English Law before the Criminal Justice Act 2003
Abstract and Keywords
This chapter examines the doctrinal history of propensity evidence, covering the development of the common law up until its replacement by the provisions in the Criminal Justice Act 2003 (CJA). An understanding of events prior to the CJA 2003 puts us in a better position to evaluate the impact of the Act and to assess whether various criticisms of it — which inevitably assume that there is a better way of going about things — are valid. For this reason, the analysis focuses on the situation just prior to the introduction of the CJA. The case law is also significant because it provides a rich set of illustrations of situations in which questions about the admissibility of propensity evidence arise, and of how courts have responded.
[T]he common sense to which the judge had recourse here was of a special kind. It posits that if two daughters say that their father behaved indecently towards them, this makes it more likely that he was guilty than if one had said so. Many citizens would agree. Yet this is precisely the reasoning which for more than 100 years the courts in England have said is too dangerous to adopt.
R v Inder (1991)
This chapter traces the history of the propensity evidence rule from its early development up until the introduction of the Criminal Justice Act 2003. There are several reasons for spending some time on a case law which has now been largely superseded by statute. If we understand what was going on before 2003, we will be in a better position to evaluate the impact of the CJA and to assess whether various criticisms of it—which inevitably assume that there is a better way of going about things—are valid. For this reason, the analysis in this chapter will pay most attention to the situation just prior to the introduction of the CJA. The case law is also significant for other reasons: it gives us a rich set of illustrations of situations in which questions about the admissibility of propensity evidence arise, and of how courts have responded. In this way, the case law provides a foundation for the next chapter in which the various factors which mould the probative value of propensity evidence—and which should therefore be central to admissibility even after the CJA—are assessed. The response of the courts to the cases is also significant for what it tells us about common ways of reasoning about similar fact evidence and their rights and wrongs. Finally, this chapter and the next one justify some of the claims made in Chapter 4, where I argued that attempts to distinguish a morally problematic ‘forbidden’ type of reasoning about character from a morally acceptable type are unconvincing.
Before moving on to the case law, some explanation is needed. When analysing probative value, earlier chapters have tended to concentrate on the example of previous convictions. But many of the cases analysed in the current chapter involve a slightly different question: cross-admissibility. D may be tried for more than one crime at the same time. In technical terms, more than one count (alleged offence) (p.92) can be joined in a single indictment. The current test for joinder is whether the allegations: when taken together amount to a course of conduct; are founded on the same facts; or form a series of offences of the same or similar character.1 Where different counts are joined together in an indictment, the question may arise as to whether the evidence supporting different counts is ‘cross-admissible’, that is, whether evidence introduced to prove count A is admissible to prove count B. For example, D may be charged with two different burglaries. If an eyewitness identifies him as the person who committed the burglary in count A, the question is whether this evidence can support the claim that D committed the burglary on count B. The more similar the two crimes, the more it makes sense to treat the evidence as cross-admissible on the grounds that evidence suggesting that D committed burglary A shows that he has a propensity to burgle, making it more likely that he committed burglary B.
Another important piece of terminology is ‘similar fact’ evidence: this was the term which tended to be used to describe bad character evidence (and ‘similar fact rule’ the principles of admissibility) prior to the introduction of the CJA. The language of propensity which has taken hold since the CJA came into force barely figures in earlier doctrine. Insofar as it does, it was often used negatively: what was to be excluded was evidence showing a propensity to commit crime. But, as I suggest in this chapter, and argue in the next, whatever the language used, courts which talked about similar fact evidence were dealing with propensity evidence.
As doctrine developed over the course of the twentieth century, courts had a number of different admissibility rules to choose from when deciding questions about the admissibility of propensity evidence. In what follows, this leads to a rather confusing narrative, but it should be apparent that, especially towards the end of the period in which admissibility questions were governed by the common law, courts were making, if imperfectly, a judgment about the probative value of bad character evidence, and using this as a guide to admissibility.
5.1 Early Cases
Although judicial objections to propensity evidence can be found as far back as the seventeenth century,2 use of the evidence in the courts seems to have been common at least until the early nineteenth century.3 Indeed, Lacey suggests that: ‘Throughout the eighteenth century, trials were dominated by evidence (p.93) about the accused’s (and witnesses’) standing and reputation.’4 At this time, one justification for using a local jury was that it would be expected to know something about the character of the parties, and more could be discovered by questioning the accused’s witnesses: ‘did he work regularly; did he support his family; was he sober and honest in his dealing with others; did he, in other words, have an established place in a community, and was he known to his respectable neighbours as a man who could be trusted.’5 ‘A man who could produce no witnesses was likely to have a difficult time in court.’6
The emergence of something like the modern rule of exclusion is usually traced to the 1810 decision in Cole.7 But any exclusionary rule that had developed by this date was not sufficiently obvious to always be noticed—there was no mention of the rule in Hawkins’s 1824 treatise.8 The rule which did materialize was subject to exceptions from the moment of its birth. As Leonard puts it, albeit with his own gloss:
At the very least, a rule admitting evidence of other wrongdoing for non-character purposes, especially to demonstrate guilty knowledge, arose at about the same time as the firm establishment of the general character rule, and might even have predated it. From this time, cases carving out specific purposes for which such evidence might be admitted began to proliferate.9
During the nineteenth century, then, the ‘categories’ approach which came to influence the law for much of the twentieth century began to develop, the idea being that propensity evidence was admissible for certain specific purposes (categories of admissibility). It is not easy to say why a categories view of admissibility emerged. In this period, treatise writers tended to write about the law of evidence as a whole in narrow terms, organizing their analysis on the basis of type of crime or type of issue to be proved, rather than by reflecting on broad principles, and this approach may have moulded the treatment of propensity evidence.10 The categories may also have appealed to a judicial mindset of cautious incrementalism.
When it came to rationalizing the exclusion of character evidence, there is little sign in the early case law that jury prejudice was considered a problem. Judges most often explained exclusion on the grounds that evidence of uncharged (p.94) misconduct would take the defendant by surprise.11 In an era before criminal record keeping, previous misconduct was hard to prove and the defendant would have had difficulty in defending himself against allegations of prior wrongdoing, especially as he could not give evidence in his own defence until late in the nineteenth century. But this was probably not the whole story. In Cole, even though D had admitted the previous behaviour (presumably homosexuality) and his tendency towards it, the prosecution was unable to use this admission against him.12 It may well be that the problem was conceptualized in terms of relevance, it being thought that previous conduct was irrelevant. That was certainly Stone’s interpretation of the history.13 By the early twentieth century, however, jury prejudice began to feature in rationalizations of the rule.14 In Bond, for example, Kennedy J complained that:
Nothing can so certainly be counted upon to make a prejudice against an accused upon his trial as the disclosure to the jury of other misconduct of a kind similar to that which is the subject of the indictment, and, indeed, when the crime alleged is one of a revolting character, such as the charge against Bond in the present case, and the hearer is a person who has not been trained to think judicially, the prejudice must sometimes be almost insurmountable.15
But this strong statement sat alongside reference to the problem of surprise,16 and most of the argument in Bond revolved around relevance.
