Fresh Credibility Evidence: Conviction Safety Depends on the Whole Evidential Case

1. Introduction

This judgment addresses when new evidence undermining a prosecution witness’s credibility makes a conviction unsafe. The appellant challenged her murder conviction after a post-trial police report revealed further reasons to question a witness who had attributed a confession to her.

The Court of Appeal dismissed the appeal. Although the report would have provided additional material for challenging the witness, her credibility had already been substantially damaged at trial. More importantly, the remaining evidence formed an exceptionally strong circumstantial case against the appellant.

The decision principally reaffirms established law rather than creates a new test: the appellate court must decide for itself whether the conviction is unsafe. The possibility that fresh evidence might have influenced the jury is relevant, but is not independently decisive.

2. Background and Procedural History

The prosecution case

The deceased lived with the appellant, his wife, and a male co-accused. The prosecution alleged that the appellant and co-accused were having an affair and jointly attacked the deceased following an argument on 25 May 2020. He died from severe traumatic head and facial injuries.

The principal evidence included:

• Relationship evidence: messages indicating that the appellant intended to leave the deceased, together with accounts from his friends that he suspected an affair.
• A disturbance: a neighbour heard an argument and banging shortly before the apparent time of the fatal assault.
• CCTV: both defendants appeared outside the flat wearing bloodstained trainers, at separate times after the apparent assault.
• Conduct when police arrived: neither defendant appeared distressed; both denied that anyone else was inside, although the deceased lay severely injured in the living room. The appellant initially said that he had fallen.
• Forensic evidence: bloodstaining on the appellant’s trainers was consistent with kicking or stamping, particularly kicking.
• Changing accounts: the appellant’s explanations differed over time and conflicted with aspects of the CCTV and physical evidence.

The appellant’s account

The appellant denied participating in the killing. She maintained that the co-accused attacked her husband and that blood transferred to her clothing and footwear while she tried to intervene. She also alleged that the co-accused subsequently raped her and prevented her from obtaining assistance.

The prosecution relied on inconsistencies between that account and the objective evidence. In particular, the CCTV showed opportunities for movement outside the flat inconsistent with her account of an uninterrupted sexual assault continuing until the police arrived. The location of her discarded clothing also conflicted with her description of where it had been removed.

Neither defendant gave evidence at trial. Both were convicted of murder on 27 November 2020. The appellant received life imprisonment with a minimum term of 14 years, less time spent on remand. The present proceedings concerned her conviction, not her sentence.

The disputed confession

A prosecution witness who had befriended the appellant in hostel accommodation alleged that the appellant confessed to killing her husband by punching, kicking and stamping on him. Significantly, the alleged confession attributed the killing to the appellant alone and excluded the co-accused’s involvement.

Cross-examination substantially weakened this evidence. The witness acknowledged an acrimonious breakdown in the friendship, hostile behaviour towards the appellant, and a lie in her police interview about having contacted the police after the alleged confession. Ultimately, when challenged about whether the confession had occurred, she said that she could not remember whether it had.

The CCRC reference

The Criminal Cases Review Commission referred the conviction under section 9 of the Criminal Appeal Act 1995. Its investigation had uncovered a police report concerning a subsequent allegation of rape made by the confession witness.

During that later investigation, CCTV contradicted important elements of the witness’s account. She admitted lying about aspects of the incident, stated that she had a problem with lying, and supplied a revised account that also did not correspond with the footage. The investigation was closed, and no action was taken against her because she was unwell.

The report was therefore potentially important credibility material. It was not, however, a conviction for making a false allegation, nor did the appeal establish through a separate trial that every aspect of the later complaint was fabricated.

3. The Court’s Decision

The Court of Appeal considered the police report de bene esse—provisionally, for the purpose of assessing its significance—and dismissed the appeal.

Its main findings were:

1. The report’s contents were capable of belief, and its post-trial origin fully explained why it had not been produced at trial.
2. The material was capable of being used to challenge the confession witness’s credibility.
3. The alleged confession concerned both participation in the killing and the appellant’s intention, so its potential relevance was accepted.
4. Nevertheless, the witness had already been extensively discredited before the jury.
5. The independent circumstantial evidence remained sufficiently powerful to leave the court in no doubt about the conviction’s safety.

The judgment does not announce a separate express order granting the application to receive fresh evidence. Its decisive conclusion is that, even considering the report, there was no basis for overturning the conviction.

