Head v Rex [2024] EWCA Crim 1487 — Bad character evidence: non-reprehensible “good Samaritan” conduct cannot be used to imply sexual motive; inadequate s.101(1)(d) pleading and misdirection render convictions unsafe

Court: England and Wales Court of Appeal (Criminal Division)
Date: 6 December 2024
Case: Graham Head (Appellant) v Rex (Respondent)

1. Introduction

The appellant, Graham Head, was convicted at Lewes Crown Court (HH Judge Gold KC and jury) of: kidnapping (count 1), attempted rape (count 2), and assault by penetration (count 4). He received an extended determinate sentence of 23 years concurrently (18 years’ custody + 5 years’ extended licence).

Two complainants featured:

• C1 (counts 1 and 2): alleged she was induced into the appellant’s car by fraud (he said he was an Uber driver) and then subjected to sexual touching and an attempted rape in/at the car.
• C2 (count 4): alleged a digital penetration assault when intoxicated; the prosecution case against the appellant was entirely circumstantial, built from CCTV and cell-site/location evidence, complicated by discrepancies about the location of the assault and by C2’s failure to identify the appellant in a video identification procedure, plus absence of DNA linkage.

The central appellate issue was the admission and use of evidence from three other women (Hanford, Glendinning, Alsop) and a police officer (PC Gipp) said to show a “pattern” of the appellant picking up intoxicated young women late at night, allegedly under a taxi/Uber pretence—deployed by the prosecution as bad character/propensity evidence.

Key issues on appeal

• Whether the “other incidents” evidence was properly admissible as bad character evidence (notably under Criminal Justice Act 2003, s.101(1)(d)).
• Whether the trial directions invited the jury to draw an impermissible inference of sexual motive from conduct that was not shown to be reprehensible or sexual.
• Secondary complaints: judicial interventions during evidence and alleged imbalance/errors in the summing up.

2. Summary of the Judgment

The Court of Appeal rejected the secondary grounds (judicial questioning and summing up complaints) as not rendering the convictions unsafe. However, it upheld the principal ground: the evidence of Hanford, Glendinning, Alsop and PC Gipp ought not to have been admitted or used in the way it was.

The court found:

• The prosecution’s written bad character application was wholly inadequate, relying only on a bald “Section 101(1)(d)” reference without articulating the misconduct, why it was “reprehensible,” the disputed issue(s), or the chain of relevance.
• The “other incidents” evidence did not support an inference that the appellant had a sexual motive in picking up intoxicated women.
• The judge’s directions nonetheless told the jury the evidence could indicate a sexual motive and could support the allegations involving C1 and C2.
• That combination—irrelevant/weakly probative material plus a direction inviting an improper inference—misled the jury.

3. Analysis

3.1 Precedents Cited

The judgment text, as provided, does not cite any prior case authorities by title. The Court of Appeal’s reasoning is instead anchored in:

• the statutory framework governing bad character evidence (Criminal Justice Act 2003, especially s.101), and
• core trial fairness principles: relevance, avoidance of jury confusion, and ensuring the jury is not invited to reason from prejudice rather than probative facts.

The absence of named authorities is itself instructive: the court treated the error as a straightforward misapplication of the statutory gateway and a misdirection on the permitted inferential use of the “other incidents” material.

3.2 Legal Reasoning

(a) The prosecution’s pleading failure and lack of analytical focus

The Court of Appeal criticised the prosecution for presenting a bad character application that could not “be said to have been prolix” but was, in substance, analytically empty: it listed “behaviour towards them when on a night out plus contact afterwards” and asserted admissibility via s.101(1)(d), without more.

The court explained why this mattered. Without clarity, the application process lacked focus on threshold questions:

• Was the conduct truly “reprehensible” so as to constitute “bad character” material at all?
• If so, what was the important matter in issue to which it was relevant?
• What was the permissible inference, and was it actually supported by the evidence?

(b) The judge’s direction permitted an inference (sexual motive) the evidence could not sustain

The trial directions framed the “other incidents” as potentially:

• indicative of “a man with a sexual motive for wanting very drunk women in his car,” and
• supportive of the prosecution case on whether C2 was “enticed into his car,” and generally whether C1/C2 allegations were true.

The Court of Appeal held the problem was not merely the direction’s structure (a conditional “if you are sure…” approach), but that it placed before the jury an inferential route the evidence did not rationally support. It thereby risked the jury supplying the missing probative link by suspicion or stereotype rather than evidence.