If propensity evidence was regarded as problematic, how was admissibility rationalized in cases which fell into the right category? To the extent that propensity evidence was regarded as irrelevant, it may be that the categories were simply seen as examples of relevant evidence being admitted, and thus as needing no particular justification. But many statements suggest that something more pragmatic was going on. Propensity evidence was often admitted in cases involving forgery and poisoning, where a defence of lack of knowledge or accident might have been plausible.17 Necessity, then, provided a justification for admission.18 As one judge put it:
Where the proof of an offence involves proof of such matters as intent to defraud, or guilty knowledge, or the like, the evidence of other transactions is often the only evidence by which that essential part of the offence can be proved. If the transaction is an isolated one a jury would seldom be satisfied of the prisoner’s guilt.19
Before the Introduction of the Criminal Justice Act 2003, the House of Lords considered the admissibility of propensity evidence in criminal cases on eight occasions.20 Despite this, the 1894 decision of the Privy Council in Makin v Attorney General for New South Wales21 influenced the case law right up until the CJA came into force. The Makins were convicted of the murder of a child whose body had been found buried, along with the bodies of other children, in the backyard of the house they were living in.22 Evidence of the finding of the other bodies, as well as of the discovery of yet more children’s corpses at other addresses where the Makins had resided, was admitted against the couple. Given the lack of consistency in the case law on the admissibility of evidence of other bad acts, the case came to the Privy Council, which found the evidence rightly admitted and upheld the convictions. As to the principles which should govern the admissibility of such evidence, Lord Herschell’s judgment was brief and cryptic:
It is undoubtedly not competent for the prosecution to adduce evidence tending to shew that the accused has been guilty of criminal acts other than those covered by the indictment, for the purpose of leading to the conclusion that the accused is a person likely from his criminal conduct or character to have committed the offence for which he is being tried. On the other hand, the mere fact that the evidence adduced tends to shew the commission of other crimes does not render it inadmissible if it be relevant to an issue before the jury, and it may be so relevant if it bears upon the question whether the acts alleged to constitute the crime charged in the indictment were designed or accidental, or to rebut a defence which would otherwise be open to the accused.23
While this endorsed an approach to admissibility based on categories which had considerable authority in the previous case law, Lord Herschell’s attempt to present an overview of principles of admissibility was problematic. As has been pointed out on numerous occasions,24 it is not easy to understand how the two sentences in the quoted passage relate to each other. Is the evidence specified as relevant in the second sentence taken to be admitted for some purpose other than showing that the accused is likely to have committed the offence because of his character? Or is the idea that where character evidence is relevant, it is admissible despite the fact that it is being used as the basis for a propensity inference? (p.96) Whatever Lord Herschell intended, in practice Makin was taken to establish an approach under which what matters is whether the evidence can be fitted into a particular category. Makin itself is vague as to what the categories were: rebutting accident, and then a broad category of ‘rebutting a defence’. But, while it was established that the categories were not closed,25 judges were reasonably cautious about expanding them, and a ‘defence’ had to be more specific than a simple plea of not guilty.26 In 1906, Bray J in Bond referred to categories of proving system, rebutting accident, and proving knowledge,27 and this was probably a reasonable interpretation of the nineteenth-century case law. As various commentators have noted,28 the solidification of the categories is not a surprising turn of events; lawyers, after all, are fond of rules, and categorization is a way of crafting a rule which will appear to be relatively determinate, and which will enable judges to avoid asking the perhaps more difficult question of whether the evidence is more prejudicial than probative.
The categories approach is best illustrated by way of example. In Harrison Owen,29 D was tried for burglary. Using keys found in a car left outside, he had let himself in to the victim’s house. At trial, his defence seemed to be that he was in a state of automatism, perhaps through intoxication. The trial judge had admitted his many previous convictions for larceny and housebreaking, but the Court of Appeal held that this was an error, and quashed the conviction. D’s defence did not fall into the recognized category of accident, but was instead that he was acting involuntarily. If bad character evidence were to be admissible in a case such as this, thought the Court, ‘the results would be very startling’.30 Contrast Mortimer,31 where D was charged with deliberately driving his car into a female cyclist. Evidence that he had knocked other female cyclists from their bicycles was admitted to show that the incident was no accident. The defendant in Flack32 was tried on counts of having committed incest with each of his three sisters. Given that his defence was a complete denial, and therefore ‘[n]o question of identity, intent, system, guilty knowledge, or of rebutting a defence of innocent (p.97) association ever arose’,33 the evidence on one count was not admissible on another. But in Hall,34 where several young men alleged that Hall had committed acts of gross indecency on them, the evidence was admissible, because Hall put forward a defence of innocent association when he alleged that he had been giving the boys medical treatment. Hall, it seems, was damned whatever defence he put forward. On one count, he denied even having met the complainant, but this merely tipped that count into the ‘identity’ category.35 Lewis provides a striking example of the difference the exact defence put forward by D can make.36 Lewis was accused of indecently assaulting his partner’s daughters. In response to one allegation, he admitted having dried one girl after a bath, but denied having done so indecently. However, when it came to an allegation that he had masturbated in front of both girls, Lewis, unsurprisingly, made a complete denial. The Court of Appeal held that evidence of Lewis’s interest in paedophilia could be admitted to prove the towelling incident, but not the masturbation incident.
These cases give a decent idea of how the categories worked. As to exactly what categories were recognized by the middle of the twentieth century, given that the categories were not closed,37 no list is likely to be complete. The brief list given by the Court of Appeal in Flack, quoted above, is indicative. Cross, writing soon after the categories were undermined by the House of Lords in Boardman, suggested the following: rebutting a defence of accident or involuntary conduct, rebutting the accused’s plea of ignorance or mistake of fact; rebutting an innocent explanation of a particular act or of the possession of incriminating material; proving identity; and rebutting a defence of innocent association.38 Whatever the categories were, enough has been said to show some of the difficulties in this way of determining admissibility: why should it matter whether Harrison-Owen alleged mistake or involuntariness? Had the complainants in Flack not been the defendant’s sisters, and he had therefore had to explain his having met them, then perhaps the evidence would have been admitted under the ‘innocent association’ category, as in Hall.39 But why should this detail in the case make a difference?
