The provenance of what is proven: exploring (mock) jury deliberation in Scottish rape trials

This article presents findings from the largest research study of the nature of mock jury deliberations in rape cases undertaken in the UK to date – and the first such study to be undertaken in the Scottish context. The study found considerable evidence of the expression of problematic attitudes towards rape complainers. These included the belief that a ‘real’ rape victim would have extensive external and internal injuries and would resist attack by inflicting injuries on her attacker and shouting for help, that even a short delay in reporting a rape is sus-picious, and that false allegations of rape are commonly made by women and difficult to refute. There was, how-ever, also evidence that some jurors were willing to challenge these attitudes and that they often relied – explic-itly or implicitly – on third-sector campaigns to do so. The article concludes by drawing out the implications of this research for policy and practice.

the UK to date. 2 It was also the first to explore in a systematic way the approach taken by Scottish (mock) jurors to key aspects of Scottish law and procedure, including decisions reached by juries of 15 people via simple majority, a second acquittal verdict (not proven), and a corroboration requirement. 3 In previous work, we have reported on the impact of jury size, majority rules, and number of verdicts, across both rape and assault trial juries. 4 In this article, we examine the tone, content, and outcome of discussions specifically within the 32 juries that observed the rape trial. Aside from the fact that juries without the not proven option at their disposal were more likely than others to discuss the meaning of 'reasonable doubt', we found no significant differences across rape deliberations as a result of variables relating to jury size, majority rules, or number of verdicts in terms of the amount or types of evidential issues discussed. However, we did find considerable evidence of jurors expressing unfounded assumptions regarding how 'real' victims typically react during and after a sexual assault. In what follows, we highlight the persistence of the view across the rape juries that a lack of physical resistance is indicative of consent, and the extent to which jurors asserted that rape allegations are 'easy' to make and frequently unfounded. Furthermore, we explore -in the context of a corroboration requirement -the lack of clarity among jurors as to whether evidence provided by a medical expert (which confirmed that the complainer's injuries were consistent with her account, but could also have been inflicted through other means, including consensual sex) could adequately support the complainer's testimony.
This discussion provides a substantial evidence base that builds on and affirms findings of previous studies that have used mock juries to evaluate the content of deliberations in rape cases. 5 It also goes beyond those findings by enabling reflection for the first time on distinctively Scottish aspects, including the corroboration requirement and the not proven verdict, against a background of a number of high-profile public campaigns spearheaded by Rape Crisis Scotland and designed to challenge misconceptions around sexual assault. 6 The article has four substantive parts. First, we situate our current focus on rape trial deliberations within the broader frame of the overarching project on jury decision making. We also outline briefly the distinctive dimensions of the Scottish legal and social context. Second, we set out our 2 Other large-scale simulation studies that have included deliberation as part of the research design have previously been research methods. Third, we discuss our findings around three core themes: force, resistance, and injury; the parties' behaviour before and after the alleged attack; and false allegations. In the fourth and final part, we reflect on the implications of our findings for law and policy responses. In Scotland, the corroboration requirement, the not proven verdict, and attrition in respect of rape complaints have been, and continue to be, the subject of substantial political and popular scrutiny. 7 So too, in jurisdictions beyond Scotland, policy makers wrangle with similar concerns regarding how to improve justice responses for rape complainers. 8 We consider how our findings might inform these debates.

DISTINCTIVE DELIBERATIONS? THE UNIQUENESS OF THE SCOTTISH CONTEXT
The Scottish jury has unique characteristics not found in other common-law systems. It has 15 members, is not required to deliberate in pursuit of a unanimous verdict but instead returns a decision by a simple majority, 9 and has at its disposal -in addition to the guilty and not guilty verdicts -a third option of not proven. The legal consequences of a not proven and a not guilty verdict are identical: the accused is acquitted and cannot normally be re-prosecuted for the same offence. 10 Beyond this, juries are given no further instruction about its meaning or how to differentiate not proven from not guilty. 11 In addition, Scots law has a corroboration requirement, which entails that, in a criminal case, there must be at least two sources of evidence in respect of each 'crucial fact' (the identity of the perpetrator and key elements of the offence). 12 This requirement has proven particularly contentious in respect of sexual offences, where it is often difficult to locate evidence corroborating the complainer's account. This has required a measure of pragmatism on the part of the courts to establish ways in which corroboration might be achieved -for example, through evidence of the complainer's distress in the aftermath of the alleged attack. 13 A review of Scottish criminal procedure undertaken by Lord Carloway in 2011 recommended abolition of the corroboration requirement, 14 9 That is, there must be eight votes in favour of conviction for the jury to return a guilty verdict. Anything less is an acquittal, and the choice of acquittal verdict will depend on the balance of votes between the two acquittal verdicts. Scottish juries cannot 'hang': for further explanation, see Ormston et al., op. cit., n. 1, p. 1. 10 This is subject to the provisions of the Double Jeopardy (Scotland) Act 2011, which permits re-prosecution following an acquittal in a narrow range of circumstances (which are not affected by whether the verdict is one of not guilty or not proven commissioned -the Post-Corroboration Safeguards Review -to examine the changes to law and practice that might be required if the corroboration requirement were removed. 15 Among other things, it recommended that '[t]he time is right to undertake research into jury reasoning and decision-making' since 'simultaneous changes to several unique aspects of the Scottish jury system should only be made on a fully informed basis'. 16 The authors, along with Ipsos MORI Scotland, were commissioned to undertake this research. The aims of the study were to explore the impact of the unique features of jury size, majority required, and number of verdicts on decision making, and how jurors understood and used the not proven verdict. 17 Given pre-existing concerns regarding sexual offence trials, the tender stipulated a specific requirement to include these as one area of investigation, alongside which we included an assault trial involving a plea of self-defence. 18 We have reported elsewhere the findings of the study in respect of these distinctive aspects of the Scottish jury system. 19 In this article, we focus attention instead on the tone and content of deliberations across the rape trials, and the extent to which the approach taken by our participants supports or challenges findings of previous research regarding jurors' use of stereotypes and myths in rape cases. 20 This study is larger in scale than any previous UK study of the dynamics and content of mock juror deliberations in rape cases, and because of this -and the measures that we took to enhance realism -it offers a particularly robust evidence base. It also provides the first opportunity to explore the prevalence of such views among Scottish jurors, and to assess how such views might interact in distinctive ways with some unique elements of the Scottish system. It has been argued that Scotland's relatively compact political geography and re-invigorated independent legislative capacity has, in recent decades, enabled a closer dialogue between thirdsector organizations (including those working in the violence against women sector) and law or policy makers. 21 Unlike in England and Wales, the shift towards a commissioning culture for campaigning and victim support services has been resisted. Organizations such as Scottish Women's Aid and Rape Crisis Scotland have received relatively secure government funding, without encountering the culture of competitive tendering that has been perceived as disruptive to, and divisive between, service providers elsewhere. This, together with an increasing recognition on the part of Scottish law and policy makers of those organizations' expertise, has facilitated the development of high-profile public awareness-raising campaigns, including Rape Crisis Scotland's 'This Is Not an Invitation to Rape Me' 22 and 'I Just Froze'. 23 Though resources to evaluate the impact of these campaigns have been limited, research conducted for Rape Crisis Scotland has indicated positive results, at least in the short to medium term, in respect of influencing social 15  attitudes. 24 Though not a primary aim of our project, the question of whether, and how, such campaigns influence jury deliberations is engaged in the present study. Our findings, set out below, give cause for some optimism, albeit suggesting that there is still substantial work to be done.

