Abstract
This article presents a socio-legal analysis of the use of non-defendant bad character evidence in Crown Court criminal trials in England. Combining an in-depth doctrinal analysis of s. 100 of the Criminal Justice Act 2003 with original qualitative empirical methods (interviews with trial counsel and observations of real Crown Court trials), the article explores the real-life practical operation of this rule of exclusion and its associated inclusionary exceptions, and the role that non-defendant bad character can have on trial tactics of counsel. In doing so, it argues that illogical Court of Appeal decisions on the use of bad character for credibility purposes are causing confusion in practice, that the ‘tit-for-tat’ gateway for defendant bad character is a more significant hurdle than s. 100 itself, and that counsel often eschew bad character applications for fear of alienating the jury.
Introduction
In England and Wales, the rules which regulate the admissibility of bad character (BC) are found in ss. 98–113 of the Criminal Justice Act (CJA) 2003. Defendant BC is admissible if it satisfies any of the ‘gateways’ within s. 101, while the gateways for non-defendant BC are found in s. 100. The BC reforms contained within the CJA 2003 were a significant departure from the prior common law approach and were not initially popular amongst legal commentators. Tapper (2004) argues that although one aim of the reforms was simplification of the law, the provisions in the CJA 2003 are anything but simple. Moreover, the inclusion of non-defendant BC evidence ultimately diluted the principles behind, and overriding objective of, the reforms, which were primarily focused on defendant BC. Further attacks concerned the nebulous term ‘reprehensible behaviour’, evidence of which is considered BC for the purposes of the CJA 2003 (Goudkamp, 2008; Munday 2005a; Waterman and Dempster, 2006).
Not all responses were hostile. Spencer (2006) and McEwan (2002) were members of a minority who were largely in favour of the reforms immediately after they were enacted. Once the BC provisions had time to develop through case law, Roberts (2022: 746 − 751) concludes that much of the immediate response from the legal community was a little harsh (although criticisms of complexity were well-founded). Similarly, Redmayne (2015) argues that when one assesses the CJA 2003 on its own merits, it is a vast improvement over the previous common law rules. Although Hunter cautions against heralding the CJA 2003 as a turning point, arguing that ‘embedded cultural norms and practices will not be altered by isolated statutory interventions’ (2007: 264), Birch (2019) argues that one of the greatest impacts has been to shift adversarial culture away from using ‘wink and nudge’ tactics in the cross-examination of non-defendants.
Other than Durston (2004a), who thought that s. 100 would have relatively little impact on the ground, and recently Birch (2019), the academic focus on BC evidence almost exclusively relates to s. 101 and defendants (Brown and Steventon, 2008; Durston, 2004b; Mirfield, 2009; Munday, 2005b, 2005c, 2005d, 2008; Redmayne, 2015; Tapper, 2004; Waterman and Dempster, 2006). A second gap in the research literature is the dearth of empirical research to substantiate critics’ theoretical and doctrinal concerns with the provisions (Roberts 2022: 750 − 751). To date, the only published empirical research project on the BC provisions of the CJA 2003 was undertaken in 2006 (Morgan Harris Burrows LLP, 2009). Court staff at three Crown Courts and three Magistrates Courts were asked to record applications to adduce BC evidence, and some interviews were conducted with lawyers, court staff and the police. 1 Ultimately, the report focused primarily on defendant BC, with 731 of the 767 applications concerning s. 101.
The findings related to the remaining 36 applications are of some interest. Of the 36 s. 100 applications, 26 concerned previous convictions, five concerned reprehensible behaviour and the remaining data was missing. Gateway s. 100(1)(a) (‘important explanatory value’) was used in 14 of the applications, s. 101(1)(b) (‘substantial probative value’) was used for 15 applications, whilst three were admitted following parties’ agreement under s. 100(1)(c). It is unfortunately not specified in the report which applications under which gateway were successful, however it is stated that 15 of the 36 applications were denied, of which eight were refused due to an ‘adverse effect on fairness’, whilst another four were excluded due to case management concerns. 2 Very little can be concluded from this very small dataset, in which court staff had to fit BC evidence and the reasons for admissibility or exclusion into pre-determined categories which did not use the statutory language. More weight can be given to the interviews, where confusion was found amongst all criminal justice professionals regarding the woolly concept of ‘reprehensible behaviour’, so that they did not know when exactly s. 100 (and s. 101) CJA 2003 applied. However, the general tenor of responses was that the BC provisions of the CJA 2003 were an improvement over the previous law and were relatively clear and simple to understand. Due to the low number of courts involved, and lack of transparency regarding the number of interviewees, caution must be employed when interpreting the findings (Spencer, 2016: 1.86 − 1.88). More positively, Spencer (2006: 1.89) and Birch (2019: 844 − 845) offer anecdotal evidence (conversations with judges and barristers) that the CJA 2003 has led to a significant change in the way that cross-examination of witnesses is conducted.
This article fills the two identified gaps in the literature by providing an in-depth, systematic analysis of s. 100 CJA 2003 which synthesises doctrinal law with original qualitative empirical data. After first noting the significance of the role non-defendant BC can play in criminal trials, the two primary gateways to admissibility within s. 100 will be scrutinised. Through this analysis, we will see how each of the two primary ‘gateways’ to admissibility for non-defendant BC are interpreted in law and practice, and how they are often used in tandem. Due to the way in which the case law has developed, the use of gateway (b) for BC going to material issues and BC going to credibility will be considered independently. The article will further consider the place and influence of non-defendant BC in trial tactics, and how tactics bear on decisions to make BC applications. Three influences are identified in the empirical data: the threat of retaliation for attacking prosecution witnesses; the possibility of alienating the jury in the process of a character attack; and decisions to call witnesses of BC.
Methodology
The empirical data presented here were generated through an empirical study designed to examine how both s. 100 CJA 2003 and s. 41 of the Youth Justice and Criminal Evidence Act 1999 were operating in practice. The relevant research questions posed were:
How do the exclusionary rules and ‘gateways’ in s. 100 and s. 41 operate? How, why, and to what extent does the availability of character evidence influence pre-trial decision-making, and trial tactics, of counsel?
This article focuses on findings related only to s. 100. As the research questions were directed at studying rules of evidence in their social setting, it was decided that a qualitative mixed-method approach was appropriate (Webley, 2010: 928 − 932), combining observations of real Crown Court trials and interviews with trial counsel.
3
Trial observations were chosen to make the exclusionary rule ‘real’ by observing the role it plays in the wider context of a full criminal trial: only by observing real trial dynamics can we understand the ways in which BC evidence is argued about, framed, and utilised in practice. Interviews with trial counsel supplemented the observation data by providing additional information concerning: counsel's understanding of s. 100, how the rule is used in specific cases, and how it influences pre-trial and eventual trial decision-making. Combining the two methods allowed triangulation between what was observed and what counsel claimed.
Observations were conducted over a period of nine months in 2018 at a large Court Centre which served a wide area. For effective trials, court would usually sit on weekdays between 10.30am and 1pm, and then 2pm and 4.30pm, resulting in a maximum five-hour observational period each day. In total, 22 cases are included for analysis in the study. This consisted of observations of approximately 450 hours of full trial hearings, and at least another 150 hours watching pre-trial hearings, sentencings and the beginnings of trials which ‘cracked’. 4 Most cases were selected through necessity: as many cases cracked there was usually only one effective trial to watch in a given week. Where there was a choice of effective trials, cases were selected randomly or, as court staff and counsel began to know more about my study, I would sometimes receive tip-offs to cases which were likely to have BC (or sexual history) issues arise. As audio or visual recording in courtrooms is illegal under s. 9 of the Contempt of Court Act 1981 and s. 41 of the Criminal Justice Act 1925, 5 observations were recorded in the form of contemporaneous handwritten notes of everything said in open court.