After Makin, questions about the admissibility of propensity evidence reached the House of Lords in Ball,40 Thompson,41 and Harris,42 but, beyond endorsing (p.98) Makin, these decisions did little to develop the law. In 1975, however, the House of Lords delivered its judgment in Boardman,43 which undertook a more detailed rethinking of the principles. Boardman was the headmaster at a male boarding school; he was tried on three counts involving sexual behaviour with the pupils. The evidence on two of the counts was held to be cross-admissible, and on conviction Boardman challenged this ruling.44 His defence was that there had been no impropriety. The case might have been forced into the category of ‘innocent association’, but the House of Lords was reluctant to take this course (as Lord Cross noted, in a sense Boardman had no choice but to associate with the boys because of his job).45 There was also some authority that cases involving homosexuality were subject to different principles; however, the House of Lords was unanimous in rejecting this view of the law. All of the judgments warned against an inflexible approach to the categories. Beyond this, it is not easy to say what the House of Lords agreed on, but as a whole the decision lends support to a new approach to similar facts: it was certainly heralded as such by commentators.46 Drawing on the Court of Appeal’s attempt to rethink the law in Sims,47 emphasis was placed on the concept of ‘striking similarity’, the idea being that one (alleged) instance of misconduct could be used to prove another if they were linked by sufficient resemblance. Most of the judgments suggest that this was meant to be a demanding test. This can be seen by references made to the test in Kilbourne:48 is there such underlying unity between the allegations of the witnesses ‘as to make coincidence an affront to common sense’?49 If the evidence excludes coincidence, of course, then unless the witnesses have colluded, the defendant is almost certainly guilty. The idea seems to be that ‘striking similarity’ is just a way of expressing this high standard. Thus, Lords Hailsham and Salmon gave the example of a burglar who leaves an unusual written mark in the properties he burgles, and Lord Hailsham suggested a case where sexual acts are performed while wearing ‘the ceremonial head-dress of a Red Indian chief’.50 After such examples, the similarity on the facts of Boardman itself was bound to disappoint. Lords Wilberforce and Cross emphasised that the case was borderline. The trial judge had put some emphasis on the fact that on both counts Boardman had been said to have shown an interest in being the passive partner in an act of buggery, but the House of Lords was less impressed by this. Other similarities, such as the fact that Boardman had woken the boys up at night and spoken to them in a quiet voice, were emphasised. The evidence was held to have been properly admitted.
(p.99) We also find another approach in Boardman. According to Lord Hailsham’s judgment—the most thorough in dealing with the previous law—the first sentence in Makin sets out an absolute prohibition:
evidence of bad character is not admissible for the purpose of leading to the conclusion that a person, from his criminal conduct or character, is likely to have committed the offence for which he is being held…[W]hat is not to be admitted is a chain of reasoning…If the inadmissible chain of reasoning is the only purpose for which the evidence is adduced…the evidence itself is not admissible. If there is some other relevant, probative purpose [and this is what the second half of the Makin rule sets out]…the evidence is admitted, but should be made subject to a warning from the judge that the jury must eschew the forbidden reasoning.51
Similarly, for Lord Salmon: ‘[T]he test must be: is the evidence capable of tending to persuade a jury on some ground other than his bad character and disposition to commit the sort of crime with which he is charged.’52
5.4 From Boardman to P
It is difficult to find a single coherent message in Boardman. Some of the judgments emphasise striking similarity, others forbidden reasoning. While all stress that the admissibility standard is set high—a policy perhaps most clearly symbolized by the test of excluding coincidence—this was immediately undercut by the fact that the House of Lords held the evidence rightly admitted. Perhaps this is one reason why the House of Lords was later to describe the admissibility test in Boardman as ‘somewhat unprincipled’.53 It is not surprising, then, to find conflicting impulses in the cases which followed Boardman.
Soon after Boardman, in Scarrott,54 the Court of Appeal appeared to read down the test of striking similarity by suggesting that it was synonymous with a criterion of ‘positive probative value’. But, while cited in a few cases involving, like Scarrott, multiple allegations of sexual offences, this new test was not terribly influential.55 This may have been because the admissibility criteria found in Boardman are sufficiently vague, and subject to manipulation, that they often did not thwart judicial intuitions about what bad character evidence should be admitted. For its part, the forbidden reasoning test frankly invited abuse. The idea here is that the jury must not reason to guilt via the proposition that the defendant has a propensity to commit crime. Because most bad character evidence does involve propensity (p.100) reasoning (an argument made in more detail in Chapter 6), judges who relied on the Boardman forbidden reasoning doctrine had to find a way of obscuring the role of the propensity inference. Sometimes, propensity was banished by mere assertion; in Rance, for example, the fact that Rance had passed other bribes than the one he was on trial for ‘went beyond merely showing a tendency on the part of Rance to commit the offence’56 (there was no explanation why). Less bluntly, judges often used the old idea of the categories to provide them with language which enabled them to describe what the evidence was being used to do, such as to rebut a defence, rather than how it did so, which was almost always via propensity. This linguistic sleight of hand enabled courts to give the impression that the bad character evidence did not engage forbidden reasoning. One of the most brazen examples of this strategy pre-dates Boardman. Straffen was charged with the murder by strangulation of a young girl, and evidence that he had strangled other young girls was used to prove his guilt.57 This was powerful evidence, and was more or less the only evidence against him. Yet according to the Court of Appeal, while it was an ‘irrefragable rule’ that such evidence was not admissible to prove that the accused has a ‘propensity for committing the particular type of crime with which he is being charged’,58 this evidence did not infringe that rule, because it ‘tended to rebut a defence which was otherwise open to the accused, that is, that he was not the person who committed the murder’.59 A more mundane example is Saunders,60 where evidence that D had acted aggressively towards the victim was said to be more than evidence of propensity, because it was ‘admissible to rebut some of the defences put forward’.61 And in Baird, the Court of Appeal rejected the argument that diaries, which detailed the defendant’s fantasies, ‘were evidence of inclination and no more…The expressions of sexual attraction and motive…went far to rebut the defence of innocent association.’62 As these decisions show, Boardman did not kill off the categories approach. It remained popular with courts which read Boardman as a ban on propensity reasoning.
Boardman’s striking similarity test was also open to manipulation. Here, for example, is the list of similarities noted by the trial judge in Shore, a case involving allegations of indecent assault by a headmaster:
(1) That all the incidents complained of arose during the course of the appellant’s duty as a headmaster.
(2) That all the incidents concerned girls and not boys, though on occasions boys were present as well.
(3) That the incidents took place in circumstances which cropped up again and again, that is, either in the hall (which was used for PE) or in a bus or at the swimming pool.