Methodological issues
In Scotland, s. 8 of the Contempt of Court Act 1981 precludes asking about 'statements made, opinions expressed, arguments advanced or votes cast by members of a jury in the course of their deliberations'. While such restrictions, common (but not universal) across many jurisdictions, honour the sanctity of the jury room, they generate a deliberative 'black hole' at the centre of the trial process. 25 While illuminating studies have gathered important data about the juror experience without straying into questions regarding the content of deliberations, 26 these limitations on researching decision making itself make it difficult to properly understand, evaluate, or more effectively facilitate the process of reaching a verdict.
To offer a glimpse inside the jury room, a range of methods have been developed. In the context of rape, a large number of studies have asked participants to respond to a series of abstract attitudinal questions, generating a field of literature centred on 'rape myth acceptance' scales. 27 This has demonstrated that a sizeable minority of people hold false and prejudicial beliefs about rape and rape victims. 28 Men are more likely to endorse such rape myths than women, 29 as are those with lower educational levels. 30 The vast majority of such work has been conducted with volunteer members of the public, but a recent study in England and Wales explored the attitudes of individuals (n = 771) who had just completed jury service, whether in a sexual offence or other type of trial. 31 This study used 'agree', 'disagree', or 'unsure' to capture views, rather than the five-or seven-point scales deployed by most previous research to capture subtleties in beliefs. Nonetheless, it found that 13 per cent of participants did not disagree that 'a rape probably didn't happen if the victim has no bruises or marks', that 15 per cent did not disagree that 'if a person doesn't physically fight back, you can't really say it is a rape', and that 27 per cent did not disagree that 'it is difficult to believe rape allegations that were not reported immediately'. 32 While the author suggests that this indicates that 'rape myth acceptance' may be less common among jurors than previous research with the public suggests, these findings also amplify the need for further exploration of the ways in which such attitudes may nonetheless still enter jury deliberations, and the potential impact that they might have in the context of discursive exchanges in which any doubt about the complainer's credibility has the potential to create substantial persuasive leverage in steering peers towards a collective acquittal.
Indeed, one of the difficulties with such attitudinal studies is that previous research has shown that even participants who have relatively low levels of 'rape myth acceptance' (based on their responses to questionnaires about abstract attitudes) can sometimes rely on or be persuaded by problematic views, grounded in those same stereotypes and misconceptions, when engaging in verdict deliberations. 33 A sizeable body of quantitative research has thus sought to explore the correlation between scores on 'rape myth acceptance' scales and judgments about blame or guilt in a specific scenario. 34 Since it is not possible to observe or analyse deliberations between jurors in real cases, researchers have relied on volunteers and simulations. While too much of this research in the past paid little attention to the reality of the trial context, relying, for example, on short written scenarios, the methodology utilized has become more sophisticated over time. Extended vignettes, or even live or video trial reconstructions, have become more common, sometimes combined with instructions to participants on the legal tests to be applied, and more studies have begun to factor in collective rather than merely individual decision making. 35 Though valuable in bridging some of the potential gap between participants' abstract attitudes and their application to concrete trial scenarios, much of this work has continued to pay insufficient attention to the content and dynamics of collective deliberative exchanges, and the ways in which careful, qualitative analysis of such discussions yields a more rounded understanding of how juror attitudes influence the outcomes of rape trials. 36 The present study sought to learn from, and build upon, this body of existing research, in order to explore mock jury deliberations in the Scottish criminal justice context. Of course, it should also be borne in mind that mock jury research has its own limitations. 37 Many of these can be avoided by replicating the real trial process as far as possible, 38 but the fact that participants know that they are role playing ultimately cannot be avoided, and this has raised concerns about whether participants will take their task sufficiently seriously. Nonetheless, there is evidence from existing studies that suggests that the fact that nobody's fate truly hangs in the balance does not prevent participants from engaging rigorously and conscientiously in deliberations. On the contrary, such research has reported heated and agonizing discussions among participants, with several 32 Id., p. 1002. remarking subsequently on how stressful they found deliberations, and suspension of disbelief to a marked degree in order to talk about the consequences for the parties. 39

