R v Mitchell (Northern Ireland) [2016] UKSC 55

Angeline Mitchell was convicted of murdering her partner Anthony Robin by stabbing. The Supreme Court considered whether non-conviction bad character evidence relied on to establish propensity must be proved beyond reasonable doubt incident by incident, holding that propensity itself must be proved, not each underlying fact.

Facts

On 11 May 2009, Angeline Mitchell stabbed her former partner Anthony Robin at his Belfast flat following an argument about his son’s arrest. Mr Robin suffered five knife wounds, one of which (a 20cm deep wound to the chest) was fatal. Mitchell initially blamed a ‘Swedish girl’ but at trial admitted the stabbing, claiming self-defence, provocation and lack of intent.

At her trial in October 2010 at Belfast Crown Court, the prosecution applied to introduce evidence of Mitchell’s bad character to demonstrate a propensity to use knives to threaten or attack others. None of the incidents had resulted in convictions. Two incidents (a 2003 dispute about mobile phones in which she chased two men with knives, and a 2007 incident in which she stabbed two women) were the subject of agreed statements read to the jury. Five further incidents were also admitted, including allegations she had threatened to stab Michael McGeown and Mr Robin.

Despite the agreed statements, Mitchell in evidence disputed that events had occurred as described. The trial judge did not direct the jury on the standard of proof applicable to the bad character evidence, or on whether they had to be satisfied that a propensity had been established.

Mitchell was convicted of murder. The Northern Ireland Court of Appeal quashed the conviction in 2015 on the basis that the trial judge had failed to direct the jury properly on the bad character evidence. At a retrial in April 2016, Mitchell pleaded guilty to manslaughter on the ground of diminished responsibility and was sentenced to ten years’ imprisonment.

Issues

The certified question was:

Is it necessary for the prosecution relying on non-conviction bad character evidence on the issue of propensity to prove the allegations beyond a reasonable doubt before the jury can take them into account in determining whether the defendant is guilty or not?

The Court had to determine: (i) whether each individual incident relied upon to establish propensity must be proved beyond reasonable doubt; (ii) whether the jury must consider such incidents in isolated compartments; and (iii) what, if anything, must be proved to the criminal standard in respect of propensity evidence under the Criminal Justice (Evidence) (Northern Ireland) Order 2004.

Arguments

Appellant (Prosecution)

Mr McCollum QC submitted that propensity evidence did not require special treatment or to be placed in a separate compartment for individual evaluation. All relevant evidence should be assessed together by the jury. The unsatisfactory situation in R v O’Dowd, where the jury effectively conducted multiple individual trials, should be avoided. Nothing in the 2004 Order required incidents said to constitute propensity evidence to be proved beyond reasonable doubt.

Respondent (Mitchell)

Mr O’Donoghue QC argued that at common law, disputed bad character evidence had to be proved beyond reasonable doubt before being relied upon. The 2004 Order had not abrogated this rule, having abolished only the common law rules of admissibility. It was inconceivable that a jury could entertain a reasonable doubt as to the underlying facts and yet rely on those facts to establish propensity.

Judgment

Lord Kerr, delivering the unanimous judgment, dismissed the prosecution’s appeal but corrected the Court of Appeal’s reasoning.

The pre-2004 common law position

Reviewing Makin v Attorney General for New South Wales, R v Armstrong, R v Kilbourne, R v Boardman, R v Scarrott, DPP v P and R v Z, the Court concluded that before the 2004 Order there was no clear or definitive statement as to how juries should treat similar fact or propensity evidence. The respondent’s submission that the common law required each incident to be proved beyond reasonable doubt could not be sustained.

The 2004 Order

Article 4 abolished the common law rules of admissibility. Article 6(1) created statutory ‘gateways’, and Article 8 made propensity a matter in issue under gateway (d). The legislation is silent on the standard of proof applicable to non-conviction bad character evidence.

The correct approach

Lord Kerr drew a fundamental distinction between proof of propensity itself and proof of the individual underlying facts said to demonstrate it. The existence of a propensity must be proved to the criminal standard, because such evidence stands apart from evidence directly bearing on guilt. However, it does not follow that each individual incident must be proved beyond reasonable doubt or considered in isolation.

The jury should consider evidence of propensity ‘in the round’ for two reasons: first, the improbability of multiple similar allegations all being false is itself a relevant consideration; second, similarities between incidents may provide mutual corroboration. Decisions about propensity should not result from review of incidents in ‘hermetically sealed compartments’.

The Court reviewed R v Ngyuen (where there was only a single previous incident, so proof of that incident effectively merged with proof of propensity), R v Lafayette (to the extent its reasoning required isolated, compartmentalised evaluation, it was disapproved), and R v O’Dowd (the requirement that the jury conduct what were effectively three mini-trials of the underlying incidents was incorrect, though the practical concern about satellite litigation overwhelming a trial was legitimate).

Application

The trial judge in Mitchell’s case had failed to give adequate directions on how to approach propensity. He had not directed the jury that they had to be sure that a propensity had been established before relying on it. The conviction was unsafe and was properly quashed.

Implications

The decision clarifies the proper approach to non-conviction bad character evidence admitted to show propensity under the Criminal Justice (Evidence) (Northern Ireland) Order 2004 (and, by parity of reasoning, the Criminal Justice Act 2003):

• Where propensity is relied on, the jury must be directed that they must be sure that the alleged propensity has been established before taking it into account.
• It is not necessary to prove every underlying incident beyond reasonable doubt; the jury should evaluate the evidence as a whole.
• Incidents may mutually corroborate one another, and the improbability of multiple allegations being false may be considered.
• Propensity is an incidental issue. It cannot alone establish guilt, and the jury must be reminded that the defendant is not on trial for the previous allegations.
• Trial judges must remain vigilant under Article 6(3) (and the equivalent English provision) about admitting volumes of disputed bad character evidence which could generate satellite litigation and skew the trial, even if technically admissible.

The Court expressly noted that specimen directions in the Bench Books for Northern Ireland and England and Wales, which suggested juries must be sure of the truth of every allegation of propensity, were misconceived and should be reconsidered. The decision is therefore of direct practical importance to trial judges, prosecutors and defence practitioners dealing with bad character applications, and significantly recalibrates the way in which juries are to be directed on this commonly contested category of evidence.

Verdict: The Supreme Court dismissed the prosecution’s appeal. Although the Court of Appeal’s reasoning was partly disapproved (insofar as it suggested each previous incident must be proved beyond reasonable doubt and considered in isolation), the trial judge had failed to direct the jury that they had to be sure a propensity had been established. The conviction was therefore unsafe and had been properly quashed.

Source: R v Mitchell (Northern Ireland) [2016] UKSC 55