(5) That on each occasion the appellant’s hand would stray into the child’s clothing and would move towards, though (with one exception) it never reached, the child’s private parts.63
The Court of Appeal was unimpressed by (4), but thought that the other details showed a ‘strikingly similar pattern’.64 The number of allegations against Shore may have meant that the case was a strong one, but it is hard to see how the common features here can be described as ‘striking’. In particular, Shore’s employment as headmaster is not something which can have much more significance than his date of birth; it is rather like the significance attributed to the fact that the defendant drove ‘a similar car’ in each of the incidents in Wilmot,65 a case in which identity was not disputed. In Grant,66 the defendant’s hurried departure from the scene was said to be one of the significant similarities between a series of burglaries where D gained entry on a pretext. In Boardman itself, the fact that two complainants were woken in their dormitory at night by the defendant speaking in a quiet voice was emphasised; although, if the complainants were fabricating, they were hardly likely to say that Boardman had shouted. Although not tightly linked to the striking similarity test, a wonderful example of finding distinctiveness where there is none is Johanssen, where the defendant’s ‘particular homosexual propensities’ were described as being ‘to handle the boys’ penises and getting them to do the same with his, fellatio and buggery’.67 Significantly, this was a case where the evidence gained much of its power through multiple allegations (eight complainants) rather than through especially distinctive behaviour. But because the available admissibility tests did not obviously accommodate this feature, it was tempting to exaggerate the distinctiveness of the behaviour.
While the striking similarity test was susceptible to manipulation, it would be wrong to suggest that it never presented a barrier to admissibility. In Novac,68 the allegations that the defendant had met the complainants in an amusement arcade, offered them a bed for the night, and then attempted to bugger them were said to be ‘commonplace’,69 while in Inder the common features in the allegations were described as no more than the ‘stock in trade of the seducer of small boys’, appearing in the ‘vast majority of cases that come before the courts’.70 Another element in Boardman which nudged towards a strict standard was the ‘exclusion of coincidence’ test.71 In Novac, the common element of being picked up in an amusement arcade was not ‘inexplicable on the basis of coincidence’ and thus could not secure admissibility.72 It was this strand in the case law, taking the admissibility test to be a demanding one, which created pressure to revisit it.
(p.102) The Court of Appeal’s 1990 decision in Brooks73 seems to have created some unease about the striking similarity test—or at least about applying the test seriously. The case involved allegations of sexual abuse made by the defendant’s three daughters.74 The trial judge, referring to the ‘positive probative value’ dictum in Rance, had ruled the allegations cross-admissible, commenting that ‘it would be flying in the face of common sense to suggest that those matters ought to be tried in isolation’.75 The Court of Appeal, however, disagreed with this decision, and quashed the convictions. Warning against reading too much into the positive probative value formulation of the rule, it stressed the need to locate striking similarities in the allegations. But in this case, it could not find any; as in Inder, the commonalities were little more than the stock in trade for this type of crime: ‘there is nothing…except that the offences were alleged to have happened at home and that the daughters submitted and kept silent through fear. Sadly, these are the common coin of evidence in cases of father daughter incest. There is nothing striking about them.’76 As for the trial judge’s reference to common sense:
the common sense to which the judge had recourse here was of a special kind. It posits that if two daughters say that their father behaved indecently towards them, this makes it more likely that he was guilty than if one had said so. Many citizens would agree. Yet this is precisely the reasoning which for more than 100 years the courts in England have said is too dangerous to adopt.77
At the time, sexual child abuse was becoming an increasingly prominent public concern. Because the reasoning in Brooks suggested that it would be difficult to prosecute cases of familial child abuse, it was not surprising that it was soon to be challenged.
This is not to suggest that there were no resources for justifying admissibility in a case such as Brooks. The case law was complex enough that few decisions were inevitable. Six months after Brooks, the Court of Appeal decided Bedford, a case where admissibility of bad character evidence was approved, although described as ‘borderline’.78 The defendant had pleaded guilty to sexual offences against two boys, and these admissions were used as evidence against him in respect of a third allegation. There was little that was striking in the connections between the allegations, and unsurprisingly the defence argued ‘stock in trade’. But with three allegations the lack of similarity was not felt to be an obstacle, because the ‘cannot be coincidence’ test could be used to overcome it (there was also a hint that the ‘rebut innocent association’ category could have been used). This time, the Court of Appeal was less disdainful of common sense: the combination of facts made ‘the explanation of coincidence in relation to a similar allegation by the third boy an affront to common sense’.79 Bedford was certainly a stronger case than Brooks, but it does illustrate the flexibility of the accumulated case law.
(p.103) Soon after Brooks, a very similar case reached the House of Lords. In DPP v P,80 P’s two daughters alleged sexual abuse. In the light of Brooks, the Court of Appeal reluctantly held that the evidence was not cross-admissible, but encouraged an appeal. Lord Mackay delivered the single judgment of the House, holding that the evidence of each daughter could properly support the other. Although overruling the stock in trade cases, Mackay was otherwise respectful of previous authority. The famous Makin passage was quoted with no disapproval,81 and long quotations from all of the Boardman speeches were given. However, the gist of the judgment was that striking similarity was not always a necessary criterion for admissibility. The clever move here, which helped to justify this without disapproving of Boardman, was to quote not just from the statements of principle in that case, but also from the passages where the judges had reviewed the facts and held the evidence admissible: for, as we saw, the application to the facts very much undercut the lofty statements found elsewhere in the Boardman judgments. Lord Mackay’s summary of the law was as follows:
the essential feature of evidence which is to be admitted is that its probative force in support of the allegation that an accused person committed a crime is sufficiently great to make it just to admit the evidence, notwithstanding that it is prejudicial to the accused in tending to show that he was guilty of another crime. Such probative force may be derived from striking similarities in the evidence about the manner in which the crime was committed…But restricting the circumstances in which there is sufficient probative force to overcome prejudice of evidence relating to another crime to cases in which there is some striking similarity between them is to restrict the operation of the principle in a way which gives too much effect to a particular manner of stating it, and is not justified in principle…I consider that the judge must first decide whether there is material upon which the jury would be entitled to conclude that the evidence of one victim, about what occurred to that victim, is so related to the evidence given by another victim, about what happened to that other victim, that the evidence of the first victim provides strong enough support for the evidence of the second victim to make it just to admit it notwithstanding the prejudicial effect of admitting the evidence. This relationship, from which support is derived, may take many forms and while these forms may include ‘striking similarity’ in the manner in which the crime is committed, consisting of unusual characteristics in its execution the necessary relationship is by no means confined to such circumstances. Relationships in time and circumstances other than these may well be important relationships in this connection. Where the identity of the perpetrator is in issue, and evidence of this kind is important in that connection, obviously something in the nature of what has been called in the course of the argument a signature or other special feature will be necessary. To transpose this requirement to other situations where the question is whether a crime has been committed, rather than who did commit it, is to impose an unnecessary and improper restriction upon the application of the principle.82
In closing, Lord Mackay noted the criticism, made in argument, that admissibility decisions had become a lottery. In the light of his guidance, he thought: ‘Judgments (p.104) properly made in the light of the appropriate principles should not…yield results which could properly be described as a lottery.’83
If all one were to take from P was that the admissibility test for bad character evidence depends on whether the evidence is more prejudicial than probative, it would be hard to criticize the decision. But P made no clean sweep of the law. The various views in Boardman were all quoted, seemingly with approval. And while striking similarity was sidelined, at least in cases not involving identity, Lord Mackay’s judgment still seemed to emphasise something like similarity in its references to the importance of a relationship between the items of evidence.84 It is no surprise, then, to find that P did little to solve the problems of the law, nor to end the admissibility lottery.