During the same period 10 semi-structured interviews were conducted with barristers. 6 Half of the interview participants were recruited via email or Twitter, and half during my trial observations. The latter five barristers appeared in observed cases, allowing detailed triangulation of their responses. The interview data are supplemented by notes of dozens of informal conversations with counsel who permitted me to record some general comments in lieu of a full interview.
Interviewees, all of whom were barristers, 7 are referred to throughout with a numerical identifier, e.g., Barrister 1. The observed trials are similarly labelled, e.g., Case 1. Whether interviewed or not, all barristers are referred to as female, while all judges are referred to as male. 8 When a particular observed trial is being described, individuals are referred to by their role (complainant/C, defendant/D, complainant's mother etc.). As a further anonymity safeguard, some non-material distinctive facts of observed trials and those discussed in interviews have been altered. Though there was a strong local community of criminal barristers, cases were covered by counsel from across England, including many from London. The high turnover of barristers from around the country goes some way to fend off criticisms of anecdotalism. 9
The analytic technique adopted for this study is informed by the trial observation research of Temkin et al. (2016). Their ‘inductive thematic analysis’ involves elements of both grounded theory and thematic analysis (Hammersley and Atkinson, 2019: Chapter 9; Silverman, 2014: 118 − 125): it acknowledges the imposition of a legal framework prior to analysis, but still allows themes to arise naturally within that framework. In practice, this led me to code my data initially according to the statutory structure of s. 100. Subsequently, a more grounded approach was taken whereby themes were allowed to arise from the data itself. As such, the analysis provided here is socio-legal in the truest sense, in that it combines and compares approaches to s. 100 generated by both: (a) empirical qualitative methods (‘socio’), and (b) classic doctrinal methods (‘legal’). 10 The discussion below therefore also contains substantial doctrinal content. 11 This is necessary because the research questions—particularly the first—require analysis of the intersection between doctrine and practitioner understanding of that doctrine. It would be impossible to investigate the latter without the former.
Prevalence of the issue
Westlaw lists over 240 appellate judgments which interpret and apply s. 100.
12
Of the 22 trials observed in this study, non-defendant bad character evidence was admitted in nine—not a trivial number. Moreover, my interviewees suggested that non-defendant BC is an issue that must at least be considered in almost every case: Barrister 10
I think [non-defendant bad character] is an issue quite a lot, but whether it actually gets taken any further is a different matter. Because one of the things that is always asked for in a defence case statement is going to be any previous convictions, cautions, or reprimands against any of the prosecution witnesses. Because that is capable of undermining the credibility of their witness. But it's once you’ve got the record and you can see if the complainants have got anything, you can then decide whether or not there's anything that assists you.
Barrister 7, Barrister 8 and Barrister 10 said that non-defendant BC issues most commonly arose in violent offences where the complainant has previous convictions for violence and the defence are claiming self-defence.
13
There were competing views as to the frequency of complainant BC in sex offence cases: Barrister 8
It's one of those things that crops up more often in sex cases than in other cases. The reason being: so many of these sex cases tend to be familial or tend to be defendants who don't have previous convictions. And in those circumstances, because the defendant has nothing to lose by attacking other people's character because there's no danger of their own going in, you see a lot more s. 100 applications in those cases than you do anywhere else.
Barrister 9
It's not particularly common in sex cases, it has to be said, that non-defendant bad character is an issue. […] When you raise issues about whether an application should be made about certain things, quite often they don't amount to bad character.
Both respondents’ claims seem plausible, and may be compatible. In sex offence cases, it may often be the case that the defendant has no prior convictions, and therefore the potential for ‘tit-for-tat’ applications under s. 101(1)(g) is limited. 14 Equally, complainants in sex offence cases (especially those involving children) may be unlikely to have previous convictions, and so there may be no BC to attack them with. 15 Whether complainant BC is in issue in sex offence cases is fact-specific. Note, though, that non-defendant BC in sex cases does not necessarily have to relate to the complainant. 16
Contested bad character
Though BC evidence is often admitted by party agreement (Thomason, 2021), contested applications relating to non-defendant or defendant BC were observed in 12 of the 22 sample cases. Focusing on non-defendants, the wordings of the admissibility gateways contained in s. 100(1)(a) and s. 100(1)(b), and their juridical interpretations, are pivotal in resolving these applications. This section considers gateways (a) and (b) as they relate both to material and collateral trial issues, from both doctrinal and practical perspectives.
If the BC evidence (or question) has satisfied one of the s. 100 gateways, Braithwaite
17
established that there is ‘no residual discretion in the judge to refuse to admit the evidence’.
18
Although apparently sweeping, it has been confirmed that BC evidence adduced by the prosecution which satisfies one of the gateways in s. 100 CJA 2003 is still subject to exclusion if it would have such an ‘adverse effect on the fairness of proceedings’ under s. 78 of the Police and Criminal Evidence Act 1984.
19
Spencer (2016: para 3.53) argues that this is unlikely to occur in practice, and in my trial observations there was not a single mention of s. 78 in this context. One of my interviewees offered the following explanation: Barrister 10
Obviously we always have s. 78 of PACE to exclude something, but once a judge has decided that the legislation for s. 100 is satisfied, it's very difficult to then stand up and say ‘well in which case, your honour, I’m making an application under s. 78 on the grounds of fairness.’ Because if they felt it was unfair they wouldn't have allowed it in in the first place. That's going to be their mind-set.
It seems that s. 78 may have limited practical use in excluding non-defendant BC which has satisfied a gateway within s. 100. However, when making admissibility decisions trial judges are also advised to be cognisant of the other available evidence in the trial when considering s. 100 applications. Specifically, if there is already character evidence admitted in relation to a particular individual, it may be unfair to admit further character evidence which would not substantially add to the existing evidence. 20
Gateway (a)—important explanatory evidence
The first admissibility gateway is contained in s. 100(1)(a) CJA 2003, whereby non-defendant BC may be admissible if ‘it is important explanatory evidence.’ As elucidated by s. 100(2), evidence is ‘important explanatory evidence’ if:
without it, the court or jury would find it impossible or difficult properly to understand other evidence in the case, and its value for understanding the case as a whole is substantial.
Further assistance can be gleaned from the Explanatory Notes, which describe the purpose of gateway (a) in terms of securing admission of ‘evidence which, whilst not going to the question of whether the defendant is guilty, is necessary for the jury to have a proper understanding of other evidence being given in the case by putting it in its proper context’.
21
This suggests that the gateway was primarily intended to admit BC going to background or collateral issues.