(c) Why the “other incidents” evidence was not probative of sexual motive

The court reviewed the essence of each witness:

The Court of Appeal’s central conclusion was stark: these accounts showed, at most, a tendency to pick up intoxicated women. That tendency, without more, was not probative of the contested sexual allegations.

(d) Relevance to C1 and C2: the court’s differentiation

• C1: identity and presence in the car were admitted; the disputed issues were fraud/consent context and sexual conduct. A tendency to offer lifts to intoxicated women was “of no probative value at all” on whether a sexual assault occurred.
• C2: the prosecution case hinged on circumstantial inference that C2 entered the appellant’s car. The Court of Appeal held that if similar-behaviour support was available, it was properly addressed by cross-admissibility between C1 and C2 (i.e., if the jury convicted on C1, that could support the inference for C2). The additional “other women” material added nothing of legitimate probative value.

(e) Misleading the jury and safety of convictions

The court accepted that not every evidential misstep makes a conviction unsafe. But this case was different because:

• there was a “substantial body of evidence” that “had no relevance” to whether sexual assault occurred,
• yet the jury were told it “might” have such relevance (via sexual motive), and
• the case—particularly count 4 involving C2—required careful handling of circumstantial strands and gaps.

The combination created an unacceptable risk the jury’s reasoning was contaminated by an impressionistic “pattern” narrative rather than proven probative facts. The Court of Appeal therefore could not be satisfied the convictions were safe.

3.3 Impact

(a) On bad character practice (applications and rulings)

The judgment reinforces practical and substantive requirements:

• Prosecution applications must do real work: a mere statutory citation (“Section 101(1)(d)”) is “wholly inadequate.” Parties must identify the misconduct, the gateway, the important matter in issue, and the precise inferential link.
• Judges must police inferential discipline: if the evidence cannot rationally sustain the alleged inference (here, sexual motive), directions must not invite the jury to consider it.
• “Reprehensible conduct” matters: the court’s analysis underscores that conduct capable of an innocent interpretation (assisting intoxicated women) is not readily classed as “bad character” without a provable reprehensible aspect.

(b) On circumstantial cases and “pattern” narratives

The case illustrates the danger of using “pattern” evidence to bolster a circumstantial prosecution where the pattern does not truly connect to the disputed issue. The more a case depends on inference (as with C2), the greater the need to ensure that “supporting” evidence is genuinely probative and not merely prejudicial.

(c) On trial directions and language

The court criticised the pejorative framing used at trial (e.g., “lure”/“entice”), highlighting the risk that loaded language can smuggle in moral condemnation and substitute for proof of the contested element (sexual intent or conduct).

4. Complex Concepts Simplified

• “Bad character” evidence (Criminal Justice Act 2003): evidence of misconduct offered to show a defendant’s propensity or disposition. It is not automatically admissible; it must pass statutory gateways and fairness controls.
• s.101(1)(d) “important matter in issue”: a gateway allowing bad character evidence if it is relevant to a significant disputed issue between prosecution and defence (e.g., identity, propensity, intent), not merely because it paints the defendant negatively.
• “Reprehensible conduct”: not every unusual or suspicious act qualifies. The conduct must truly be blameworthy; otherwise it risks becoming character assassination rather than evidence.
• Propensity vs opportunity: showing someone had an opportunity to offend (e.g., being alone with intoxicated people) is different from showing a propensity or intention to commit sexual offences. Opportunity without additional evidence rarely proves motive.
• Satellite litigation: side disputes about uncharged incidents that consume trial time and distract the jury. Even where manageable, the bigger question is whether the side incidents truly help decide the main issues.
• Cross-admissibility: where evidence on one count may support another count (e.g., if one incident is proved to involve sexual wrongdoing, it may support inference about another). The court viewed C1↔C2 cross-admissibility as the proper route—rather than importing weak “other incidents.”
• Unsafe conviction: the appellate test is whether the Court of Appeal is sure the conviction is safe. If jury reasoning may have been materially misled by inadmissible/misdirected evidence, safety is undermined.

5. Conclusion

Head v Rex is a clear appellate warning against stretching “bad character” beyond its legitimate evidential function. The Court of Appeal held that evidence showing only that the defendant sometimes picked up intoxicated women late at night could not, without more, be used to support an inference of sexual motive or sexual propensity. The prosecution’s inadequate application and the judge’s direction inviting the jury to consider that impermissible inference combined to mislead the jury, rendering the convictions unsafe.

The judgment’s significance lies in its insistence on: (i) disciplined articulation of relevance in bad character applications, (ii) careful judicial control of inferential reasoning in directions, and (iii) heightened sensitivity to prejudice risks when circumstantial cases are bolstered with “pattern” material that is not genuinely probative.