4. The Statutory Framework

Section 2: the ultimate question of safety

Under section 2(1) of the Criminal Appeal Act 1968, an appeal against conviction must be allowed where the Court of Appeal thinks the conviction is unsafe. That is the governing question; the discovery of new evidence does not itself answer it.

A CCRC reference brings the conviction before the appellate court for consideration. It does not establish that the conviction is unsafe or require the court to adopt the Commission’s assessment.

Section 23: receiving fresh evidence

Section 23 of the 1968 Act governs the receipt of evidence on appeal. The statutory considerations include whether the evidence:

• appears capable of belief;
• may afford a ground for allowing the appeal;
• would have been admissible at trial on an issue relevant to the appeal; and
• has a reasonable explanation for not having been presented at trial.

These considerations guide the court’s discretion; they are not a substitute for deciding whether the conviction is unsafe. Here, credibility, potential admissibility and the explanation for non-production were substantially uncontroversial. The real dispute concerned the report’s significance for the verdict.

Challenging the witness: sections 78 and 100

The appellant argued that the report could have supported either an application to exclude the witness’s evidence under section 78 of the Police and Criminal Evidence Act 1984 or an application to introduce evidence of her bad character under section 100(1)(b) of the Criminal Justice Act 2003.

The court regarded exclusion under section 78 as a possible argument, but admission as bad-character evidence as the more likely route. It did not decide that the report necessarily required exclusion of the alleged confession.

A statutory citation requires care: paragraph 65 refers to section 101(1)(b) when discussing the witness’s bad character. That appears to be a miscitation. Section 100 concerns the bad character of a person other than the defendant; section 101 concerns a defendant’s bad character. The submissions at paragraph 60 correctly identified section 100(1)(b). The judgment should not be understood as establishing a new route through section 101 for attacking a non-defendant witness.

5. Precedents Cited and Their Influence

R v Mushtaq Ahmed [2010] EWCA Crim 2899

This supplied the central formulation of the appellate approach. At paragraph 24, Hughes LJ rejected the proposition that a conviction becomes unsafe merely because new material might have affected the jury’s decision.

The court must form its own judgment about safety, while considering the issues at trial and the possible reasoning of the jury. Asking what effect the new evidence might have had is a useful check, but the court cannot transfer its statutory responsibility to a hypothetical jury.

That distinction was decisive here. The police report plainly offered another line of cross-examination. The question was whether it made the conviction unsafe when assessed alongside the evidence already heard.

R v Garland [2016] EWCA Crim 1743

At paragraph 55, Lloyd Jones LJ applied the same approach to withheld material: possible jury impact must be considered, but the ultimate question is whether the material causes the appellate court to doubt the conviction’s safety.

The present court relied on that reasoning as further confirmation of the governing principle. The factual distinction matters: the material here arose after trial, rather than being identified as evidence withheld from the original defence.

R v Barker (Marcus) [2021] EWCA Crim 603

At paragraph 25, Edis LJ explained that the approach in R v Mushtaq Ahmed, even if once arguably incidental to the outcome, had become established law through repeated application.

This authority also clarified the relationship between sections 2 and 23. The fresh-evidence considerations must be addressed, but they are neither exhaustive nor individually determinative of safety. It therefore supported the court’s refusal to treat the report’s credibility and potential admissibility as sufficient reasons to allow the appeal.

Authorities discussed through the quoted precedents

The judgment also incorporated the following authorities through its quotations and discussion of the three principal decisions:

The judgment did not independently reconsider the factual circumstances of these underlying cases. Their importance lies in the consistent doctrinal line they establish.

6. Analysis of the Court’s Reasoning

Relevance did not establish indispensability

The court accepted that the confession evidence addressed both central issues: whether the appellant participated in the killing and whether she possessed the required intention. It did not dismiss the report as irrelevant or peripheral.

Instead, it distinguished evidence that bears on an important issue from evidence on which the prosecution case materially depends. A confession may be highly relevant without being essential to the safety of a conviction supported by compelling independent evidence.

The independent evidence was assessed cumulatively

The court’s conclusion rested on the combined force of the relationship evidence, the overheard argument, CCTV, conduct after the assault, inconsistent explanations and forensic findings. No single feature was treated as automatically proving guilt.

The forensic evidence was particularly important. Blood had penetrated deep recesses in the appellant’s right trainer, with multiple spots also present around the relevant areas and underneath the soles. The scientist considered that pattern consistent with kicking or stamping and more likely kicking.