5.5 After P
It is not easy to generalize about the law in the wake of P. In the case law we find all of the tests which we have come across so far still being used. The post-P position, therefore, can only be appreciated by examining a number of cases.
One point is fairly straightforward. In P, Lord Mackay had suggested that the test of striking similarity would still play a role in cases where identity was in issue. No doubt he had in mind ‘hallmark’ cases like Straffen, where the prosecution case largely depended on character evidence and thus on showing that D had a peculiar modus operandi, one shared by very few people. But where there is evidence other than the character evidence to link D to the current crime, there is no need for striking similarity.85 In John W,86 D was charged with two attacks in which he had grabbed young women with what appeared to be a sexual motivation. He denied any involvement. There was identification evidence on one count, and a decent amount of circumstantial evidence on the other count. The attacks were similar, but there was nothing linking them which could be described as striking. The Court of Appeal held that the evidence on each count was cross-admissible, and rightly so. Unlike the situation in Straffen, the character evidence did not have to do nearly all of the work by itself, so there was no need for a strong connection in terms of the similarity of the attacks.
The simplest reading of P (and one endorsed in subsequent House of Lords decisions87) is that the admissibility test involves balancing probative force against (p.105) prejudicial effect: the idea which seems to be set out right at the start of passage quoted above. But in fact it was rare to find the Court of Appeal applying a pure balancing test as a means of deciding admissibility; reference to those sentences in P was usually accompanied by an assessment of how much similarity there was between one allegation, or conviction, and another, or of whether the evidence was evidence of mere propensity, or some such. And those other criteria were usually portrayed as doing most, if not all, of the work. Bell is an exception, where admissibility was justified because the similarities ‘were such as to have forceful probative value considerably outweighing any question of any prejudice that there would be as a result of the jury hearing of the other allegations at the same time’.88 The strength of the case—47 allegations of indecent assault by a number of different boys—perhaps explains why the Court of Appeal here felt comfortable with the language of balance. In Harrison, the Court made the rare move of accepting that the disputed evidence merely showed propensity, and portrayed P as making a ‘fresh start’, where the trial judge’s assessment of the balance of competing considerations was key.89 By way of contrast, the Court in Thomas T subordinated the balancing language to the older tests of non-coincidence and going beyond propensity: ‘it is a precondition for the admission of similar fact evidence that the evidence strongly supports the charges faced by the defendant. If that precondition is not satisfied there is no balancing exercise to be carried out. That precondition was not met in the present case.’90
If P was not usually read as introducing a basic balancing test, what difference was it seen to make? It is not easy to say. In Barney, the Court of Appeal agreed with what seems to have been a widely held view: that ‘P follows Boardman as regards the degree of probative force required in similar fact evidence, but widens the manner in which the probative force can be established. It is not limited to cases where the offences display a striking similarity.’91 The reference to Boardman allowed the Court to portray admissibility as still exceptional, requiring something more than ‘mere propensity’. In John Allen V, however, P was depicted as having made more of a change; it ‘diluted’ the similar facts rule. The rule’s ‘now more broad-ranging and contextual nature undoubtedly offers judges a greater degree of latitude in assessing whether a combination of features, not striking in themselves, amount to a sufficient connection for the purposes of the rule’.92 This reading of P helped to justify admissibility in a ‘borderline’ case, involving sexual assaults on two young girls in a domestic context.
As we have seen, one thing which can definitely be said about the law after P is that striking similarity was no longer needed. However, courts often felt that some degree of factual similarity was necessary. In Kumar,93 D was convicted of two rapes involving different women. The Court of Appeal thought this a borderline case. It stressed the importance of ensuring ‘that the prosecution case amounts to more than mere reliance on a number of accusations which have been (p.106) made’; the jury itself should be warned ‘that it was not enough merely to rely upon the fact that the two had made allegations of rape, lest they fall into the error of thinking that because two have made such accusations there must be something in them’.94 It was held that there was just enough similarity in the allegations that D, a mini-cab driver, had had non-consensual sex with two fares to justify cross-admissibility, but that lack of careful jury instruction made the convictions unsafe. In Musquera,95 D was convicted of various sexual offences involving three young complainants. As to the law, the Court of Appeal considered that: ‘While the decision in DPP v P has eliminated the necessity to identify a “striking” similarity, it is still necessary to invoke some identifiable common feature or features constituting a significant connection and going beyond mere propensity or coincidence.’96 Here, the similarity was insufficient.
In these cases, then, the courts were looking for more than similarity of crime category; but precisely what extra element was required is hard to say. Beyond the sentence just quoted, the court in Musquera could add little, other than that there should be ‘significant similarity’ or some other relationship between the events. Given the vagueness here, it is hard to shake off the feeling that, despite Lord Mackays’s hopes, in the wake of P admissibility remained something of a lottery. It is obviously not easy to distinguish Musquera from P itself, or from John Allen V. In Massey,97 D was convicted of sexual assaults on three young men; the court distanced itself from Musquera, suggesting, somewhat disingenuously, that there the principal problem had been deficiencies in the summing up. The convictions were upheld.
A group of cases involving multiple allegations of indecent assaults by doctors provide a good illustration of the continuing unpredictability. In Cowie, the convictions were quashed on the grounds that the complaints were not cross-admissible.98 P’s deference to Boardman was used to justify relying on statements in the latter case which emphasised the exceptional nature of admissibility.
In our view we do not think that the nature of these allegations are such that, to use the words of Lord Salmon in Boardman, ‘common sense makes it inexplicable on the basis of coincidence’. The number of complaints has to be assessed against the number of patients seen and also against the fact that the occasional misinterpretation or exaggeration by the patient is not a fanciful explanation for allegations of indecency in these circumstances.99
In the factually similar Roy,100 however, the Court of Appeal considered that the recent decision in P meant that there could be no argument about cross-admissibility. The trial judge in Haslam101 ruled that there was no cross-admissibility between a number of complaints against a psychiatrist; the only facts he did think could support each other were found on appeal—not without some justification, as one (p.107) incident was consensual—to be insufficiently similar. Carman102 lies somewhere between the extremes. The trial judge had not thought that all complaints of sexual assault by the doctor could be used to support each other, instead placing them in groups involving similar types of behaviour. The Court of Appeal remarked that some of the case law—referring to Simpson,103 a case involving familial sexual assault—might have justified full cross-admissibility. In the event, though, it worried that the judge had not directed the jury sufficiently clearly that the allegations in one group could not be used to support those in another, and ended up quashing seven of the ten convictions—all of those where the behaviour was vaguely equivocal.