Our methodological choices
The present study used a trial simulation method. A total of 64 mock juries (863 mock jurors) watched a filmed rape or assault trial reconstruction (lasting approximately 60 minutes) and deliberated for up to 90 minutes (without the presence of a researcher). Deliberations were audio and video recorded, and then subsequently transcribed, coded, and analysed. In addition, participants completed brief pre-and post-deliberation questionnaires.
In this article, we focus on the 32 juries (431 participants) who observed the rape trial video. 40 All of these juries watched the same trial simulation. However, in line with the aims of the main study, which focused on the unique features of the Scottish jury, 41 the conditions in which the juries deliberated varied. Half had the full range of verdicts available -guilty, not guilty, and not proven -while the remainder had to choose between guilty and not guilty. Those juries that had the option of not proven were given relevant judicial directions in line with those used in real trials. 42 Half of the juries had 15 members and the other half 12. Half were required to strive for a unanimous verdict, while the others were advised that they could return a verdict based on a simple majority (8 of 15 or 7 of 12). In unanimity conditions, if a jury was unable to reach a verdict after 70 minutes, they heard a supplementary judicial direction informing them that a near unanimous verdict (10 of 12 or 13 of 15) would be acceptable. If the jury remained unable to reach a verdict after 90 minutes, it was recorded as 'hung'. As noted above, these variables did not have a discernible impact on the number or type of evidential issues discussed across the juries, or on the tone and content of contributions regarding expected complainer behaviour, and so we analyse them as a cohort here.
The scenario for the rape trial involved a complainer and an accused who had been in an eightmonth relationship, which ended two months before the alleged offence. Both agreed that the break-up was cordial and that, in the intervening period, the complainer made two short telephone calls to the accused to ask if he would like to go for a drink with her and her friends (which he declined on both occasions), but that they had no other contact. On the night in question, the accused called at the complainer's home (which they previously shared) to collect some possessions. They each drank a glass of wine and some coffee as they chatted. A few hours later, as the accused made to leave, the two kissed. It was the prosecution's case that the accused then tried to initiate sexual intercourse with the complainer, touching her on the breast and thigh, and that the complainer made it clear that she did not consent to this by telling the accused to stop and pushing away his hands. The prosecution alleged that the accused ignored these protestations and went  40 There were 431 jurors rather than 432 due to a single participant who became unwell and left before deliberations commenced. 41 See text attached to ns 9-13 above. 42 All legal directions used in the study -including those on the not proven verdict -were adapted from the Judicial Institute for Scotland's Jury Manual, the latest version of which is available at <https://www.judiciary.scot/docs/ librariesprovider3/judiciarydocuments/judicial-institute-publications/jury_manual.pdf?sfvrsn=8c9918e4_6>. on to rape the complainer. The complainer testified that after the accused left, she telephoned her sister but got no answer, and, after 40 minutes, she telephoned the police to report the rape.
When questioned by the police, the accused admitted intercourse with the complainer, but maintained that this was consensual. He testified that after they had sex, he had told the complainer that he did not want to resume their relationship, at which point she became angry and told him to leave. A forensic examiner testified that the complainer had suffered bruising to her inner thighs and chest, and scratches to her breasts, that were consistent with the application of considerable force, but that -as was not uncommon in rape cases -there was no internal bruising. Her testimony concluded that these injuries were consistent with rape, but that she could not rule out alternative explanations.
The nature of our project and funding that we received from the Scottish Government meant that we were able to take steps to replicate the trial experience as far as possible. This is in contrast to previous studies, which have, often for understandable reasons of cost or convenience, relied on student samples and/or written stimuli rather than live re-enactments or videos, and have faced criticism as a result for a lack of realism. Among other things, in scripting the trials, the researchers worked closely with a group of legal advisors to ensure that the scenario and courtroom exchanges were as realistic in content, tone, and procedure as possible. The trial videos involved professional actors in the roles of witnesses and advocates. A retired judge acted as the trial judge, giving legal directions replicating those used in real trials. The trials were filmed in a real courtroom and legal practitioners were present throughout the filming to provide guidance. 43 We emphasized the solemnity of proceedings and jurors took the standard juror affirmation before viewing the film.
Our jurors were recruited from jury-service-eligible members of the general public, with the assistance of specialist recruiters from Ipsos MORI Scotland. Those recruiters worked to quotas to ensure that juries were representative of the Scottish population in terms of age, gender, education level, and work status. They also took account of any significant demographic differences in the local populations from which jurors at the Glasgow and Edinburgh High Courts would be summonsed, and modified where necessary to reflect, for example, the greater proportion of individuals from minority ethnic groups within Scotland's 'Central Belt' regions. While there is no research on the representativeness of Scottish juries, research in England and Wales has shown that juries there are 'remarkably representative of the local population'. 44 Potential participants were approached through door-to-door or street recruiting, rather than through reliance on any pre-existing lists of 'frequent' market research volunteers, and were not solicited to participate in response to any advertisement for participants on social media. 45 Juries were run on Saturdays over a two-month period in an effort to widen the pool of potential participants.
When approached by recruiters, potential participants were simply told that the study was 'to learn about the Scottish jury system' and that it would involve a four-hour session during which they would watch a trial video and then be asked to deliberate as a group towards a verdict. They were advised that 'we will be asking you to behave just as real jurors would, considering the evidence and deciding the verdict with the same degree of seriousness as you would in a real court' 43 A short extract of the trial is available to watch online at <https://youtu.be/kDAGaSedje8>. The judge's opening and closing directions can be viewed at <https://youtu.be/ecemRns-gDk>. judicial-institute/sites/judicial-institute/files/diversity-fairness-in-the-jury-system.pdf>. 45 Cf. for example Willmott, op. cit., n. 2, p. 227, where jurors were recruited by an advertisement that stated 'Ever wanted to sit on a jury? Never been asked? Strong views about crime? Now is your chance.' and -for ethical reasons -were informed in advance that they would watch either a rape or serious assault trial. Those who agreed to participate were paid £50 to cover any travel expenses and compensate them for the substantial time involved.
There remain, no doubt, limitations to this method, including -as noted above -the unavoidability of its role-playing nature. Nonetheless, in line with previous research, 46 we found that participants quickly became immersed in the task before them, engaging seriously with the trial reconstruction and approaching deliberations conscientiously. Jurors spent considerable time discussing the demeanour of the accused and the complainer, and in the vast majority of those discussions it was clear that they had suspended disbelief and forgotten that the roles were being played by actors. This assessment is underscored not only by their remarks in the debrief after deliberations about how difficult they had found the process, but also by the fact that in juries that were required only to reach a simple majority verdict, though the initial preponderance of verdict preferences was almost always such that they could have returned a decision almost immediately and left, participants always opted to stay and engage in detailed discussions.
Aside from the question of engaging seriously in the study, there is also, of course, the question of whether mock jurors might provide 'socially desirable' responses that they consider more palatable to the researchers since they are aware that their contributions are being recorded. We saw little evidence of this, to the extent that, once immersed in deliberative exchanges, a substantial range of views and perspectives were voiced, challenged, defended, and appraised within the jury room.
It is also important to bear in mind that there are advantages to simulation methods of the sort undertaken in this study -methods that could not be replicated in research with real juries, even if legislative prohibitions were relaxed to permit it. They enable researchers to hold constant factors within the trial scenario to isolate variables across different deliberating groups. This was particularly important in the present study since it allowed the effect of the unique features of the Scottish jury system to be investigated. By exposing a large number of participants to the same scenario -varied only by those procedural considerations -we were also able to reflect with higher levels of confidence on the substantive attitudes to the rape complaint exhibited by participants than would be possible by any reliance on real deliberations specific to individual trials.

