core reasoning is found in the following key propositions:

LW’s credibility was pivotal to the issue of consent. The case was classic “word against word” on the central question of whether intercourse was consensual.
The lies related directly to matters on which:
- the defence had cross-examined LW at trial (suggesting that she had been flirtatious, instigated sexual contact, put on pornography, etc.); and
- she had given firm denials under oath.
The new evidence now confirms much of what the defence had asserted.
Upon LW’s own account in 2023, the factual picture at the outset would have been very different before the jury:
- a “party atmosphere”,
- mutually flirtatious conduct,
- her voluntary participation in drug-taking and pornography,
- her active sexualised behaviour (lap-sitting, grinding), and
- her role in calling an escort agency she worked for.
The Court accepts that LW had personal reasons for lying (avoiding moral blame, fear of being thought to have “led them on”), but this does not mitigate the forensic impact: the jury were misled about core aspects of the events and about LW’s reliability.
Had the jury heard LW’s present account, they might have evaluated the consent issue differently. The Court finds it untenable to argue that the jury would “undoubtedly” have convicted even with this evidence.

Thus, applying the Dial and Dottin test, the Court concludes that LW’s admitted dishonesty realistically places Stephenson’s guilt in reasonable doubt and that the 2015 verdict can no longer be regarded as safe.

5.4 Role of the prosecution’s stance

The prosecution, represented by Miss Knight KC, accepted that:

it could no longer rely on LW as a witness of truth in any proceedings against Stephenson;
the new material might well have impacted the jury’s verdict;
they therefore did not seek to oppose the appeal.

The Court is clear that such a concession is not determinative. It remains for the Court alone to decide whether a conviction is unsafe. However, the prosecution’s stance:

confirms the seriousness with which LW’s dishonesty was regarded by those presenting the Crown’s case;
reinforces the Court’s own view that the case could not responsibly proceed to a retrial; and
illustrates modern prosecutorial ethics in responding to fresh evidence undermining a key witness.

5.5 No retrial ordered

There is no application for a retrial, and none is ordered. The key reasons are implicit:

the prosecution itself accepts that LW cannot be presented as a reliable witness of truth;
without LW’s testimony, the Crown case would be unsustainable;
LW still maintains that she was the victim of serious sexual offending, but the necessary evidential underpinning for a safe conviction, particularly in terms of her overall reliability, is no longer present.

The Court thus allows the appeal and quashes the conviction, bringing the proceedings to a close.

6. Precedents and Previous Decisions

6.1 Dial and Dottin v The State of Trinidad and Tobago [2005] UKPC 4

As noted above, Dial and Dottin articulates the key principle governing appellate review when fresh evidence shows that a material prosecution witness lied. The Court of Appeal in Stephenson faithfully follows and applies this principle:

it does not treat proof of lying as automatically fatal to a conviction; but
it asks whether those lies, in the context of all the evidence, now render the verdict unsafe.

Stephenson is a practical application of Dial and Dottin, showing how the principle operates in a consent-based rape case where the complainant admits to having lied about the prelude to the alleged assault.

6.2 The earlier appeal: R v Stephenson [2017] EWCA Crim 338

The 2017 decision is important in two ways:

It illustrates that the Appellate Court had previously found no error in the application of s.41 YJCEA at trial and had rejected arguments that LW was a fantasist. At that time, there was no evidence that LW herself had lied about the central events.
It underscores that appellate and CCRC review are limited by the material then available. The 2017 court cannot be criticised for failing to detect what, in reality, only emerged in 2023 when LW saw her ABE interview again and decided to correct her account.

The 2025 judgment carefully distinguishes these earlier attempts to challenge LW’s credibility (which failed due to lack of sufficiently weighty evidence) from the genuinely fresh and powerful admissions she made in 2023.

7. Impact and Broader Implications

7.1 Sexual offences, credibility, and the centrality of context

Stephenson highlights the acute sensitivity of rape cases to:

the complainant’s credibility; and
the detailed context in which sexual activity begins and develops.

In many consent-based cases there is no significant forensic evidence (e.g. injuries or DNA pointing one way or the other). The verdict often turns on a comparative assessment of the complainant’s account and the defendant’s. In such cases:

deliberate misrepresentation by the complainant about the consensual or non-consensual nature of the early interaction is likely to be of great significance;
the jury must have the opportunity to hear the true picture—whether this supports the Crown or the defence—to carry out its task fairly.

The judgment recognises that a complainant may lie for reasons that do not necessarily mean the allegation of rape is false (for example, to avoid reputational damage), but insists that this cannot be ignored where those lies went directly to the issue of consent.

7.2 Section 41 YJCEA and evidence of sexual behaviour

Although the Court in 2025 does not reopen the earlier s.41 rulings, the judgment inevitably prompts reflection on:

the line between impermissible “sexual history” evidence and admissible contextual evidence directly relevant to consent on the occasion in question;
how far the operation of s.41 may prevent a jury from hearing about a complainant’s contemporaneous sexualised behaviour, where that behaviour forms part of the alleged incident.

Stephenson suggests that:

where a complainant has actively instigated or participated in sexualised conduct at the time of the alleged offence, concealment of that behaviour (or its exclusion by s.41) can substantially distort the jury’s understanding of what occurred;
courts and practitioners must scrutinise whether particular questions about behaviour are properly about “sexual history” (ordinarily inadmissible) or about the immediate factual matrix of the incident (which may be essential to a fair trial).

The case does not loosen s.41; rather, it underlines the importance of accurately distinguishing between prohibited and essential evidence, and the potential consequences when crucial contextual behaviour is not truthfully placed before the jury.