The use of the words ‘important’, ‘impossible’, ‘difficult’ and ‘substantial’ in the relevant provisions give the distinct impression of a high bar to admissibility. One might infer that the gateway would be little used. Although there are no reliable recent statistics regarding usage, there are relatively few appellate decisions addressing gateway (a) compared to gateway (b). Generally, interviewees were unwilling to discuss gateway (a), claiming that they either never, or rarely, used the gateway and so had nothing to say. This supports the impression of low usage. Two respondents were willing to expound a little, though only to complain about gateway (a)'s strictness: Barrister 6
The one problem I have with s. 100 is when it comes to important explanatory evidence. The definition that it would have to be ‘impossible for the jury to understand the case without it’. And I think ‘impossible’ is far too high a threshold. Because it's very difficult to imagine anything which would render it impossible for the jury to understand the case without it. And I think that should be a much lower threshold. Something like ‘more difficult’ or ‘potentially unfair’ or something along those lines. Because I think the approach to it should be: ‘is this likely to make a difference to the way the jury view the case?’ And if the answer is yes, then they should know about it. But that's very different from saying it makes it impossible for them to understand the case otherwise. And so I do find that bit of s. 100 troubling. And I know that a lot of my colleagues do as well, and so do a lot of judges to be honest
Barrister 8
I think ‘important explanatory evidence’ either in s. 100 or s. 101 is almost always a red herring. Because if you look at the definitions of that, it basically needs to be impossible to understand the case without it […]. I think that every CPS application that always says ‘important explanatory evidence’ is because they don't understand what it means.
Both barristers stated that gateway (a) is rarely used because the bar to admissibility is impossibly high, though neither seem to be aware that the exact formulation suggested (‘difficult’) is already provided for in the legislation as an alternative ground.
Does the case law support practitioners’ strict interpretation? The first significant case was Buaduwah-Esandol, 22 where Waller LJ restated the s. 100(2) criteria in defining what evidence should be considered ‘important explanatory evidence’, before concluding that BC evidence going to an individual's prior knowledge would not be admissible through gateway (a), though it might pass gateway (b) depending on the circumstances. 23 Buaduwah-Esandol supports the suggestion in the Explanatory Notes that BC evidence going to material issues would not be admissible via gateway (a). Subsequent cases align: the Court of Appeal has stated that BC evidence relating to an individual's motive to lie (a collateral issue) might be admissible via (a), 24 whereas BC suggesting a manslaughter victim's propensity for violence (a material issue) could not. 25
Barrister 8 offered the following example of a case where she had successfully argued the admission of BC evidence under gateway (a): Barrister 8
[T]he defendant assert[ed] that he had never met the complainant. And the complainant changed his account from ‘we went to school together’, to ‘actually, we used to deal drugs together, that's how I knew him’. And then produced a photograph of the two of them holding out a fan of money with a pile of drugs. It was quite compelling, as you can imagine! But the issue there was, it wasn't actually probative because it was a robbery—[the defendant had] robbed him. So the drug dealing wasn't probative to whether he had a propensity, but it was important explanatory evidence because it was the only way that he could explain that he knew him.
This case appears to be a reasonably straightforward application of gateway (a), as the BC had no direct relevance to the current charge of robbery, but was indirectly relevant in explaining the background relationship between the complainant and defendant.
Applications of s. 100 are not always so clear-cut. In Luckett, 26 the defendant had been convicted of putting a person in fear of violence by harassment. On appeal, the defendant argued that the trial judge erred in refusing to admit evidence of the complainant's association with drug dealers which was relevant to his defence that it was these drug dealers, rather the defendant, who had intimidated and attacked the complainant. 27 Allowing the appeal, the Court of Appeal considered that ‘without this evidence the jury would find it difficult properly to understand other evidence in the case, and its value for understanding the case as a whole was substantial’, 28 therefore satisfying s. 100(2). However, the court also thought that the BC evidence could have been admitted under gateway (b), 29 suggesting that the two gateways are not mutually exclusive. Given the statutory wordings this is not surprising, as evidence which has ‘substantial probative value’ in relation to a matter in issue may also be ‘important explanatory evidence’. As in this case, evidence of the complainant's association with drug dealers went indirectly to a material issue, namely whether it was the defendant who had done the intimidating and assault, but could also be considered to be contextual information which ‘set the scene’ for the litigated facts.
Luckett might therefore indicate that gateway (a) is not solely for BC evidence going to background or collateral issues. However, the multiple purposes of the evidence, and the Court of Appeal's ruling, warrant caution. More recently, in Edwards 30 and Ibrahim, 31 the court clearly demarcated gateways (a) and (b), suggesting that the admissibility requirements are distinct and must not be conflated. Though equally, it may be that on a given set of facts both gateways may be applicable for different purposes. Notably, the only s. 100 applications made under gateway (a) during my trial observations were argued in combination with gateway (b). 32
Gateway (b)—substantially probative to a material issue
The second admissibility gateway is contained in s. 100(1)(b) CJA 2003, whereby the evidence will be admissible if:
(b) it has substantial probative value in relation to a matter which— (i) is a matter in issue in the proceedings, and (ii) is of substantial importance in the context of the case as a whole.
Section 100(3) then sets out this list of factors which must be considered when assessing the probative value of the evidence:
(a) the nature and number of the events, or other things, to which the evidence relates; (b) when those events or things are alleged to have happened or existed; (c) where—
(i) the evidence is evidence of a person's misconduct, and (ii) it is suggested that the evidence has probative value by reason of similarity between that misconduct and other alleged misconduct, the nature and extent of the similarities and the dissimilarities between each of the alleged instances of misconduct; (d) where—
(i) the evidence is evidence of a person's misconduct, (ii) it is suggested that that person is also responsible for the misconduct charged, and (iii) the identity of the person responsible for the misconduct charged is disputed, the extent to which the evidence shows or tends to show that the same person was responsible each time.
For gateway (b) the evidence must not only be probative but ‘substantial[ly]’ probative. 33 Though the BC evidence itself does not have to be important in the context of the case, the issue in the case that the BC evidence is ‘substantial[ly] probative’ of must be of ‘substantial importance’. 34 The Explanatory Notes offer two examples of issues which may be of substantial importance depending on context. The first is the credibility of a witness (see below). The second is where a defendant claims that a witness was the individual who committed the crime charged (‘it was him, not me’ defences). Both are issues which commonly arise in trials, but the Notes are silent on how judges are to determine whether these issues are of ‘substantial importance’ in a given case.
Regarding material issues, it was held in H 35 that a non-defendant witness's propensity to do something can count as a ‘matter in issue’, 36 which may be of ‘substantial importance’ in the context of a case (especially in a ‘him, not me’ situation), and that previous convictions (or other BC) may be of ‘substantial probative value’ in supporting the alleged propensity. However, the alleged propensity must itself relate to a material issue which is of ‘substantial importance’. In Muhedeen, 37 although a complainant's four prior convictions (each involving the use of knives) were thought to be of ‘substantial probative value’ regarding his propensity to carry and use knives, that propensity was held not to relate to an issue of ‘substantial importance’ as it had no bearing on the primary issues in the case (the aggressor's identity and intention). 38
Although most reported cases deal with BC evidence going to propensity, evidence can be admissible under gateway (b) for a plurality of material issues. Barrister 4 gave the following example: Barrister 4
I made [a s. 100 application] in an allegation of GBH/attempted murder. [The complainant] was said to have assaulted her husband in response to him abusing her in the kitchen. I had to make an application [for her first husband's] previous conviction in Scotland for breach of the peace where he had abused, shouted and hit her in a car when she was 7 or 8 months pregnant. […] I needed it in to show that my client was a victim of domestic violence.