JURY DELIBERATIONS: COMMON SENSE, CONSENT, AND CONSEQUENCES
Previous studies have documented a belief on the part of jurors that 'genuine' victims of rape will physically resist an attack, and consequently either display injuries themselves or be able to provide evidence of having injured an assailant. 47 In contrast to this belief, an extensive body of psychological literature establishes that individual reactions to traumatic events vary significantly; and that for many individuals a 'freezing' or dissociative response is common, as too may be the more intentional and calculated decision not to physically resist in the hope of avoiding additional violence to oneself or others. In addition, previous studies have suggested that members of the public hold other mistaken beliefs about rape victims, including the expectation that a 'true' victim would always report an attack to the police quickly 49 and would invariably be observably distressed while recounting events, 50 both at the time of the incident and at trial. Survey data has also demonstrated a high prevalence of 'victim-blaming' attitudes, such as suggesting that women should take some responsibility for having been raped where they have been drinking alcohol, flirting, or dressed 'provocatively'. 51 Finally, and allied in many respects to these expectations, research has also revealed a relatively frequent and resilient perception that women commonly make false rape allegations and that these are hard for a man to contest. 52 In this section, we explore how our findings support or challenge that existing research, drawing on the 32 rape trial deliberations, which involved 431 individual jurors. Transcripts of the deliberations were coded using NVivo in order to analyse their substantive content, while videos were observed to explore jury discussion dynamics. Though crude to the extent that one concise and well-executed interjection may have had more impact on deliberations than a number of lengthy, rambling, or repetitive ones, we also triangulated that analysis with indices of the frequency with which certain views were presented and challenged. Before turning to our findings, it is important to address a preliminary point about verdict outcomes.

Verdict outcomes: what lies beneath?
Across our 32 rape juries, four returned a verdict of guilty, 24 a verdict of acquittal (12 not guilty and 12 not proven), and four were hung. This pattern is different to the actual pattern of acquittals in Scottish courts, where the overall proportion of convictions is higher and where not proven accounts for a lower proportion of acquittal verdicts. 53 However, these results were not surprising. The scenario with which participants were presented was designed to be finely balanced, and in particular to make it likely that jurors who had the option of a not proven verdict would discuss its appropriateness. 54 Participants in the present study were inevitably provoked by the specificities of the particular trial scenario. To the extent that jurors' assessments of credibility are apt to be influenced by factors such as the 'likeability' of trial parties and perceived 'cues' associated with certain sociosexual interactions, their responses were also linked to choices that we made in the casting and direction of actors and the narrative of the scenario (parties in their twenties who had been in a previous relationship, with the incident taking place in the complainer's home). 55 Thus, while we can describe how our juries reacted, for example, to evidence that the complainer invited the accused into her home, we cannot say what difference that evidence made, since we did not run separate experiments involving different locations, contexts, or parties.
That is not to say that we cannot offer important insights into the ways in which our participants approached such factors. However, it speaks to the problems with seeking to uncritically extrapolate general claims from one trial simulation. It also highlights the difficulties of drawing firm conclusions about the attitudinal contours of jurors' discussions based solely on verdict outcomes -real or simulated -in a context in which a wide range of factors influence the translation of individual preferences into collective verdicts, including jury dynamics and the persuasiveness of trial advocates. Thus, while the verdict may initially appear to be a useful measure of what goes on in the jury room, it remains a crude index of the complexity, fluidity, and nuance of what lies beneath. It is often more illuminating, in terms of developing a sophisticated understanding of juror attitudes and how they interact with the relevant legal tests, to focus attention on the content, tone, and dynamics of deliberations.

4.2
Troubling the trinity: force, resistance, and injury

Force, resistance, injury, and consent
In line with previous studies, 56 we found that many participants were inclined to view as suspicious a failure on the part of the complainer to resist an attacker. It was frequently suggested that if the sexual activity in question were genuinely non-consensual, the complainer would have resisted physically, and in turn the assailant would have had to use extensive force. In our scenario, the complainer had bruising and scratching to her thighs and upper body, though there was no evidence of injury to the accused. Nonetheless, many jurors suggested that the complainer's injuries could have been sustained in alternative ways and so were of limited probative value, and/or maintained that they would have expected her to have sustained more serious injuries.
In 28 of the 32 juries, the view was expressed that a failure to physically resist an attack may be indicative of consent. For some jurors, this translated into an expectation that a 'genuine' victim would have injured the accused: 'surely if it was rape, you would expect scratch marks' (M01F); 57 'if you're being attacked . . . then to me you would scratch, you would scream, you would try and do anything possible to get him off' (M01G), or 'in spite of being weaker and smaller, she could have tried to defend herself -you can always defend' (M01H). For others, and despite expert testimony confirming that its absence was not uncommon in rape cases, internal trauma was expected: Even though the doctor said that it doesn't necessarily have to be internal trauma, if it had been sort of a violent rape like she was making [out], like there was bruises on her body, there would have been internal trauma as well. (M01E) 55 We did attempt to control for this as far as possible, through actor selection. For example, the actors chosen to play the roles of the prosecuting and defence advocates were both female and of a similar age and appearance. 56 See the sources cited in n. 5. 57 Each jury was given a unique code, which we use here to demonstrate that views were expressed across a wide range of juries. This is concerning in a context in which the law (and jury directions) on rape is now clear in Scotland: the offence requires a lack of consent on the part of the victim, coupled with a lack of any reasonable belief in consent on the part of the accused. 58 There is no requirement for the use of force or evidence of resistance. Indeed, in our trial scenario, the complainer had sustained a level of injury and bruising that was consistent with her account, even though it could have been inflicted in other ways. Nonetheless, jurors often expected higher levels of resistance and injury. The implications of this are significant given that in many rape cases there are no injuries at all.
Nonetheless, it is worth emphasizing that in the present research -seemingly more so than in previous studies, which involved a trial scenario with comparable levels of injury upon the complainer 59 -there was a degree of willingness on the part of jurors to challenge these expectations when voiced. In 79 per cent of the juries where these views were expressed (n = 22/28), they were directly challenged. We see examples in the following two exchanges: In some cases, these efforts at rebuttal drew on personal experience: In others, jurors pointed instead to peers' lack of personal experience: M: I would assume there would be a bit of shouting and screaming if she didn't want it, seriously. F: More of a struggle. M: Struggle, more bruises, he would need to use two hands if she was struggling against a man who is not that much bigger than her. F: Have you seen anybody that has actually been raped? Has anybody been raped before? M: I've never witnessed a rape, no, I've never witnessed one. F: Well, you can't say what a rape looks like if you have never witnessed one. (M04F)