If P did not make admissibility decisions as predictable as Lord Mackay had hoped, it did not put paid to the older approaches to similar fact evidence either. After P, courts continued to rely on the categories approach, as well as on an interpretation of the admissibility rule as a ban on propensity reasoning. The categories approach could offer a way to engineer admissibility in cases where similarity was lacking. Thus, in Buono,104 evidence that D had been driving dangerously shortly before his car crashed was used to prove that dangerous driving caused the crash, on the grounds that it rebutted the defence of accident. In Baird,105 diary entries showing a sexual interest in boys were held admissible to rebut a defence of innocent association. And in Simpson,106 cross-admissibility was justified by the idea of a ‘course of conduct’—presumably a reference back to the old category of ‘system’—a concept vague enough to justify admissibility in almost any case, certainly all of the doctor cases.
Reliance on a ‘no propensity reasoning’ rule was more common as a means of excluding, rather than admitting, evidence, especially if the evidence involved incriminating objects. While in Furze107 the trial judge allowed evidence of dishonesty to prove conspiracy to defraud, on the grounds that it rebutted the defence of innocent association, the Court of Appeal held the evidence inadmissible, noting, among other things, that it only showed propensity. (In Baird,108 however, priority went the other way, and the innocent association category was used to trump the argument that the diaries were merely evidence of inclination.) The defendant’s interest in homosexual pornography was held inadmissible as pure propensity in the indecent assault case of B(RA),109 as was possession of child pornography in Luke B,110 and the defendant’s interest in young girls expressed in a ‘poem’ in John R.111 Incriminating objects were not always held inadmissible, however. Despite the Court of Appeal’s expression of the rule as a propensity based one in Clarke,112 a robber’s kit discovered in the defendant’s car was held admissible because of its significance when added to other evidence of identification.
(p.108) Given the state of the law in the years following P, it is easy to see why the Law Commission concluded that ‘[i]t is hard for parties and courts to establish exactly what the law is in this area’,113 and that ‘[i]t is difficult for prosecutors to predict whether similar fact evidence can be adduced in a particular case; the defence does not know whether the prosecution’s similar fact evidence will be admitted’.114 It would be wrong, though, to use the decisions discussed here to paint a picture of complete arbitrariness. Although there was a degree of indeterminacy, and a relatively broad borderline where cases could go either way, many of the decisions roughly track the strength of the case; Clarke, for example, was certainly a better case to leave to the jury than B(RA). But it is reasonably clear that the admissibility decisions were not determined by the rules; creative judges had so much choice between similarity, probative value, categories, forbidden reasoning, and other criteria that almost any decision could be justified. It is just that judges would usually not want to justify admissibility in weak cases.
Questions about propensity evidence reached the House of Lords four times between 1991 and 2005, dealing with issues such as how to deal with the possibility of collusion between witnesses,115 the similar facts rule in civil cases,116 and whether an allegation of criminal activity which had been prosecuted and resulted in acquittal could be used to support a new, similar allegation.117 For present purposes, the most interesting of these decisions is Randall.118
Randall was not a case involving propensity evidence in the sense we have been focussing on in this chapter—where the prosecution seeks to admit such evidence to prove guilt, and the evidence must satisfy the common law admissibility test. Randall involved a dispute between co-defendants, Randall and Glean, charged with murder. As each blamed the other for the crime, both ‘lost their shield’ under section 1(3) of the Criminal Evidence Act 1898. This meant that the jury would be instructed to consider the defendants’ criminal records as evidence of their credibility; the disputed question was whether the judge should also instruct the jury that they could consider Glean’s record of violent crime as evidence of his propensity to violence. The House of Lords held that such an instruction was appropriate. What is interesting about Randall is the way in which Lord Bingham’s judgment clearly embraces the relevance of the criminal record as evidence of violent propensity. Randall was not a case where the prosecution would have been able to rely on evidence of Glean’s criminal record as evidence of murder. There was little similarity between the offences, and the evidence could (p.109) not easily be fitted into a category. Any attempt to introduce the evidence might well have met with the response that this was evidence of ‘mere propensity’. A dismissal of the evidence in these terms had sometimes been used in cases involving admissibility under the Criminal Evidence Act: ‘There is a clear general principle, that, in general, evidence of propensity to commit a crime is not evidence that the man with that propensity committed the crime on the particular occasion.’119 In Randall, however, Lord Bingham was not swayed by such unrealistic reasoning. The evidence was seen as relevant and admissible to show propensity. Randall was decided shortly before the 2003 Act came into force. Had legislative reform not intervened, it is an interesting question whether Randall’s openness about propensity reasoning would have had any impact on similar fact cases, where, as we have seen, even after P reference to ‘forbidden reasoning’ was still common. If the issue had returned to the House of Lords it might well have done, although it is possible that the analysis in Randall would have been considered to be confined to cases involving co-defendants where issues of relevance are sometimes seen differently.120
The discussion in this chapter has been largely expository. The various admissibility tests for propensity evidence found in the post-Makin case law have been described, partly so that they can be submitted to more critical scrutiny in the following chapter. Where I have been critical, this has partly been to draw attention to the fact that, as other writers have noted, the admissibility tests have not always worked on their own terms: for example, propensity reasoning has not been avoided, even by courts taking a ‘forbidden reasoning’ approach to admissibility. More generally, the history shows a tendency for high admissibility standards to be proclaimed but then abandoned in practice. Thus, the categories were expanded, and similarities were declared to be striking when they were not—most notably in Boardman itself.121 When the rules were applied strictly, there was pressure to abandon them in order to admit probative evidence, as in the move away from striking similarity in P and later in identity cases in John W. While some might be critical of this drive towards lower standards,122 another conclusion which might be drawn from this aspect of the history is that strict admissibility standards are simply unworkable.
In this chapter I have also tried to show how indeterminate the case law was, especially in the period between P and the introduction of the Criminal Justice Act 2003. This is a significant point. As we will see in Chapter 7, one criticism (p.110) of the law under the CJA is that there are few clear rules and that admissibility of propensity evidence is unpredictable. There may be considerable truth in that criticism, but it would be wrong to presume that the law was clear before the CJA. As applied in the courts, it was anything but.