In this case, the defendant's previous husband's conviction was adduced in order to portray her as a victim of domestic violence, which would shed light on her state of mind and assist in explaining her reactions to her current husband's abuse (thereby supporting a self-defence argument). 39
Non-defendant BC went to a different issue in Case 2. D was charged with raping and sexually assaulting his partner's young daughter. The mother had two other daughters, both of whom had been sexually abused by two previous partners of hers. One of these men had been convicted following trial, whilst the other had committed suicide following an allegation. Defence counsel applied under s. 100(1)(b) to adduce these two prior instances of sexual abuse in the family, primarily to: deny D had the opportunity to commit the abuse due to the heavy presence of Social Services; and suggest that D was not ‘so stupid’ to be the third man in this family to sexually abuse one of the daughters. The judge ruled that the BC evidence was admissible under gateway (b) as the defence could not run their case properly without it. Plainly, whether an issue is of ‘substantial importance’ will depend on the specific defence advanced.
The factors listed under s. 100(3) CJA 2003 are imperative factors which a judge must take into account when considering gateway (b), but the list is non-exhaustive. In Dizaei 40 the Court of Appeal suggested that trial judges could, and should, also have regard to the factors in s. 100(2), which formally only relate to gateway (a), when using gateway (b). It was also mentioned in Hodkinson 41 that when determining admissibility under gateway (b), it might be legitimate for a trial judge to take into account the potential prejudicial effect (to the defendant) which may occur in the opening of the tit-for-tat gateway in s. 101(1)(g) CJA 2003. 42 Though this suggestion has potentially sweeping ramifications, the ruling was case-specific: the defendant was convicted of sexually assaulting two male complainants, both of whom, like the defendant, were inmates in a specialist sex-offenders prison. Here, one can plainly see the dangers of prejudice and distraction if the jury were told of each person's prior convictions. The decision in Hodkinson might consequently be limited to situations where all parties involved have significant offending histories. 43
The potential for juries to get distracted from the main issues of a case is an additional consideration which may influence trial judges’ practical decisions. Several cases have expressly stated judges are entitled to exclude evidence using gateway (b) in order to avoid satellite issues from assuming much prominence and distracting the jury from the primary issues in the trial. 44 Yet whilst this is a legitimate consideration, justice and rigorous fact-finding should not be sacrificed for the sake of expediency.
One must also scrutinise what propensity prior conduct actually suggests. The Court of Appeal has refused to accept that misconduct which does not include any element of violence, such as drug taking or non-violent sexual abuse, 45 establishes a propensity to act violently. When it comes to differing ‘levels’ of misconduct, so long as the misconduct is sufficiently similar in nature the Court of Appeal has sometimes been willing to infer that less extreme misconduct may establish a propensity to act in more extreme ways, though decisions are not fully consistent. 46
Under s. 100(3)(a), courts are directed to consider the number of prior instances of misconduct when assessing their probative value. If multiple prior instances of misconduct are being adduced to suggest a propensity, each does not have to be proved to the criminal standard. Rather, the prior instances should be considered together by the jury in order to determine whether a propensity has been established. 47
In relation to propensity, the Court of Appeal stated in Hanson:
48
There is no minimum number of events necessary to demonstrate such a propensity. The fewer the number of convictions the weaker is likely to be the evidence of propensity. A single previous conviction for an offence of the same description or category will often not show propensity. But it may do so where, for example, it shows a tendency to unusual behaviour or where its circumstances demonstrate probative force in relation to the offence charged.
49
Therefore, it appears that a single conviction for a common offence, such as assault or theft, may not be considered sufficient to demonstrate a propensity, but a single conviction for an ‘unusual’ offence (or a common offence with an ‘unusual’ modus operandi) may be considered sufficient. Whilst Hanson proposed two examples of committing child sexual abuse and arson as ‘clear’ instances of ‘unusual behaviour’, 50 the court did not further define the concept of ‘unusual’. As such, it remains unclear whether ‘unusual’ should be taken to mean behaviour which is statistically rare, or whether there is a class of behaviour which is inherently ‘unusual’.
The recency (s. 100(3)(b)) and similarity (ss. 100(3)(a), (c)) of the prior conduct are articulated factors in considering the probative value of BC evidence to establish a propensity. For example, in Berry 51 a defendant charged with assault wished to adduce the complainant's prior convictions for assault, in order to suggest a propensity for violence. In rejecting the defendant's appeal, the Court of Appeal considered it significant that the complainant's prior convictions were over 10 years old and were of a relatively minor nature. 52 Similarly, the previous violent convictions of one of the two complainants in Case 10 were excluded largely on the grounds that they were 25 years old, and so according to the trial judge were not of ‘remote arguable relevance’ to the current alleged propensity for violence. If prior behaviour is particularly uncommon or unusual, it is possible that this may outweigh the passage of time, 53 but one cannot generalise too far.
Regarding similarity, my interviews indicate that details really matter: Barrister 9
I always want the details. Not just of the convictions, but the details of what happened, the details of the offence itself, whether they pleaded guilty or not etcetera.
Barrister 8
Generally, I think people don't go too much into the detail either, even though I think the authorities say you should, to point out similarities. But the difficulty is that often, particularly convictions more than a few years old, the reason we don't go into the details is because we haven't got any.
These quotations highlight the concern that relevant and probative BC may not be admitted owing to lack of information, whether due to poor record-keeping or a lack of proper prosecution disclosure. 54
Section 100(3)(d) applies in ‘him, not me’ situations. Though it does not specifically direct that the ‘nature’ or ‘similarity’ of the prior conduct should be considered in determining admissibility, the Court of Appeal in H
55
stated that such factors might still be taken into account in establishing ‘the extent to which the evidence shows or tends to show that the same person was responsible each time’. This seems eminently sensible. Moreover, in interview, barristers felt that gateway (b) was easiest to satisfy in ‘him, not me’ situations: Barrister 6
I’m going to have to deploy [s. 100] in a case I’ve got starting next week. Because my client's charged with murder. He is saying it wasn't him, and the deceased was attacked two weeks previously. We don't know who by, but we’re saying there's a potential candidate who's got previous [convictions] for robbery and possession of a bladed article. And the deceased was stabbed to death.
An additional factor to be borne in mind when assessing probative value for the purposes of gateway (b) is whether the evidence of misconduct is hearsay. In M, 56 it was held that if BC evidence qualifies as hearsay, this may erode its probative value (even if admissible via a hearsay exception in ss. 114 − 136 CJA 2003).
We have seen that the Court of Appeal in Luckett 57 thought that a complainant's prior association with drug dealers could be admissible under gateway (b), as well as (a), as being substantially probative regarding the defendant's case that it was those same drug dealers who intimidated and assaulted the complainant rather than the defendant. However, a clutch of similar cases has been decided the other way. In both Lewis 58 and Alyson, 59 vague suggestions that the complainants’ lifestyle made them vulnerable to violence were held to not reach ‘substantial’ probative value. Therefore, though it is possible to have a propensity to be vulnerable, for BC evidence of that propensity to be considered substantially probative a specific and identifiable threat may need to be shown.