Freezing reactions as an explanation for non-resistance
When challenging peers' preconceptions regarding the inevitability of injury and physical resistance, some jurors drew on their awareness of the fact -now well evidenced across psychological studies 60 -that people react differently to traumatic events and that freezing responses are common. This was raised in 25 of the 32 juries. It sometimes appeared to stem directly from the Rape Crisis Scotland 'I Just Froze' campaign. 61 One juror referenced 'that ad campaign saying "Don't feel guilty if you have just frozen because lots of people do"' (M01G). While other jurors did not express the connection to the campaign directly, they used language mirroring it. One said that 'everybody reacts totally differently' (M04F), while another noted that 'a lot of women are scared, some people just think they can't fight back or do anything, they get scared, they freeze up' (M05E). Another observed that 'it's a normal reaction to trauma to go into complete shock and to do absolutely nothing, your body just shuts down' (M06F).
It is difficult to assess the influence that such interjections had on the direction and outcome of deliberations. For jurors who were entrenched in their preconceptions, they may have been unpersuasive. There were certainly occasions where it was clear that the interventions had limited observable impact, with some participants maintaining their initial presumption that a 'genuine' victim would have struggled physically. Part of the difficulty here, as identified in previous research, 62 may be that for some jurors the feasibility of a freezing response was less plausible in a 'non-stranger rape'. A number suggested, for example, that 'I could understand that if it was a stranger, maybe she would have frozen more, but she knew him' (M06H), or noted that 'there is no history of violence' in order to infer a lesser level of fear compared to a stranger attack (M04F). Equally, however, there was a willingness among some jurors to reject the suggestion that freeze reactions would only be confined to 'stranger rapes'. Again, the scale of this debate and challenge in the jury room appeared to go beyond that seen in previous research, indicating perhaps the effects of the Rape Crisis Scotland campaign. That was reflected in the following exchanges: M: You're more likely to be frozen in fear if it is a complete stranger, an attack in the street, you can see that more so. M: Rather than somebody you have already had a relationship with as well. F: I don't agree with that. ( Struggling, screaming, and seeking assistance For many participants, there was an expectation that a victim, even if incapable of physical resistance, would still call out verbally for assistance. In our scenario, the complainer testified that she told the accused to stop but did not scream to get the attention or assistance of others. The incident took place in her flat, so this is perhaps not surprising given the likely futility of her doing so. Nonetheless, this became a prominent feature of discussion. The belief that a 'real' rape victim would shout for assistance was expressed in half of our 32 juries. As one juror put it, 'if you were getting held down and raped, you would be screaming your lungs out' (M02E); another insisted that 'if you're getting raped, you're going to shout, you're going to scream' (M02E). Some jurors brought this belief closer to home, indicating that it was advice that they would expect people connected to them to heed: I've got three daughters and I have always said that if anything happens to you, in the street or anything like, if you get attacked, you scream. Now she stays in a flat, right, if she is being attacked, a serious assault, why does she not scream? (M01G).
In contrast to the rates of rebuttal in respect of physical resistance and injury, this view in relation to calling verbally for assistance was only challenged in 44 per cent of juries where the view was expressed (n = 7/16). Where jurors did challenge it, it was often again on the basis that a person might freeze during a traumatic event, which would potentially also preclude them from shouting. There may, of course, be other reasons why a victim might be unable or consider it pointless or unwise to scream. Nonetheless, it was evident that such interventions did not always have an impact. For example, having raised the expectation that his daughters would scream, the juror above was challenged by peers who said that 'you don't know what you would do', but he maintained that 'I'm just saying this is how I am feeling. To me, you would scratch, you'd scream' (M01G).

Conundrums of corroboration and credibility
The implications of this discussion around resistance, injury, and calling out for assistance may be particularly pronounced in the Scottish context, where -as noted above -the corroboration rule dictates that, in a criminal case, there must be at least two sources of evidence in respect of each 'crucial fact'. 63 Our mock jurors were directed on this requirement, based on the guidance used by judges in directing real juries: The law lays down that nobody can be convicted on the evidence of one witness alone, no matter how strong that evidence may be. There must be evidence from at least two separate sources which you accept and which taken together point to the guilt of the accused. There are two essential matters that must be proved by corroborated evidence. These are that the crime charged was committed, and that the accused was responsible for committing it.
The corroborating evidence need not be unequivocal; it does not matter if there is a possible explanation for the potentially corroborating evidence that is inconsistent with guilt, as long as the jury, taking the evidence as a whole, is persuaded beyond reasonable doubt of the accused's guilt. 64 In our scenario, the complainer's testimony was accompanied by evidence from a forensic examiner who confirmed that her external injuries were consistent with rape but acknowledged that alternative explanations could not be ruled out. This sufficed in law to satisfy the corroboration requirement. However, in 23 of the 32 juries, the view was expressed that the evidence was inadequate for conviction. Jurors spoke of the medical evidence being '50/50' (M03E), 'not helping' (M01F), 'not having huge value' (M01F), not being 'very strong' (M02E), and not being 'valid as a categoric way of convicting' (M01H). Not all jurors mapped their evaluations directly onto the corroboration requirement. In many cases, it was pinned more loosely to a conclusion that there was insufficient evidence to be 'sure beyond reasonable doubt' which, despite judicial directions to think of this standard as 'less than certainty, but . . . more than a suspicion of guilt, and more than a probability of guilt', was often interpreted as requiring something considerably closer to, if not actually, 100 per cent certainty. For some jurors, however, it was the interaction with the corroboration requirement that provided a crucial first hurdle and ensured that an acquittal had to follow. As one put it: it's only facts that can be corroborated that should be taken into account. For me, the only person that could corroborate was the doctor, and the doctor was not able to state that these injuries were sustained as a result of rape, therefore she couldn't corroborate the evidence. (M04G) A similarly striking example came from another juror: 'that's it, there is no corroborating evidence, the injuries have been dismissed by the doctor, full stop' (M05H). In this context, it is perhaps less surprising that so many of the juries that had the not proven verdict at their disposal elected to make use of it, since this was interpreted by many participants as the verdict that was most appropriate when they suspected that the accused may have been guilty but there was not enough evidence to prove it, 65 or felt that a compromise was needed. 66 At the same time, however, this negative appraisal of the impact of the forensic examiner's testimony on the corroboration requirement did not always go unchallenged. In 74 per cent of the juries (n = 17/23) where the sufficiency of the evidence was questioned in this way, other jurors put forward a contrary view. This was sometimes done by suggesting that an impossible standard was being applied: I would be amazed if there is a crime that happens anywhere in the world, that physical evidence is provided for, that the defence can't turn round and say, 'But could something else not have caused that injury?' I don't think there's any injury that you can't say that for. (M01G) Alternatively, the challenge was made rhetorically by asking what else could have been offered as corroborating evidence: 'how are rape cases proven then?' and 'how is anybody going to prove a rape case?' (M05H) 64  While the tender under which this jury research was conducted arose as a result of the Post-Corroboration Safeguards Review, it was not designed to test directly for any impact of the corroboration requirement itself. As such, it is impossible to know whether, or how, jurors' deliberations might have differed if they had been asked to assess the evidence without this requirement. That said, it is worth reiterating that, for many jurors, the relationship between corroboration and the standard of proof was a very close one, and their ambivalence about the medical evidence translated more broadly into a feeling that there was a lack of adequate proof and a sense that the doctor's evidence was not sufficiently 'definitive'.