It might, though, be argued that the problems of the law were of form rather than substance, that the rules, despite their constant manipulation and indeterminacy, operated well enough in practice, with admissibility roughly tracking probative value. Hoffmann, writing shortly after Boardman, and reading that decision as one that allowed strongly probative evidence, suggested that ‘there are virtually no cases which can be said in the light of Boardman to have been wrongly decided—a tribute to the power of common sense over the forms of legal reasoning’.123 And Dennis has echoed this point, noting that there are few clear examples where very prejudicial propensity evidence of low probative value has been admitted.124 This may well be true; it is hard to find reported cases where defendants have been unfairly disadvantaged by the rules on propensity evidence. But that, of course, is only half the story: defendants may well have been unfairly advantaged by the exclusion of propensity evidence which should have been admitted. Even on the brief review in this chapter, it is possible to point to cases where there can be little dispute that this has happened: Harrison-Owen is an obvious, if rather dated, example. And we simply have little idea what sort of evidence was being excluded by the courts under the propensity evidence rule. Because until recently the prosecution lacked appeal rights against admissibility decisions, a decision to exclude cogent propensity evidence could not be appealed, and so would rarely come to the attention of commentators.125 At this stage, then, we are simply not in a position to say whether the Courts pre-CJA were generally getting things right. We will be in a better position to judge this when we have looked at the post-CJA case law.
(1) Criminal Procedure Rules 2014, para 14.2.
(2) eg Hampden’s Trial (1684) 9 How St Tr 1053; and Harrison’s Trial (1692) 12 How St Tr 833.
(3) This summary of the early history of the rule against propensity evidence draws heavily on D. P. Leonard, The New Wigmore: Evidence of Other Misconduct and Similar Events (Frederick, MD: Aspen, 2009), chs 2–3. For other useful historical material, see J. Stone, ‘The Rule of Exclusion of Similar Fact Evidence: England’ (1933) 46 Harvard L Rev 954; T. P. Gallanis, ‘The Rise of Modern Evidence Law’ (1999) 84 Iowa L Rev 499; and J. H. Langbein, The Origins of Adversary Criminal Trial (Oxford: Oxford University Press, 2003), 190–6.
(4) N. Lacey, ‘The Resurgence of Character: Responsibility in the Context of Criminalisation’ in R. A. Duff and S. Green (eds), Philosophical Foundations of Criminal Law (Oxford: Oxford University Press, 2011), 159.
(5) J. M. Beattie, Crime and the Courts in England, 1660–1800 (Oxford: Oxford University Press, 1986), 440.
(6) Beattie, Crime and the Courts in England‚ 440.
(7) Unreported. See Leonard, The New Wigmore, 4.
(8) Hawkins wrote that ‘after a crime hath been proved in the county in which it is laid, evidence may be given of other instances of the same crime in another county, in order to satisfy the jury’ (cited in Leonard, The New Wigmore, 22).
(9) The New Wigmore, 35. The interpretation of the exceptions as allowing evidence for non-character (ie roughly, ‘specific’ propensity) purposes is Leonard’s own.
(10) The New Wigmore, 28.
(11) The New Wigmore, 8; and J. H. Wigmore, A Treatise on the Anglo-American System of Trials at Common Law (3rd edn, Boston, MA: Little, Brown & Co, 1940), § 194.
(12) Quoted in Leonard, The New Wigmore, 28.
(13) ‘Rule of Exclusion’.
(14) See also Wigmore, A Treatise on the Anglo-American System, §194, arguing that prejudice is a reason for exclusion.
(15)  2 KB 389, 398.
(16)  2 KB 389, 397.
(17) Poisons were widely available and widely used in the nineteenth century; indeed, arsenic was voluntarily taken on health grounds. It also seems that poisons were a popular instrument of murder. See J. C. Wharton, The Arsenic Century: How Victorian Britain was Poisoned at Home, Work, and Play (Oxford: Oxford University Press, 2010).
(18) Leonard, The New Wigmore, 39.
(19) Ollis (1900) 2 QBD 758, 781 (Channell J).
(20) Ball  AC 47; Thompson  AC 221; Harris  AC 694; Boardman  AC 421; DPP v P  2 AC 447; R v H  2 AC 596; R v Z  2 AC 483; and Randall  1 WLR 56,  UKHL 69.
(21)  AC 57.
(22) For a narrative account of the case and its historical background, see A. Cossins, The Baby Farmers: A Chilling Tale of Missing Babies, Shameful Secrets and Murder in 19th Century Australia (Sydney: Allen & Unwin, 2013).
(23)  AC 57, 65.
(24) The passage has been said to ‘contain an internal logical contradiction which would appear to render it unworkable’: C. R. Williams, ‘The Problem of Similar Fact Evidence’ (1979) 5 Dalhousie LJ 281, 283.
(25) Harris,  AC 694, 705.
(26) See Thompson,  AC 221, 232: ‘The mere theory that a plea of not guilty puts everything material in issue is not enough for this purpose.’ Lord Sumer continued: ‘No doubt it is paradoxical that a man, whose act is so nakedly wicked as to admit of no doubt about its character, may be better off in regard to admissibility of evidence than a man whose acts are at any rate capable of having a decent face put upon them, and that the accused can exclude evidence that would be admissible and fatal if he ran two defences by prudently confining himself to one. Still, so it is’ (232–3). Cf Noor Mohammed  AC 182, 191–2: ‘An accused person need set up no defence other than a general denial of the crime alleged. The plea of not guilty may be equivalent to saying “Let the prosecution proves its case, if it can,” and having said so much the accused may take refuge in silence. In such a case it may appear (for instance) that the facts and circumstances of the particular offence charged are consistent with innocent intention, whereas further evidence, which incidentally shows that the accused has committed one or more other offences, may tend to prove that they are consistent only with a guilty intent. The prosecution could not be said, in their Lordships’ opinion, to be “crediting the accused with a fancy defence” if they sought to adduce such evidence.’
(27)  2 KB 389, 414.
(28) eg P. Mirfield, ‘Similar Facts—Makin Out?’  CLJ 83, 88.
(29) (1951) 35 Cr App R 108.
(30) Harrison Owen, 113.
(31) (1936) 25 Cr App R 150.
(32)  1 WLR 937.
(33) Flack, 943.
(34)  1 KB 302.
(35) Hall, 308.
(36) (1983) 76 Cr App R 33.
(37) See Harris,  AC 694, 705.
(38) R. Cross, Evidence (5th edn, London: Butterworths, 1979), 378–93.
(39) The description of the facts in Hall is sparse but we are told that D ‘was attached to an institution which young men used to attend’ (Hall, 302), which makes the distinction between the cases even thinner: it is not as if there was necessarily anything suspicious about Hall’s association with the complainants which might have given the prosecution case some initial plausibility. And as Lord Cross noted in Boardman ( AC 421, 458), it is not really fair to categorize a case as ‘innocent association’ when D’s profession gives him a reason for having known the complainants.