A particular point of contention in the s. 100 case law is how to deal with allegations and out-of-court disposals which qualify as BC evidence. Charges are a form of allegation which are considered to be misconduct, and therefore BC, by virtue of their being assumed true under s. 109. This assumption of truth holds, not just for the categorisation of the evidence as BC, but also for the purposes of assessing its probative value. 60 However, under s. 109(2) this presumption can be rebutted if the judge rules that ‘no… jury could reasonably find it to be true’. This limitation was utilised in Bahaji, 61 where the Court of Appeal agreed with the trial judge's assessment that for a charge to have any probative value a jury would have to conclude that the individual was guilty of that charge, which would be too much of a departure from the main issues in the current case. 62 It was also acknowledged in Bahaji that although the position may differ regarding the admissibility of charges as evidence of BC against defendants under s. 101, the difference is statutorily justified as s. 100's gateways are ‘much narrower’. 63 Unproven allegations of BC may pass the admissibility threshold in (b) if evidence can be adduced to substantiate them. Barrister 10 described a case, not in my trial sample, where she was able to support allegations concerning the complainant's violence (which had not attracted any prior police attention) by calling his alleged victims to give evidence of that violence.
Gateway (b)—substantially probative to credibility
Adducing BC evidence going to specific credibility (i.e., the veracity of the content of testimony) seems relatively uncontentious, since this merely involves rebutting a factual statement made by a witness (cf. Keane, 2015). This section will instead focus on credibility as it relates to a witness's general ability or propensity to tell the truth.
Although the wording of gateway (b) created some initial doubt as to whether credibility could be considered a ‘matter in issue in the proceedings’, it did not take long for the Court of Appeal to confirm this possibility in Yaxley-Lennon.
64
This logic was reflected in interviews: Barrister 9
The creditworthiness of a witness is always a matter in issue.
Whether it is also an issue which is of ‘substantial importance in the context of the case as a whole’ will largely depend on the case facts. In addition, it was suggested in Clarke 65 that misconduct which occurs post-conviction may theoretically be relevant and admissible for a witness's credibility for the purposes of an appeal. It seems that, unlike BC going to material issues, 66 the individual in question needs to be both alive and a witness (live or hearsay) in order for their credibility to be an issue in the case. 67
The evidence which is most likely to meet the test of ‘substantial probative value’, though probably rare, is a previous conviction for perjury or any offence against the administration of justice: Barrister 10
Obviously, if someone's got a conviction for perverting the course of justice, you’re going to be thinking: jackpot!
Another form of BC evidence easily adduced, which featured in both my interview data and trial observations but is neglected in the case law, is a previous conviction for any offence where that individual gave sworn testimony in their defence. Depending on the details of the testimony, the logic here is that being found guilty is evidence of lying under oath:
68
Barrister 8
One thing that's dramatically underused […] is when defendants have previous convictions and they’re found guilty after trial. Just adduce them for the fact that the defendant lied. I always try and use them, but lots of people don't. I think that they can be quite helpful […] One of the difficulties is […] that often the data as to whether or not the defendant gave evidence in their own trial is missing. So you can't actually be sure that they gave evidence on oath that was disbelieved.
Again, the difficulties in sourcing the details of prior convictions were highlighted. If the relevant facts can be ascertained, it seems relatively straightforward to satisfy the test in gateway (b). In Case 9, one of the complainants had a previous conviction for criminal damage where she had given evidence in her defence, but was found guilty. This evidence was initially subject to a s. 100 application, but was ultimately admitted under s. 100(1)(c) by agreement after both the prosecution and judge concurred that this ‘obviously’ met the test
Other than these relatively straightforward examples, the case law in this area is messy and contradictory. A primary sticking point appears to be whether there is a difference between ‘untruthfulness’ and ‘dishonesty’. Addressing the credibility of defendants under s. 101, Hanson 69 drew a firm distinction between untruthfulness and dishonesty, holding that the former is relevant to credibility while the latter is only relevant to certain material issues (for example, the identity of a thief).
Although logical assessments of relevance apply equally to defendants and non-defendants, the Court of Appeal has equivocated on whether Hanson should be applied in s. 100 cases. S 70 and Garnham 71 appear to follow Hanson; both held that prior convictions for offences of dishonesty could not be used to undermine a witness's credibility unless those offences also involved untruthfulness. 72 Conversely, in Stephenson 73 it was held that Hanson was limited to s. 101 cases and, unlike with defendants, there is less need for caution when considering the credibility of a non-defendant witness as they do not bear the risk of conviction. Stephenson can be criticised for failing to engage seriously with the Hanson rationale for distinguishing between dishonesty and untruthfulness, falling back instead on the old illogical common law assumption that any recent or grave previous conviction may be relevant to a witness's credibility. 74
The Court of Appeal ostensibly ‘resolved’ (Spencer, 2016: para 3.23) this conflicting authority in Brewster,
75
where the Court stated: It seems to us that the judge's task will be to evaluate the evidence of bad character which it is proposed to admit for the purpose of deciding whether it is reasonably capable of assisting a fair minded jury to reach a view whether the witness's evidence is, or is not, worthy of belief. Only then can it properly be said that the evidence is of substantial probative value on the issue of creditworthiness… Whether convictions have persuasive value on the issue of creditworthiness will, it seems to us, depend principally on the nature, number and age of the convictions. However, we do not consider that the conviction must, in order to qualify for admission in evidence, demonstrate any tendency towards dishonesty or untruthfulness.
76
The clear message was that a fair-minded jury, in assessing the credibility of a non-defendant witness, may justifiably take into account previous convictions for offences which involved no dishonesty or untruthfulness.
As Birch (2019: 851–853) notes, some early post-Brewster cases on s. 100 continued to utilise the Hanson distinction, 77 but more recent decisions suggest that courts are willing to admit previous convictions which involved no dishonesty or untruthfulness at all for the purposes of impeaching credibility. 78 In Docherty 79 a complainant's prior convictions for rape and assault were thought to be ‘of great significance’ 80 to his credibility regarding his allegation that the defendant threatened to ‘[d]o in [the complainant's] car proper’. 81 Similarly, in Hussain 82 a witness's laundry list of previous convictions (including robbery, assault, shoplifting and dangerous driving) was held to be of substantial probative value, though unlike in Docherty admissibility here seemed to be a function of the number, rather than nature, of the previous convictions. 83 For previous offences involving no dishonesty or untruthfulness, the Court of Appeal has stressed that they need to be relatively recent if they are to have substantial probative value. 84
The Brewster doctrine appears to have had some impact on the ground. Interviewees were flush with examples of adducing offences of dishonesty for credibility purposes:
85
Barrister 7
I defended someone accused of buggering his nephew—under the old [pre-Sexual Offences Act 2003] law—and the complainant had 26 previous convictions for fraud. We wanted, in essence, to show that they were a thoroughly dishonest person who would lie to get whatever they wanted. Ultimately that was successful, and I think it made a big difference in the context of the case.
In Case 5, Barrister 3 was particularly keen to admit C's prior dishonesty offences (largely for theft and robbery) to attack his credibility. But in addition, Case 5 involved the admission (by agreement) of C's convictions for drug offences which were also used to attack his credibility (Thomason, 2021: 741–742). During the initial s. 100 argument, Barrister 3 submitted that, unlike for defendants, previous convictions need not have involved dishonesty or untruthfulness in order to be relevant to a non-defendant's credibility in law. This was an explicit invocation of Brewster reasoning. Other barristers admitted using this tactic, though in tones suggesting shame in appealing to moral prejudice over logical relevance: Barrister 1
Most members of the jury will have no experience being the victims of crime save for burglary. So, if they realise [you have] burgled other people's houses, you may not attract a very sympathetic hearing from a jury.