Context matters, but in what context?
Another prominent theme across deliberations was reflection on the behaviour of the complainer (and to a lesser extent the accused) in the period before the alleged attack. This was undertaken by jurors to test the credibility of competing claims regarding consent and to consider the extent to which the accused might have understood the complainer's conduct to signal consent. The fact that the complainer had offered the accused a glass of wine on his arrival, and the impact that alcohol may have had on both parties' subsequent behaviour, was a prominent focus. In 21 of the 32 juries, the wine was referred to in a way that went beyond merely fact setting; it was cited to make inferences about the parties' motives and behaviour. Jurors spoke, for example, of how the accused 'got something of a come-on' by the complainer giving him a 'royal reception', signified by the offer of wine (M01F). Many spent time debating whether that was a reasonable interpretation on his part: female jurors reflected on whether they would have offered alcohol to an ex-partner in this way; male jurors reflected on what meaning they would have given to such an offer. While some jurors emphasized, for example, that 'you wouldn't, when your ex is coming up to pick a TV up, pour wine . . . unless you want something' (M06H), others pointed to personal experience -'I have a relationship with my ex-boyfriend and I would offer him a glass of wine if he came over' (M05G) -and insisted that 'I don't feel as though if you offer somebody a glass of wine that means you want to have sex with them' (M05E).
Nonetheless, references to the complainer 'plying' the accused with drink persisted (M05E, M08F), with jurors often pointing to a perceived irresponsibility on her part for having offered alcohol, given that the accused was intending to drive afterwards: jurors maintained that 'I wouldn't offer somebody a glass of wine that came in a car to visit me' (M04F) and highlighted that the complainer 'let' (M05E) the accused drink 'when she knew he had a car outside' (M08F).
In addition, while it was agreed in evidence that the parties only drank one small glass of wine each, it was not uncommon for jurors to exaggerate the consumption, such as stating that they had 'a few glasses of wine' (M01G) or even that the accused had 'finished the bottle' (M04F), and such mistakes were not always corrected by peers. There were also references to them being 'steaming' (in other words, heavily intoxicated) (M06F) and an insistence by some jurors that even one glass of wine could have severely impacted both parties' reasoning processes, to make it more likely either that the complainer would consent to sex or that the accused would struggle to register her resistance. Such views were often juxtaposed alongside discussions about the relevance of the complainer having consensually kissed the accused. Jurors were, on the whole, relatively quick to challenge any assertion from peers that 'people can say the consent is her kissing him' (M01E) or that 'she kissed him back . . . so that might be like giving consent, sort of' (M05F). They typically did so by insisting that 'consent can be withdrawn at any given time' (M01E). However, while they were relatively clear that consent to a kiss does not equate to consent to sex, they were far more uncertain as to whether the kiss could have been sufficient to have 'confused' the accused (M01E) and made him mistakenly think that consent remained, even in the face of the complainer's objection. This is evidenced in the following exchange: M: Maybe he has thought that he has got permission to have sex because she has actually kissed him back. In this respect, some jurors insisted that 'we all know that actions speak louder than words' and that the kiss in itself might 'make a man feel that it's green, all go' (M03G). For some, this was allied to a perception that -for men -sex can be a mechanical process, and that if the accused was 'too passionate' (M01F), 'stuck in the moment' (M05E), or 'completely in his own world' (M07F), he may have struggled to register the complainer's refusals, and perhaps 'didn't even hear what she was saying over and over because he was so involved in what he was doing' (M07F). Though such behaviour seems far removed from the ideal of communicative sexuality that law and modern society seeks to encourage, jurors who took this view were often keen to underscore that the accused's actions would not have been 'premeditated' (M01G) and that he was not 'a big bad guy out on the streets roaming about wanting to rape a woman' (M01F). Instead, the suggestion was that the accused might have become 'overwhelmed by the kiss' (M01G) and, as a result, 'it's went hazy for him' (M01G) such that 'he has not bought into the concept of no is no' (M05E), he was not able to 'fully understand the meaning of stop' (M05E), or 'for whatever reason, he couldn't stop' (M01F).
While such assertions were often challenged by peers who insisted that it was nonetheless the accused's responsibility to stop, others maintained that it can be hard for men to mechanically shut down their sexual arousal and that 'once it starts, they hear nothing else except for when it comes to a conclusion' (M03G).
In addition, jurors often looked to the behaviour of the parties after the incident for further cues that might assist them to construct a narrative of what had taken place. In this respect, a key focus was the timing of the complainer's disclosure. The fact that she had delayed calling the policealbeit by only 40 minutes -was seen by some jurors as suspicious. This was raised as a concern in 13 of the 32 juries. As one juror put it, 'the time delay factor when she reported to the police for me was a bit sketchy' (M02G). This finding is in line with existing research, which has indicated that jurors can assume a lack of credibility where a report is delayed, even briefly. 68 In the present study, it is striking that this view was aired when (1) the delay was so short and (2) jurors received a direction from the trial judge advising that there can be good reasons why a person against whom a sexual offence is committed might delay in reporting, and that this does not necessarily indicate that the allegation is false. 69 At the same time, it is important to note that where suspicion regarding delay was raised, it was challenged in most cases. In ten of the 13 juries (77 per cent), it was questioned, often through direct reference to, or reliance upon, the judge's direction. As one juror put it, the judge 'said something about in the rape cases the time not being an issue. . . . [H]e said don't be swayed by that' (M01H). However, even in situations where it was challenged, this did not always assuage peers' suspicions. For example, when a female juror retorted to the suggestion above from her male peer that the 40-minute delay was 'sketchy' (despite his acknowledging that the complainer may have been in shock), reminding him that the complainer 'tried calling her sister and stuff in-between times', he maintained that 'I take that point, but I thought maybe the police would have been the first point of contact, so eventually I decided not guilty' (M02G).
For some jurors, it was not only the timing but also the complainer's decision to contact her sister first that raised suspicion, being seen as indicative of a plan to fabricate a false complaint (M03F). As one put it, the complainer 'feels she has been raped, so why would you phone your sister and not the police?' (M01F). Another observed that 'I'm literally on the edge of guilty and not guilty, but what I would like to know is why she phoned her sister before she called the cops' (M08F). Again, some jurors responded by maintaining that they too might, in that situation, have called a family member or friend before the police (M04E, M06G), while others insisted that it is impossible to know how one would react (M05E). However, the impact of such assertions was unclear. In one case, for example, while a number of jurors responded to peers' suspicions about the complainer's choice to call her sister with assertions that 'I would do the same', 'it's the right decision for any woman', and 'you would talk to your sister about it first', these suspicions were restated later in the deliberation by one juror who maintained that 'it's the phone call to the sister that kind of threw me' and that 'you would have phoned the police straight away I think' (M08F).
A further aspect of the complainer's behaviour in the aftermath of the alleged attack that merits mention, since it appeared to cause consternation among several jurors, was that she changed her clothes but did not shower or take a bath. There was a strong expectation on the part of many jurors, informed by media depictions, that someone who had been raped would feel compelled to shower. As one juror put it, why would she want to change her clothes? Right, she felt dirty, she felt used, but she never mentioned anything about going for a bath. Now in my experience in watching things on telly . . . that's one of the first things that a person would do after getting raped. . . . [W]hy change clothes and not go in for a bath if she felt that dirty and used? (M04G) Similarly, in another jury, it was insisted that 'after a woman has been raped, the first thing that she wants to do is to take a bath. . . . I was quite surprised that didn't happen because usually that's what they say is the first thing that a woman does' (M05E), while in another it was observed that 'we have all watched TV when it comes on and the first thing that the lady does is go and take a shower' (M06G).
At the same time, other jurors appreciated that taking a bath or shower might have detrimentally impacted on the forensic evidence and suggested that perhaps the complainer had this in mind. For example: F: If you're so traumatized you're not going to change your clothes, you just sort of stay how you are. F: Yes. M: No, I think you would change your clothes because you feel dirty. F: You would go straight to the shower. F: Yes, you shower. M: I'm not sure of the shower because . . . F: She would have got in her head that she was going to call the police. M: She would know that she is going to call the police. If she is going to basically call the police, then they're going to do an examination. Therefore she wouldn't shower, but she would change her clothes because . . . Though such discussions rarely came to a definitive outcome, they underscore the extent to which popular culture media -and the expectations of 'normal' behaviour that they propagate -can infiltrate the jury room and provide a litmus against which credibility is tested.

