Authored By: Nafeesah Rahman
SOAS, University of London
Introduction
In two trials for identical offences, two defendants stand before a jury. Neither has a relevant criminal history in any meaningful moral sense. Yet one receives a full good character direction, instructing the jury that his testimony is more credible and that he is less likely to have committed the offence, while the other receives nothing.¹ This divergence is not a hypothetical. It is the predictable consequence of the current good character framework in English criminal law.
Under R v Vye [1993] 1 WLR 471, a defendant of good character is entitled to a two-limbed jury direction: that their absence of prior convictions supports their credibility as a witness, and that it makes them less likely to have offended.² This entitlement was affirmed in R v Aziz [1996] AC 41, where the House of Lords confirmed the direction is not discretionary for defendants of absolute good character — it is a right.³
This article argues that the guidelines in R v Hanson [2005] EWCA Crim 824, hardened in R v Hunter [2015] EWCA Crim 631, have entrenched rather than remedied these inequities. By vesting trial judges with broad, effectively unreviewable discretion while failing to define the conceptual basis for the direction itself, the framework produces outcomes that are structurally arbitrary, epistemically unsound, and incompatible with the equitable administration of criminal justice. This article examines the legal framework and its definitional fault line, interrogates the Court of Appeal’s reasoning across Hanson, Doncaster, and Hunter, and demonstrates that neither limb of the direction rests on a sound evidential rationale.
The Legal Framework and Its Definitional Fault Line
“Absolute good character,” as confirmed in Hunter, means a defendant with no previous convictions or cautions and no other reprehensible conduct alleged, admitted, or proved.⁴ Such a defendant is entitled as of right to both limbs of Vye. “Effective good character” — a judicial construct without statutory footing — applies where prior convictions are old, minor, and irrelevant to the current charge, giving the judge discretion to give a modified direction.⁵
The threshold is entirely conviction-based. English law defines good character not by reference to a defendant’s actual conduct or moral history, but by what the state has formally recorded against them. A defendant who has committed acts of dishonesty or violence that were never detected is treated identically to one who has genuinely never transgressed. Conversely, a defendant with a single spent conviction from fifteen years prior may be denied a direction entirely.
This conflates moral character with the administrative record of the criminal justice system — a record shaped by policing priorities, prosecutorial discretion, and, as the ongoing debate around disproportionate stop and search practices illustrates, structural inequalities in enforcement.⁶ The direction does not measure what it purports to measure. It measures successful evasion of institutional attention, and that foundational incoherence runs through every subsequent difficulty the framework encounters.
III. Hanson, Doncaster, Hunter, and the Manufacture of Unreviewable Discretion
R v Hanson [2005] EWCA Crim 824 arose in the context of the Criminal Justice Act 2003’s bad character provisions, but Rose LJ’s guidelines directly shaped good character directions by addressing what courts must do when a defendant’s character is not “unequivocal.” The guidelines created a tripartite structure — full direction, modified direction, or none — with the boundaries between categories left entirely undefined.⁷
R v Doncaster [2008] EWCA Crim 5 confirmed that a modified direction, typically confined to the credibility limb, could be given where convictions were insufficiently serious to negate good character entirely but too recent to warrant the full direction.⁸ It supplied no criteria for that determination, simply affirming the breadth of discretion Hanson had created.
The five-judge Court of Appeal in Hunter then tightened these boundaries decisively. Effective good character now requires that prior convictions be simultaneously old, minor, and irrelevant — a conjunctive test excluding many defendants with technically minor records.⁹ Where that threshold fails, the defendant has no entitlement to any direction, only the possibility of one reviewable solely on Wednesbury grounds.¹⁰ That standard — requiring the challenger to show no reasonable judge could have decided differently — is constitutionally inadequate for a direction that materially affects how the jury evaluates the defendant’s testimony. It insulates judicial discretion from challenge rather than disciplining it.
This produces a troubling asymmetry. The prosecution’s ability to adduce bad character evidence is tightly regulated by the CJA 2003: admissible only through defined gateways, requiring judicial leave, and subject to a probative value threshold.¹¹ The defendant’s ability to benefit from good character remains unreviewable common law discretion. The defendant faces greater evidential uncertainty in the very area the direction was designed to protect.
The Court of Appeal in Hunter was candid about its scepticism, observing that many had questioned “whether the fact someone has no previous convictions makes it any the more likely they are telling the truth.”¹² The court left the direction intact not because it was persuaded of its merits, but because it was bound by Vye and Aziz — a remarkable admission from an appellate court of its own doctrine’s fragility. It narrowed the direction’s reach while declining to examine its premise.
The Evidential Incoherence of Both Limbs
Even if the discretion problem were resolved, a deeper objection remains: neither limb of Vye rests on a defensible evidential rationale.