In response to my asking whether all forms of BC, including those not involving dishonesty, are used to attack witness credibility, Barrister 9 responded: ‘[I]f a witness's bad character has gone in, that's all about their credibility.’ Barrister 8 agreed that juries tend to accept the largely discredited link between criminality and dishonesty,
86
but claimed she refused to utilise this to her advantage. She expressed commitment to the Hanson distinction: Barrister 8
I think jurors tend to come to the conclusion that if you’ve got previous convictions it's likely that you did the same thing. And I think they also come to the conclusion that if you’ve got previous convictions, whatever type they are, you’re more likely to tell fibs. I suspect that's what juries think. […] [But] it seems to me you can be a thug, and it wouldn't necessarily make you dishonest […] You can drink-drive a car full of cocaine, punch someone in the face. But as long as you’re honest about those things […] it doesn't necessarily affect your truthfulness.
Barrister 10 also endorsed the Hanson distinction. She described a case where her client was charged with assault by (digital) penetration of an employee in a shop. Barrister 10 called another employee of the shop to give evidence about the general layout of the store, to suggest that there was nowhere that the defendant could have committed the assault without anyone else seeing. This witness had a very lengthy criminal record: Barrister 10
Anything that's contrary to the Theft Act was there! But, as far as I was concerned it didn't really matter. Because stealing something is not the same as coming to court to lie. […] But the prosecutor made an application that the judge granted. […] But he was acquitted anyway, so it didn't matter in the end! But it could have mattered […] because if my client had been convicted it would have been a ground of appeal, I think. 87
Despite Barrister 10's protestations that committing offences of dishonesty is not relevant to one's testimonial credibility, the judge in this case admitted the evidence which, following Brewster, is a correct application of the law. Even on this limited sampling, it is striking that counsel are taking such different approaches to the uses of previous convictions to undermine witness credibility. Inconsistent approaches are somewhat inevitable when rational inferential reasoning clashes with doctrinal fudges.
Regarding non-conviction BC going to credibility, the Court of Appeal has not been so relaxed. For example: exaggerating events to friends, 88 making complaints to the police which were not believed, 89 allegedly committing fraud and theft, 90 and allegedly lying about sexually abusing a third party 91 have all been held not to meet the threshold of substantial probative value in relation to a witness's credibility. Two generic rationalisations have been advanced: lack of similarity between the prior instance of untruthfulness and testifying under oath; and concerns over engendering satellite litigation. Both reasons indicate more intense judicial scrutiny towards non-conviction BC as opposed to conviction BC. More recently, the Court of Appeal suggested in Moody 92 that non-conviction BC (such as evidence of racism) could be admissible for credibility purposes if it is thought that the evidence would likely impact the jury's assessment of the witness. 93 Though there is no obvious inferential link between being racist and being dishonest, Moody may evidence the first ripple effect of Brewster impacting on the use of reprehensible behaviour under s. 100(1)(b) for credibility purposes. 94
The admissibility of prior false allegations has received inconsistent judicial treatment. Most cases concern false sexual allegations and raise additional issues concerning s. 41 YJCEA 1999, which are beyond the scope of this article.
95
There is, however, some case law on false allegations of violence. In Walsh
96
and Hawkins,
97
evidence of the complainants making prior false allegations of violence against third parties was excluded due to dissimilarities in the types of violence in the prior and current allegations, and concerns over satellite litigation. In V,
98
by contrast, the complainant's prior false allegation of being mugged was held to bear substantial probative value for her credibility on the current allegation of rape. It is not clear how these cases can be satisfactorily reconciled. An example suggested by Barrister 3 may imply that Walsh and Hawkins would have greater resonance with practitioners: Barrister 3
I did a kidnap/false imprisonment trial and the victim had made two false allegations when she was 14, She was kidnapped at 16. The one was to the school about her mum holding a knife to her throat, and one was to the police saying that her house had been robbed and she had been robbed, when in fact she had had a party without her parents knowing and stuff had been nicked by her friends. We managed to get both of those false allegations in because the crux of the case was the credibility of the complainant, and here she was lying to authorities and the police at the age of 14. […] And, the crux of that case, in each of the false complaints she was saying someone had a knife and was holding a knife to her, or she was threatened by a knife. And in the current case, the clinching feature, my client had apparently held a knife to her and had struck her with a knife. So there was the link of knife-knife-knife.
For non-sexual prior false allegations, it may be that there needs to be some specific similarity, such as the knives in the case described by Barrister 3, in order to satisfy the substantial probative value test
Dual applications under both gateways
When considering which gateway to apply under, it appears that counsel often hedge their bets and try both. New empirical data are not needed to make this argument, as many appeals on s. 100 concern evidence which is being argued under both gateways.
99
Though Barrister 8 strongly disapproved of this practice, believing gateway (a) virtually impossible to satisfy,
100
the rest of my interview and trial observation data support the impression that both gateways are often utilised in tandem:
101
Barrister 9
(a) and (b) I quite often do [together]. I had a case where my client was charged with arranging to have her ex-boyfriend beaten up and robbed. He was a serving prisoner, a drug dealer, by the time of the trial. And we made an application to put in his previous convictions which included, obviously, drug dealing and violence to show why she would have been afraid of him, therefore not likely to be party to what he was suggesting in the way he was suggesting it had been carried out […] So we successfully applied to put all of that in against him [using gateways] (a) and (b).
A s. 100 application made in Case 8 reflects an indiscriminate approach. The trial concerned historic sexual abuse, allegedly committed by a female baby-sitter on two young brothers. The defence was a complete denial of any sexual contact, and counsel made an application to adduce the complainants’ mother's alleged violence against her then partner, in order to suggest the current allegations were fabricated as misplaced revenge on their violent and absent mother. The application was made under both gateways (a) and (b), with defence counsel admitting in oral submissions that she had somewhat conflated the gateways but that ‘your honour gets the point’. Though the application failed in the teeth of robust prosecution opposition, there was no negative consequence to applying under both gateways; the judge made no criticism of the defence and dealt with both limbs in ruling the evidence inadmissible.
This was a desperate and confused application that had little chance of success—indeed, it took prosecution counsel and the judge several minutes to grasp the alleged relevance of the evidence. But this case is worth mentioning precisely because of the lack of rigour of the defence's application. It appears the confusion was attributable to the defence being forced to make a hastily written application in writing during a short break, with possibly less thought and reflection than it ideally merited.
The reference to the judge ‘getting the point’ seems to be an allusion to the fact that many barristers consider the statutory wordings of the gateways mostly inconsequential. These decisions are made more on lawyers’ instinct and experience:
102
Barrister 7
I think in practice, […] we have these gateways and this statutory tick box, but we tend to take quite a practical approach to a lot of this. […] You tend to have a feel for what's going to go in and what isn't.
If, like these barristers, one has practised for several decades and made many BC applications, then it seems reasonable that to draw on that wealth of experience in assessing whether particular BC evidence is likely to be admitted or excluded. If doctrinalists might be disappointed that my empirical data reveal that few counsel differentiate between the two gateways, it is not surprising. As previously discussed, Court of Appeal judgments have been inconsistent in boundary-drawing and tolerating overlaps.
The realities of the working lives of barristers are also pertinent. Given the high level of discontinuous representation (Thomason, 2021: 747–748), it seems reasonable to suppose that if pre-trial counsel has the time to make a s. 100 application, but does not expect to actually conduct the trial, applying under both gateways will preserve the greatest flexibility for trial counsel. More generally, given the preparation required to run a case, it is no wonder that the technical niceties which distinguish gateways (a) and (b) are overlooked in favour of ‘gut instinct’ (Barrister 1), especially if there is no downside to hedging one's bets.