It happens all the time: the spectre of false allegations
The suggestion that false allegations of rape are common arose in 19 of the 32 juries (59 per cent), 70 often linked to a suggestion that the complainer's allegation could have been fabricated. Among the explanations afforded by jurors for why the complainer might have made a false complaint were that 'she wants her man to come back' (M01H), that she was 'obsessed' with the accused (M04F), that she might be a 'psycho bunny' out for revenge (M07F), or simply that 'women can be vindictive' (M04G), 'bad', 'mad', or 'horribly scorned' (M06H). Jurors who sought to bolster the feasibility of a false allegation sometimes suggested that the complainer's injuries could have been self-inflicted. As one put it, 'pressing yourself very hard to make yourself bruised, you can do it' (M01H), while another suggested that it was 'fairly easy to scratch yourself if you really want to' (M07G). It was here that expectations regarding resistance and injury coalesced with a perception of the regularity and ease of false complaints to produce a powerful scepticism.
Importantly, attempts were made across the juries to rebut the suggestion that false allegations are common. This claim was challenged in 74 per cent of the juries (n = 14/19) where it was made. Sometimes this was done via a flat rejection: 'it's very rare for a woman or anyone to make an accusation of rape to the police like that if it's not true' (M08E). Other jurors pointed to the difficult process involved in making a complaint: 'it's a big step to take, for women to take, unless it happened' (M05E).
Despite this, it was clear that for some jurors the prospect of false allegations still loomed large. As one concluded, having participated in a discussion in which the view was challenged by peers, 'yes, but some women do just use it as a tool' and 'you can always say "Why would they lie?", but people do' (M05E). Others maintained that 'there [are] hundreds of cases coming out where women have lied about rape' (M06H) or that it is a 'scary thing for the guy, I mean a terrifying thing for the guy' because 'it does happen' (M04F). Some who took this view went further, arguing 70 We excluded from these figures bare acknowledgements of the existence of false allegations; only statements that false allegations were for example 'routine' or 'common' were included. that 'there is a thing in modern society now, isn't there, because of the MeToo movement, that you believe a woman no matter what' (M06H) and suggesting that support to make an allegation can increase this risk: 'there are now organizations that encourage women to go forwards when they have been raped' (M05H).
This evidence of juror buy-in to concerns about false rape allegations, which has been identified in previous research, 71 suggests that the so-called 'MeToo watershed' may have had a limited impact. High-profile social media and public campaigns, including #MeToo but also #BelieveAll-Women and #IBelieveHer, have been cited by commentators as indicating a turn (and, some practitioners have argued, a turn too far) towards the presumed veracity of allegations and increased awareness of the pervasiveness of sexual victimization. 72 However, it seems that, within the jury room, the spectre of female fabrication still looms large, fuelled by disproportionate media coverage of rare but sensational cases in which allegations are shown to be false. 73 This stands in contrast to our more optimistic findings, discussed above, regarding the impact of targeted public information campaigns on the frequency of freeze responses, for example. More work is clearly required to understand what media and messages are likely to be most effective in informing prospective jurors. What this underscores, though, is that the mere existence of a high-profile public rhetoric that 'calls time' on sexual misconduct does not necessarily reflect a more embedded attitudinal shift that will take hold in the jury room. Indeed, the prevalence of such rhetoric may have counter-productive effects, creating complacency or even suspicion of 'band-wagoning'.