7.3 The role of the CCRC and systemic delay

Stephenson also illustrates both the value and the limits of the CCRC mechanism:

Stephenson sought CCRC review once in 2017. That application was refused in 2021, correctly based on the evidence then available. The CCRC cannot “create” evidence; it is dependent on what emerges.
Only after LW, in 2023, voluntarily corrected her earlier account did powerful fresh evidence arise, enabling the CCRC to refer the case.
By the time the conviction was quashed, Stephenson had served his entire custodial term. The Court rightly expresses “profound regret” at this outcome, but notes the timeline: the exculpatory material simply did not exist earlier.

The case may fuel debates about:

whether there are better ways to monitor or re-check critical witness testimony, especially where later prosecutions (such as that of Byron Smith) depend on the same account; and
how quickly the justice system can respond once potentially exculpatory information comes to light.

7.4 Prosecutorial duties, disclosure and ethics

The Court acknowledges and commends the prosecution, counsel, and police for their “entirely appropriate” reaction once LW’s lies were disclosed:

they reassessed the viability of the prosecution against Byron Smith and discontinued it;
they promptly disclosed LW’s 2023 material to Stephenson’s representatives;
they adopted the principled stance that they could not safely rely on LW in any proceedings against Stephenson, and therefore did not oppose the appeal.

This demonstrates good practice:

when prosecution witnesses recant or significantly alter their accounts, the Crown has a duty not only to reassess ongoing prosecutions but also to consider whether previous convictions relying on that witness remain safe;
full and prompt disclosure to former defendants is essential, especially where CCRC review is appropriate.

7.5 Anonymity and nuanced treatment of complainants

Notably, the Court:

maintains LW’s legal anonymity under the 1992 Act;
avoids branding her as a fantasist or as having fabricated the entire allegation;
expressly notes that she continues to allege that she was the victim of serious sexual offending, and that she has not said that the allegations of sexual offending were lies.

This careful approach:

protects complainants from being publicly vilified where their motives for earlier dishonesty may be complex (shame, fear of blame, social stigma); and
demonstrates that the quashing of a conviction need not entail a judicial conclusion that no offence occurred, but rather that the conviction cannot be upheld as safe on the available evidence.

8. Complex Concepts Simplified

8.1 “Unsafe” conviction

In England and Wales, the Court of Appeal can only allow a criminal appeal against conviction if it considers the conviction “unsafe”. This does not mean the Court must be sure of innocence, but rather:

that it can no longer be confident that the jury would have convicted had the trial been fair and all relevant evidence been available; or
that some error or new evidence creates a real possibility that the verdict is wrong.

8.2 Fresh evidence on appeal

“Fresh evidence” is material not put before the jury at trial, which:

either did not exist at the time of trial, or
was not then reasonably obtainable.

The Court must decide whether to admit it and then whether, taken with the rest of the evidence, it might reasonably have affected the jury’s verdict. In Stephenson, LW’s 2023 statement and interview are the classic example: they simply did not exist in 2015.

8.3 The CCRC

The Criminal Cases Review Commission is an independent body that investigates possible miscarriages of justice. If it considers that:

there is a real possibility that the Court of Appeal will quash a conviction, and
this is based on new evidence or argument not previously raised,

it can refer the conviction back to the Court. The CCRC does not quash convictions; it opens the door for the Court of Appeal to reconsider them.

8.4 Achieving Best Evidence (ABE) interviews

An ABE interview is a video-recorded interview with a vulnerable or intimidated witness (often including complainants in sexual offences), conducted under guidelines designed to:

maximise the quality of evidence;
minimise the risk of contamination or suggestion; and
reduce the stress of giving evidence.

In this case, LW re-watched her ABE interview in 2023 in preparation for Byron Smith’s trial, which prompted her to acknowledge that her earlier account contained lies and omissions.

8.5 Section 41 Youth Justice and Criminal Evidence Act 1999

Section 41 restricts questions and evidence about a complainant’s sexual behaviour, including:

sexual behaviour with people other than the defendant; and
often, previous sexual history.

Exceptions allow some such evidence if it is strictly necessary to ensure a fair trial—for example, if:

it directly explains physical evidence; or
it shows that the complainant’s account cannot be true in a significant respect.

The policy aim is to prevent unfair “character attacks” on complainants based on their sexual behaviour and to avoid deterring victims from reporting offences. However, as Stephenson illustrates, difficulties may arise when:

the complainant’s own sexualised behaviour forms part of the incident itself; and
the complainant later admits that she lied about that behaviour to appear more “innocent” to the jury.

9. Conclusion

R v Stephenson [2025] EWCA Crim 1600 is a stark illustration of how the discovery of fresh evidence, particularly admissions by a complainant that they lied about critical aspects of the surrounding circumstances, can fundamentally alter the safety of a conviction in a rape case.

Building on Dial and Dottin, the Court of Appeal reaffirms that:

the mere fact that a key witness has lied is not, by itself, enough to quash a conviction;
the question is whether, in the light of the lies and the true narrative, the conviction remains safe;
where the lies go to the heart of the issue in dispute—in this case consent—and significantly bolster the defence account, the conviction may well be unsafe.

The case also emphasises:

the importance of the CCRC as a safeguard, even though its actions are constrained by the timing and emergence of new evidence;
the ethical responsibilities of prosecutors and police to act swiftly and transparently when key witnesses recant or materially change their accounts;
the tension between the protective aims of s.41 YJCEA and the need for juries to hear accurate contextual evidence on consent.

Perhaps most importantly, Stephenson shows that the justice system can, albeit belatedly, correct its course when it becomes apparent that a jury was misled on essential factual matters. The Court openly acknowledges the “profound regret” that Stephenson served his entire custodial sentence before his conviction was overturned. Yet it also demonstrates that, once the truth emerged, the prosecution, CCRC, and Court acted as they should, culminating in the quashing of a conviction that can no longer be regarded as safe.