Bad character and trial tactics
The decision to make a s. 100 application is not one that is taken lightly. Indeed, it is not axiomatic that the existence of BC evidence against a witness will necessarily lead to an application to adduce it at trial. The following risks are examined in detail below: the threat of retaliation by the operation of s. 101(1)(g) (the ‘tit-for-tat’ gateway); and the concern that attacks on witnesses may backfire. Bad character also plays a major role in decisions to call witnesses.
Tit-for-tat
Under ss. 101(1)(g) and 101(3), defendant BC is admissible if ‘the defendant has made an attack on another person's character’, and if it would not have ‘such an adverse effect on the fairness of the proceedings that the court ought not to admit it’. A successful s. 100 application from the defence qualifies as an ‘attack’ triggering potential retaliation. This holds whether the application concerns a witness or non-witness, though in Nelson 103 the Court of Appeal suggested that if the defence was attacking a non-witness the ‘fairness’ threshold in s. 101(3) would rarely be met.
This situation arose in Case 2. The defence successfully applied under s. 100 to ask questions concerning two of C's sisters (non-witnesses), who had allegedly been abused by their mother's prior partners. As previously discussed, the intention behind this was to suggest that D would not be ‘so stupid’ to be the third man in the family to commit abuse. Already in evidence were two of D's prior convictions for assaulting a child and criminal damage. In the middle of D's evidence, the prosecution applied under s. 101(1)(g) 104 to admit two further prior convictions of domestic assault. In submissions, Barrister 1 argued for the defence that the overriding fairness caveat in s. 101(3) applied, and that there was already sufficient BC of D before the jury. The judge agreed and denied the prosecution's application.
A related, but somewhat novel, issue arose in Case 13. D was charged with several drug offences, and claimed that the real owner of the drugs, whom he alleged was a serial drug user and dealer, was his friend (F) whom he was unwilling to identify. The prosecution alleged F was fictional. Barrister 8 nonetheless applied under gateway (g)
105
to admit D's previous drug offences. In ruling on this gateway, the judge stated that the prosecution could not both claim that F was fictional and that he is a person whose character has been attacked. The application consequently failed under that gateway. For gateway (g) to be activated, the person whose character is being attacked must be real (and, ideally, identifiable).
106
Barrister 8 later confided that this aspect of her application may have been a stretch: Barrister 8
I had one bit in it that shouldn't have been in it—you know, the attack on someone's character was a bit dodgy.
Vital to this discussion is the level of the risk for the defendant. Significantly, the Court of Appeal has been willing to allow dishonesty BC to be admitted against the defendant via gateway (g) for credibility purposes, 107 departing from the reasoning in Hanson. 108 The tit-for-tat deterrent is therefore very real.
Almost all barristers I talked to, both in interview and during observations, referred to gateway (g): Barrister 6
One of the realities of life is that, generally speaking, a defendant is likely to have previous convictions. And the minute you engage s. 100, you’re then opening the door for your client's ‘form’ [BC] to go in! So, unless the difference between the prosecution witness's character is markedly worse than the defendant's, tactically you tend not to deploy s. 100 anyway.
Indeed, gateway (g) was often seen as the most significant hurdle—over the admissibility tests within s. 100 itself—in attacking a non-defendant's character, due to most defendants being repeat offenders. The bare fact of one's client having prior convictions, though, was not seen as a complete bar to making a s. 100 application, as much would depend on the specific offences, and the importance of attacking that other person's character:
109
Barrister 1
You have to make a judgement call, because the cost to you of the jury learning about your character could completely outweigh the attack on a non-defendant's character. You’ve got to judge each case on its merits.
This cost/benefit analysis must be undertaken for every BC application, where counsel's judgement will depend largely on their experience in previous cases, the nature of the current case and the types and seriousness of the BC in question.
Conversely, it was clear from my interviews that s. 100 applications were more likely in cases where tit-for-tat was not an option—because the defendant had no previous convictions: Barrister 1
It's usually the case in historic [sexual] cases that the defendant's of good character. So by attacking a non-defendant's character, you aren't risking putting your own character in because you haven't got a character to put in.
Both Barrister 1 and Barrister 8 110 highlight (historic) sex cases as ones in which gateway (g) is less likely to be a factor. With no prior record to reveal, the defence has ‘nothing to lose’ by attacking the character of a prosecution witness (or non-witness). The same tactic was deployed in Case 14 where D was charged with assaulting his ex-partner with a knife. D was of good character, and so faced no consequences in attacking C throughout the trial as a habitual drug user.
Another situation where there is ‘nothing to lose’ in making a s. 100 application is where the defence believe they have already lost the character battle. Where there is no hope of opposing a s. 101 application, there is no forensic penalty for attacking the character of the main complainant:
111
Barrister 10
If your client has got previous convictions that are relevant, and the prosecution are going to probably make an application against you that is likely to succeed, then you don't want to miss the opportunity. So, if your character is going to go in anyway, you may as well make an application against the [prosecution] witnesses because you know your character is going in.
Bad character backfiring
A cold reading of the statute and case law might imply that if there is relevant evidence of BC relating to an individual, it is likely that one of the parties will seek to admit that evidence if they think it will help their case. However, even if a witness has a string of prior convictions, it may not be in that party's best tactical interests to seek their admission. This is the worry that attacking a witness's 112 character in cross-examination will not discredit the witness in the eyes of the jury, but rather will discredit the party doing the attacking.
Barrister 10 described a case in which she was defending a man who was alleged to have digitally penetrated a woman in the shop where they both worked. She called a witness to describe the layout of the shop in order to suggest that there was no location in the shop where the defendant could have done this without other people seeing. The prosecution successfully applied under s. 100 to adduce this witness's lengthy history of committing dishonesty offences. Barrister 10 considered this overkill, and was able to flip the narrative in her closing speech: Barrister 10
In my closing I said to the jury: ‘Well why do you even need to know that? All you know is that this man has come to court and been humiliated in the witness box by the prosecution because they’re clutching at straws. Does it really matter that he's got previous convictions for theft? What are the prosecution actually saying? They don't seem to be saying that these witnesses have come to court to lie, to cover up a sexual assault that they witnessed. Because the prosecution have never suggested that they did witness it. It's always been the prosecution case that this managed to happen without them seeing it. So why does it matter that he's previously been a thief and a burglar? He's just here to assist you with what happened on that day.’ And the jury were…well it's always impossible to know, but the jury were very much nodding their heads in agreement when I was making those points. And at the time that the witness was being cross-examined about his previous convictions, there were audible frustrated sighs from the jury. So, I think it backfired—I would never have made the application personally.
Evidently, some barristers are alive to the prospect of appearing callous, or even desperate, in the eyes of the jury, especially if the witness in question is young, or otherwise vulnerable. My data also suggest that this concern is more pressing when deciding whether it is worth making a s. 41 YJCEA 1999 application to cross-examine a complainant on sexual history, as the perceived likelihood of losing the jury is even higher.