Importantly, the scientist had accepted that movements made while attempting to separate an attacker from a victim could potentially explain aspects of the staining. The appellate court therefore did not treat the science as conclusive proof of kicking. Its significance lay in the overall pattern and its fit with the other evidence.

Similarly, the appellant’s failure to summon help and her statements to police were evidential circumstances, not freestanding substitutes for proof of participation and intent. The route to verdict required the jury to be sure that she deliberately inflicted blows causing or contributing to death, or encouraged or assisted the co-accused, with the required intention to cause death or really serious harm.

The new material added to an already substantial credibility challenge

The report would have further damaged the witness’s credibility. But the jury already knew of the breakdown in her relationship with the appellant, retaliatory behaviour, an admitted lie and uncertainty over whether the alleged confession had occurred.

The critical distinction was between exposing a previously unsuspected reliability problem and adding another example to a reliability problem already laid bare. The court considered this case to fall into the latter category.

That reasoning should not be reduced to an assumption that mental ill-health makes a witness dishonest. The concrete credibility difficulties were the witness’s demonstrated hostility, admitted falsehood and equivocation about the confession itself.

The trial directions reduced the confession’s likely significance

The judge directed the jury to disregard the alleged confession against the appellant unless sure that it had been made. The jury also had to consider whether any admission was reliable, given the appellant’s reported consumption of alcohol and cannabis.

The confession had a different potential use for the co-accused. Because it purported to exclude his involvement, he could rely on it if it might be true and therefore raised a reasonable doubt about his participation. This reflected the prosecution’s burden of proof, not an inconsistency in the directions.

The conviction of both defendants supported the appellate court’s view that the jury probably placed little weight on the confession and instead relied on the circumstantial case. The court did not claim certainty about the jury’s actual deliberations. It treated the verdicts as a contextual indicator, not a conclusive reconstruction of their reasoning.

The court distinguished its own assessment from the jury-impact check

The judgment considered likely jury reaction, as the precedents permit. But its ultimate conclusion was independent: the remaining evidence was sufficiently powerful that the new report did not cause the court to doubt the conviction’s safety.

This is why the appeal failed despite the acknowledged usefulness of the report to the defence.

7. Practical Impact and Limits

The decision offers several lessons for future fresh-evidence appeals:

• Identify the incremental effect. An appellant must explain what the new material adds to the credibility challenge actually presented at trial.
• Assess evidential dependence. The importance of an undermined witness depends on how much the prosecution case relied on that witness, not simply on whether the evidence concerned a central issue.
• Examine the whole trial. Cross-examination, judicial directions, forensic concessions, competing accounts and the remaining evidence all matter.
• Separate admissibility from safety. Evidence may be credible and potentially admissible without justifying reversal of the conviction.
• Treat later allegations precisely. A police report recording contradictions or admitted lies may bear on credibility without constituting a judicial determination of a separate offence.

The judgment does not establish that further impeachment evidence is always merely cumulative, or that strong circumstantial evidence invariably cures an unreliable confession. A different result may follow where the witness provides an indispensable link, the original credibility challenge was limited, or the new material exposes a fundamentally different defect.

Its broader significance is therefore methodological: fresh credibility evidence must be evaluated in context, through the statutory safety test, rather than through either automatic deference to a hypothetical jury or automatic dismissal because the witness was already imperfect.

8. Complex Concepts Simplified

Unsafe conviction

A conviction the appellate court considers cannot safely stand after assessing the relevant evidence and legal issues.

Fresh evidence

Evidence not presented at the original trial which a party asks the appeal court to consider. Here, it concerned events occurring after trial.

De bene esse

Considering evidence provisionally so that its potential effect can be assessed without first finally resolving its formal admission.

Jury-impact check

Asking how new evidence might have affected the jury to test the court’s own view. It is not the ultimate legal test.

Circumstantial evidence

Evidence from which guilt may be inferred, rather than a direct account of the offence. Several mutually supporting circumstances may form a compelling case.

Bad-character evidence concerning a witness

Evidence of misconduct potentially relevant to assessing a witness’s reliability, admitted through the applicable statutory gateway rather than merely to portray the witness unfavourably.

Cumulative credibility evidence

Additional material reinforcing a credibility concern already exposed. Its importance depends on how much it changes the overall assessment.

9. Conclusion

The Court of Appeal dismissed the appeal because the new police report, although damaging to the confession witness, did not undermine the powerful independent case against the appellant.

The central lesson is that fresh evidence must be assessed for its effect on the safety of the conviction as a whole. Its possible influence on a jury matters, but the final responsibility remains with the appellate court.