REFLECTIONS FOR THE ROUGH GROUND OF LAW REFORM
In the previous sections, we have highlighted the tenacity of participants' misconceptions regarding the inevitability of force and injury in 'genuine' rape complaints, and the extent to whichthough educational campaigns that highlight the feasibility of freeze responses appear to have penetrated the public consciousness to a degree -there is more work to be done in ensuring that jurors are appropriately informed about the complexities of sexual assault. We have also explored the ways in which -notwithstanding important shifts in cultural dialogue precipitated by social media movements, including #MeToo -socio-sexual scripts continue to circulate in the (mock) jury room that attribute responsibility to rape complainers for any 'mixed signals' that they may have communicated. These, and other behaviours -for example, failure to respond to an attack in expected ways -can translate into unwarranted suspicion over the veracity of the complainer's account, a suspicion that is legitimated further by continued misconceptions regarding the prevalence of false rape complaints. Ours is not the first mock jury study exploring the content and dynamics of jury deliberation in rape cases to yield such results, though it is the first on such a large scale, and the first in Scotland, where there are a number of distinctive legal, procedural, and social factors that interact with these juror attitudes. In other jurisdictions, concerns over such attitudes have led to the suggestion that juries might be removed from rape trials, either to be replaced by judge-only hearings or determination from a panel of experts. 74 This would constitute a substantial departure from current practice in Scotland. It would also raise challenging questions about the importance of public participation in the justice process and whether it is sustainable to restrict such departures to sexual cases. Its appeal as a solution is also dependent on confidence that alternative procedures for determining sexual assault complaints would be better positioned, in the final analysis, to ensure just results. 75 For now, at least, jury trials remain -and thus the question is one of what steps might be taken to increase the prospects for justice within that process. Proposals have been made to 'screen' prospective jurors to ensure that those with high levels of 'rape myth acceptance' are not eligible to deliberate in sexual assault cases. 76 This brings its own challenges, however, not only by undermining the fundamental principle of juries as a constituency of unfiltered peers, but in respect of its ability to contend with the ongoing potential for those who display low 'rape myth acceptance' in response to abstract attitudinal questionnaires to nonetheless rely upon, or be persuaded by, problematic assumptions in the concrete context of deliberation. 77 Moreover, while research is limited, there are indications from this study that well-resourced and carefully designed education campaigns can be effective in informing public attitudes and shifting the tone of jury deliberations, without any move to screening. This approach underscores, rather than forecloses, the importance of provoking public dialogue around consent and socio-sexual norms, with the ultimate objective being not only to ensure more informed jury deliberations but also fewer instances of sexual abuse.
At the outset, we identified two other features that are unique to the Scottish system that may have particular relevance in terms of how they interact with juror attitudes in rape trials: the corroboration requirement and the not proven verdict. In a context in which both features have been -and continue to be -the focus of considerable debate, it is worth reflecting on how our findings might inform those law and policy discussions. In respect of corroboration, although it is often thought to be a requirement that has most impact at early stages of the process, acting as a filter that prevents cases proceeding to trial where they depend on one piece of evidence only, it was clear that it featured in jury discussions in ways that suggested that it imposed additional obstacles to conviction. This was particularly so where jurors were sceptical of the value of the testimony provided by the medical examiner, upon which the prosecution's assertion of corroboration depended, or where they chose to interpret the requirement to be sure of guilt beyond reasonable doubt as requiring virtual certainty. As noted above, we did not test in this study a condition in which jurors were not directed regarding the requirement for corroborated evidence, and so cannot assert with any confidence what difference its removal might have made. We can, however, conclude that having been advised of it, some of our jurors relied upon it to support their 74 See for example New Zealand Law Commission, The Justice Response to Victims of Sexual Violence: Criminal Trials and arguments in favour of acquittal, and that this also interacted with their selection of a not proven verdict.
In 16 of the rape juries, participants had the choice between guilty, not guilty, and not proven verdicts. Twelve of these resulted in acquittals, and in every case this was via not proven rather than not guilty. As noted above, this is a substantially higher incidence of not proven than in real cases in the Scottish courts, reflecting the intentions of the researchers in devising the trial scenario. Nonetheless, it also underscores the concerns raised by many third-sector campaigners in Scotland that not proven verdicts are returned disproportionately in rape cases. 78 In other work, we have discussed in greater detail the reasons given by mock jurors for favouring a not proven verdict. 79 Among the key themes arising from that analysis were that some participants saw the verdict as a 'compromise' outcome, appropriate in cases where they believed the complainer but were not sufficiently sure to be able to convict the accused. 80 Jurors sometimes suggested that the not proven verdict gave them an opportunity to send a signal to the accused that he had not been fully exonerated (even though the consequence was still one of a complete acquittal) or to the complainer that they believed that she was probably telling the truth. 81 It is far from clear, however, based on narratives provided by complainers of rape who have received not proven verdicts, that these messages are being received, let alone valued. On the contrary, research suggests that those narratives speak of the ways in which a not proven verdict has undermined complainers' faith in the justice process, indicating recognition of their sexual victimization but a refusal to punish the perpetrator, and cementing a perception that their injuries do not matter. 82 The future of the not proven verdict in Scotland, though liable to be the subject of more scrutiny, is yet to be written. Its past is one of anomaly within the system, and its present is one in which its retention as part of a modern legal system requires increasing justification. Crucially, though, its operation cannot and should not be understood in a vacuum. It interacts in complicated ways with other structural features, including the corroboration requirement and the capacity to return verdicts by simple majority, as well as the complex and sometimes contradictory norms, assumptions, and attitudes that -as this article has shown -jurors often bring with them as they participate in, and forge the outcomes of, deliberations.

CONCLUSION
In this article, we have presented findings from the largest research study of the nature of mock jury deliberations in rape cases undertaken in the UK to date -and the first to be undertaken in Scotland. We found considerable evidence of the expression of problematic attitudes towards rape complainers during deliberations. These included the belief that a 'real' victim would have extensive external and internal injuries and would resist by inflicting injuries on her attacker and shouting for help, that even a short delay in reporting is suspicious, and that false allegations are commonly made by women and difficult to refute. These findings are consistent with many previous studies that have explored more abstract levels of 'rape myth acceptance' and bolster the existing evidence base that jury deliberations in rape cases may proceed on problematic assumptions about complaints and complainers. However, there was also some evidence -more than in previous research -that jurors in our study were willing to challenge problematic attitudes, sometimes drawing on third-sector campaigns to do so. We have argued that this gives cause for measured optimism in regard to the potential for better-informed deliberations, and thus strengthens the case for resisting abandonment of trial by jury, or any proposed move to juror screening, in sexual assault cases. Equally, however, we have emphasized that the scale of the challenge should not be underestimated. Misconceptions and prejudicial attitudes of the sort discussed in this article were not systematically challenged in all juries, and even where they were, it was hard to confidently determine the extent of their lasting impact upon peers or verdicts. It was clear that -even in this 'post-MeToo moment' -suspicion in respect of the veracity of women's claims of sexual victimization often remained alive in the (mock) jury room. In the distinctive Scottish context, the ways in which such attitudes interact in concrete cases with the existence of the not proven verdict, as well as the corroboration requirement, to produce particular outcomes are likely to be complicated -but our findings suggest that this is a pressing issue that cannot be ignored any longer.

A C K N O W L E D G E M E N T S
We are enormously grateful to all who participated in the jury simulations. Particular thanks are also due to Rachel Ormston and Lorraine Murray (Ipsos MORI Scotland) who led the Scottish Government tender and with whom we produced the project report.