Worries about BC evidence backfiring are perhaps more salient when there is already other material undermining that individual. In Case 12, an issue arose concerning C being bullied at school, including threatening letters being left in her locker. The school had investigated, concluding that C had likely authored the letters herself to seek attention. Barrister 1, for the defence, sought to adduce the letters and the school's report under s. 100 in order to undermine C's credibility. However, this application was made in tandem with two applications to admit sexual history under s. 41 of the Youth Justice and Criminal Evidence Act 1999. In the course of a dialogue with the judge, Barrister 1 admitted that she was not sure whether cross-examining C on this relatively marginal issue, as well as on two matters of sexual history, would make D appear desperate. The judge concurred and denied the application at the invitation of the defence. A witness's BC will evidently not always be brought out.
Calling witnesses of bad character
Another aspect of trial strategy implicating BC evidence is the decision to call witnesses. From my data, the salience of BC seemingly varies depending on whether one is defending or prosecuting. Concerning the defence; many of my respondents claimed that on receiving the antecedents of any potential witnesses, the defence often gets cold feet about calling them: Barrister 6
The case I’m [prosecuting] at the moment is a paedophile sex case, and the defence were planning on calling a witness who himself had got a previous conviction for indecent assault on a girl under 14. And the defence were proposing to call him to say: ‘The defendant's a nice guy, and I wouldn't expect this sort of thing of him’. So, once we’d disclosed that previous conviction, he very quickly disappeared off the defence list of witnesses!
Barrister 10
Prosecutors don't tend to make [s. 100] applications because defence witnesses are very few and far between. And often, even when there is a defence witness, once the prosecution then disclose […] to the defence ‘Oh you’ve told us that you’re going to call this witness, here's a copy of their record so you know’ [laughs]. And once you know, you might think ‘Oh God now I can't call them’, if there's a problem. So tactically the defence will just not call that witness if they don't have to.
Such reactions seem tactically sound. If a witness is likely to have their credibility undermined by reference to BC, calling them may end up doing more harm than good. Indeed, for many defence barristers calling witnesses other than the defendant is ill-advised in general, and any potential BC related to a witness would take them out of consideration completely, unless that witness was vital:
113
Barrister 3
You only call a defence witness if they’re perfect without any previous convictions, and if they’re very, very good. You lose so many cases by calling ‘a friend’ who is shit, or a mother who is not relevant. Have I ever had a s. 100 against me? No, because I don’t think I would call someone unless […] I have to or they are bloody good.
(I ask whether she would ever call a witness with previous convictions)
Depending on the circumstances. If they knew the defendant, then no. If they were completely independent and just happened to be there, then fine. But those scenarios are not common.
These tactical assessments do not appear to depend on whether counsel think the BC is likely to be admissible (having satisfied either gateway (a) or (b) in s. 100). The mere possibility of the BC being brought up seems to be sufficient for counsel to discount calling most witnesses. It is possible, therefore, that defence witnesses who could give exculpatory evidence are not being adduced for this reason, which is problematic where prior convictions are irrelevant or would be inadmissible on judicial application of s. 100.
The salience of BC for prosecution counsel appears to be much less, since witness selections are generally made by CPS reviewing lawyers: Barrister 8
Prosecuting, I don't think it would make any difference. The reason being that because by the time as counsel you come to the case, the decision has already been made to charge, and you already have the witnesses lined up that you’re going to rely upon. And you then just have to deal with the fact that they might be undermined by whatever information you have, or convictions they have.
When prosecuting, then, barristers must make do with the witnesses they are given and ‘deal with’ any s. 100 applications against them by mitigation. This is especially true when the witness in question is the complainant—one cannot drop a case merely because the alleged victim has previous convictions. This reality is evidenced by Case 19, where D was alleged to have been the head of a drug-dealing operation, who had knifed three of his dealers. All three complainants, as well as a witness to one of the stabbings, had lengthy criminal histories. 114
An alternate view was proposed by Barrister 7, who suggested that, as trial counsel, the ultimate decision to call a witness rests with her, and that she might decide not to call a witness if she thought their BC outweighed their usefulness. There are two potential explanations for this view. The first is that Barrister 7 was a relatively senior prosecutor, and regularly took on major sex cases. She may therefore have a greater influence on pre-trial decisions than more junior barristers. Secondly, if a barrister is selected as trial counsel for a case early on (as is more likely in major cases), then they are also more likely to be involved in pre-trial decisions than a barrister who picks up the case as a late return. Unfortunately, the lack of CPS involvement in this study means that the effect of BC on CPS reviewing lawyers’ decisions to select witnesses for trial remains unexamined, but would be worth pursuing in future research if access could be secured.
Conclusion
Lack of scholarly interest in s. 100 CJA 2003 could perhaps be justified if the section was an obscure, rarely used or litigated rule of admissibility. This is plainly not the case. Whether considering the case law or my empirical data, non-defendant BC is an issue with which counsel commonly grapple.
Concerning the admissibility gateways in s. 100, the Court of Appeal's unprincipled attempts to differentiate non-defendants from defendants as regards the relevance of BC to credibility clash with lawyers’ natural instincts and, it is submitted, flout logical analysis. Yet some barristers are taking advantage of the current legal position. The Court of Appeal must surely revisit this issue and set Hanson as the standard for both s. 100 and s. 101. 115 Aside from this issue of credibility under s. 100, my data suggest that technical doctrinal distinctions between gateways (a) and (b) are often lost in practice where counsel commonly apply under both. Barristers’ attitudes are unlikely to change unless they can spend more time on case preparation, which cannot occur unless there are significant reforms to case listings and a reversal to funding cuts to both legal aid and Her Majesty's Courts and Tribunals’ Service (Thomason, 2021: 744–748).
Away from s. 100, there are also a raft of tactical factors which affect counsels’ decisions to make non-defendant BC applications. Principally, it appears that the most significant barrier to adducing the BC of a prosecution witness is not the strictness of the gateways in s. 100, but the threat of the defendant's BC being admitted in retaliation under s. 101(1)(g). Roberts (2022: 727) allegation that the ‘tit-for-tat’ gateway is the ‘least satisfactory’ of the CJA's BC regime in principle seems empirically well-founded and raises questions as to whether this gateway needs to be reformed, if not removed entirely. A second tactical barrier concerns the fact that counsel must make a judgement as to whether attacking the character of a witness will, on balance, be more beneficial or harmful to the jury's assessment of their case. The risk of character attacks backfiring is very real.
Given the relatively small scale of the study, it is unclear to what extent the identified issues are widespread across England and Wales. What is clear is that the problems identified can and sometimes do occur. Limited though it may be, unless and until further empirical data are presented this study contains the best (and effectively only) information currently available on the day-to-day use of s. 100. Moreover, it is hoped that this article illustrates the benefits of studying rules of admissibility beyond the black letter. Evidence scholars must undertake research which places both the rules of evidence and the evidence itself in its fuller socio-legal and institutional contexts. As such, the blended methodology of doctrinal research, trial observations and interviews with trial counsel 116 employed in this study offers a template for examining other rules of evidence or aspects of trial procedure. Only if this call is taken up will we begin to truly understand the relationships between doctrine, practice and tactics in the production of criminal evidence.
Footnotes
Acknowledgements
My thanks to Professor Paul Roberts, Professor John Jackson, Professor Ian Dennis, and Dr Candida Saunders for their comments and suggestions on prior versions. Particular thanks must go to Professor Di Birch for our many discussions about s.100 over the years, and for providing detailed comments on the final draft. Your presence at Nottingham will be very much missed as you enter retirement.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Economic and Social Research Council, (grant number ES/J500100/1).