The propensity limb, that a person of good character is less likely to have committed the offence, depends on an inference from absence of prior convictions. The Law Commission’s report on Evidence of Bad Character recognised that propensity reasoning is unreliable even from positive prior misconduct evidence, recommending strict admissibility gateways to control prejudicial over-reliance.¹³ The propensity limb asks juries to draw an inference epistemically weaker still: a positive propensity conclusion from the mere absence of a recorded adverse history. If direct propensity evidence requires careful regulation, an inference built on its absence demands at least equal scrutiny.
The credibility limb demands separate examination. The proposition that a defendant is more likely to be telling the truth because they have no prior convictions conflates two legally distinct character traits: law-abiding conduct and personal honesty. A person may have a clean record and be a habitual liar; a person with a minor historic conviction may be scrupulously truthful.
A counter-argument, derived from Lord Steyn in Aziz, is that the direction is not evidential but protective — a procedural safeguard against wrongful conviction, giving the defendant “a position of advantage” before the jury.¹⁴ This submission fails because protective rules require predictable application. A direction dependent on the trial judge’s unreviewable classification of a conviction as “old, minor, and irrelevant” cannot reliably protect anyone. Protection without predictability is not a safeguard — it is arbitrary privilege.
Furthermore, if absence of prior convictions were genuinely probative of testimonial truthfulness, one would expect the law to recognise this wherever lay witnesses testify. Expert witnesses and police officers appear in virtually every criminal trial without any equivalent credibility enhancement. The direction exists exclusively for defendants, suggesting it functions not as a principled evidential rule but as a procedural protection dressed in evidential language. If it is a protection, it requires explicit criteria and consistent application. If it is an evidential rule, it must meet the standards of logical relevance the law of evidence demands. Currently it satisfies neither description adequately. As J R Spencer has observed, the direction “rests on a folk psychology that empirical research does not support.”¹⁵ Empirical work on jury decision-making suggests that character directions may operate as general “halo” effects rather than the focused propensity/credibility distinction the law assumes.
Conclusion
The good character direction is a doctrine in search of a rationale. Hanson introduced discretion without criteria. Doncaster confirmed it without narrowing it. Hunter narrowed the direction’s reach without questioning its premise — leaving a framework over-protective of defendants whose records are clean however that record was acquired, and systematically unfair to those with minor or irrelevant convictions outside the narrow band of effective good character.
It is submitted that the propensity limb should be abolished. The Law Commission’s analysis demonstrates such inferences are unreliable even from positive evidence; they are weaker still from its absence. The credibility limb is capable of principled application and its removal would leave defendants of genuine good character without means to draw the jury’s attention to a relevant consideration. The appropriate reform is targeted: statutory abolition of the propensity limb via amendment to the Criminal Justice Act 2003, with the credibility limb preserved but codified — subject to defined criteria replacing the open-textured Wednesbury standard Hunter has produced. Codification within the 2003 Act would resolve the asymmetry identified above, placing good character within the same statutory framework as bad character evidence, subject to equivalent scrutiny and criteria accessible to defendants before trial.
The good character direction was designed to protect the innocent. In its current form, it protects the undetected.
Footnote(S):
¹ This hypothetical represents the structural range of outcomes permitted by the current framework rather than a single reported case.
² Vye, R v [1993] 1 WLR 471 (CA), 476–477.
³ Aziz, R v [1996] AC 41 (HL), 50.
⁴ Hunter, R v [2015] EWCA Crim 631, [74].
⁵ ibid [80].
⁶ See Roberts, R (on the application of) v Commissioner of Police of the Metropolis [2015] UKSC 79, [12]–[18].
⁷ Hanson, R v [2005] EWCA Crim 824, [10]–[18].
⁸ Doncaster, R v [2008] EWCA Crim 5, [14].
⁹ Hunter (n 4) [80].
¹⁰ ibid [97]; Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948] 1 KB 223 (CA).
¹¹ Criminal Justice Act 2003, ss 101(1)(d), 103(1)(a).
¹² Hunter (n 4) [67].
¹³ Law Commission, Evidence of Bad Character in Criminal Proceedings (Law Com No 273, 2001) [9.1]–[9.4].
¹⁴ Aziz (n 3) 53 (Lord Steyn).
¹⁵ J R Spencer, Evidence of Bad Character (2nd edn, Hart 2009) 87.
Table of Cases
Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948] 1 KB 223 (CA)
Aziz, R v [1996] AC 41 (HL)
Doncaster, R v [2008] EWCA Crim 5
Hanson, R v [2005] EWCA Crim 824
Hunter, R v [2015] EWCA Crim 631
Roberts, R (on the application of) v Commissioner of Police of the Metropolis [2015] UKSC 79
Vye, R v [1993] 1 WLR 471 (CA)
Table of Legislation
Criminal Justice Act 2003, ss 101–106
Bibliography
Choo A, Evidence (5th edn, OUP 2018)
Law Commission, Evidence of Bad Character in Criminal Proceedings (Law Com No 273, 2001)
Spencer J R, Evidence of Bad Character (2nd edn, Hart 2009)
Tapper C, Cross & Tapper on Evidence (13th edn, OUP 2018)