(40)  AC 47.
(41)  AC 221.
(42)  AC 694.
(43)  AC 421.
(44) The conviction on the third count had been quashed in the Court of Appeal, on account of an insufficiently careful direction on the corroborating role of distress.
(45)  AC 421, 458.
(46) L. H. Hoffmann, ‘Similar Facts after Boardman’ (1975) 91 LQR 193, describing the decision as an ‘intellectual breakthrough’ (at 193). It must be said that it is a rather sad indictment of the state of evidence law when the reconceptualization of an issue in terms of relevance is hailed in this way. See also M. E. Turcott, ‘Similar Facts: The Boardman Legacy’ (1978–79) 21 Crim LQ 43.
(47)  KB 531.
(48)  1 WLR 1365.
(49) See Boardman,  AC 421, 453, 444, 439.
(50)  AC 421, 454.
(51)  AC 421, 451–3.
(52)  AC 421, 462.
(53) O’Brien v Chief Constable of South Wales Police  2 AC 534,  UKHL 26, .
(54)  3 WLR 629. This approach was prefigured in Rance and Herron (1976) 62 Cr App R 118.
(55) It played some role in Inder (1987) 67 Cr App R 143 (although here it did not seem to result in any relaxation of the admissibility standard); Lunt (1987) 85 Cr App R 241; Barrington  1 WLR 419; and Wilmot (1989) 89 Cr App R 12. Taking the statement out of context—ie as being other than a synonym for striking similarity—was cautioned against in Brooks (1991) 92 Cr App R 341.
(56) (1976) 62 Cr App R 118, 121.
(57) Straffen  2 QB 911.
(58) Straffen, 914.
(59) Straffen, 914–15.
(60)  1 WLR 1163.
(61) Saunders, 1165.
(62) (1993) 97 Cr App R 308, 317.
(63) (1989) 89 Cr App R 32, 38. For another example, see Pattenden’s discussion of Mustafa (1977) 65 Cr App R 26: Rosemary Pattenden, ‘Similar Fact Evidence and Proof of Identity’ (1996) 112 LQR 446, 466–7.
(64) Shore, 42.
(65) (1989) 89 Cr App R 341, 348.
(66)  2 Cr App R 272, 274.
(67) (1977) 65 Cr App R 101, 103.
(68) (1977) 65 Cr App R 107.
(69) Novac, 112.
(70) (1978) 67 Cr App R 143, 149.
(71) The coincidence test was sometimes interpreted as a less demanding standard, eg in terms of ‘unlikelihood of coincidence’ in Osmanioglu  EWCA Crim 930, .
(72) (1977) 65 Cr App R 107, 112.
(73) (1991) 92 Cr App R 36.
(74) One of the three withdrew her allegation during the trial.
(75) Brooks, 39.
(76) Brooks, 42–3.
(77) Brooks, 43.
(78) (1991) 93 Cr App R 113, 117.
(79) Bedford, 117.
(80)  2 AC 447.
(81) DPP v P, 454.
(82) DPP v P, 460–2.
(83) DPP v P, 463.
(84) See Mackay’s later interpretation of P, noted at n 87.
(85) See Pattenden’s astute analysis (‘Similar Fact Evidence’), cited in John W.
(86)  2 Cr App R 289.
(87) See R v H  2 AC 596, 611 (Lord Mackay), 613 (Lord Griffiths), 621 (Lord Mustill), and 626 (Lord Lloyd) (although note Lord Mackay’s slightly more complex formulation at 603: ‘the requirement then is for a particular relationship to exist between the allegation in issue and the allegations in evidence sought to be adduced as similar fact evidence’); Randall,  1 WLR 56,  UKHL 69,  (Lord Steyn); and O’Brien v Chief Constable of South Wales Police  2 AC 534,  UKHL 26,  (Lord Phillips). The Law Commission agreed with this reading of P: Evidence in Criminal Proceedings: Previous Misconduct of a Defendant, LCCP No 141 (London: HMSO, 1996), 32.
(88)  EWCA Crim 1719, .
(89)  EWCA Crim 1792, .
(90)  EWCA Crim 2915, .
(91)  EWCA Crim 1385, .
(92)  EWCA Crim 236, .
(93)  EWCA Crim 3549.
(94) Kumar, , .
(95)  Crim LR 857.
(96) Musquera, transcript p 9.
(97)  EWCA Crim 2850.
(98)  EWCA Crim 3522.
(99) Cowie, .
(100)  Crim LR 185.
(101)  EWCA Crim 1840. The trial judge’s decision is a good example of how the concept of similarity could sometimes distort judgment: the judge thought the incidents similar because they were said to have taken place in the same room.
(102)  EWCA Crim 540.
(103) (1994) 99 Cr App R 48.
(104)  EWCA Crim 1313.
(105) (1993) 97 Cr App R 308.
(106) (1994) 99 Cr App R 48.
(107)  EWCA Crim 2706.
(108) (1993) 97 Cr App R 308.
(109)  2 Cr App R 88.
(110)  EWCA Crim 3231.
(111) CA 25 May 2000.
(112)  2 Cr App R 425.
(113) Law Commission, Evidence of Bad Character in Criminal Proceedings, Law Com No 273, Cm 5257 (London: TSO, 2001), 50.
(114) Law Commission, LCCP No 141, 167.
(115) R v H,  2 AC 596.
(116) O’Brien v Chief Constable of South Wales Police,  2 AC 534,  UKHL 26.
(117) R v Z,  2 AC 483.
(118)  1 WLR 56,  UKHL 69.
(119) Neale (1977) 65 Cr App R 304, 307.
(120) See the discussion of ‘broad’ and ‘narrow’ views of Randall in R v B(C)  2 Cr App R 34,  EWCA Crim 1254, .
(121) As Pattenden notes: ‘Similar Fact Evidence’, 465.
(122) See the discussion in the following chapter.
(123) ‘Similar Facts after Boardman’, 204.
(124) I. H. Dennis, The Law of Evidence (2nd edn, London: Sweet & Maxwell, 2002), 624.
(125) One exception is an example used by McEwan: the case of Michael Maloney. ‘Fear of causing excessive prejudice led Bathurst-Norman J. to insist that five accusations of rape made against Maloney by five different women be tried before separate juries, although the defence was consent in each case. The judge felt that he had to look for striking similarities between the incidents…In the event, each jury considered only one allegation of rape in isolation, and the defendant was acquitted by all but one.’ J. McEwan, ‘Law Commission Dodges the Nettles in Consultation Paper No. 141’  Crim LR 93. See also the examples recounted at HC Debates 2 April 2003, Col 1024, and in J. R. Spencer, Evidence of Bad Character (Oxford: Hart Publishing, 2009), 4–5.