3. Jurors’ Perceptions of Pre-Recorded Evidence Given by Child Witnesses
This chapter outlines key findings arising from a number of studies that have focused on the impact of the use of special measures in general, and live-links and pre-recorded testimony in particular, by child witnesses. Broadly speaking, these have demonstrated that jurors are not significantly better able to reach the truth and discern signs of deception when children testify in open court as compared to testifying via live-link or pre-recorded testimony. Despite this, some – but by no means all – studies have suggested that jurors may prefer children’s testimony to be delivered live in court. There is no clear evidence, however, that this impacts in any significant way upon verdict outcomes. Group deliberation appears to be key in this respect, reducing any pre-deliberation propensity by jurors to rate less highly the credibility and likeability of child witnesses who give evidence via live-link or pre-recorded video. Research has also established, though, that findings in respect of child witnesses are complicated by the interaction of mode of evidence delivery with factors such as the perceived emotionality of the child, his or her age and level of understanding, and jurors’ preconceptions regarding the reliability of children’s memory, particularly over time.
It has been argued that “the traditional adversarial cross-examination is not a reliable method of either testing the truthfulness of what a child has previously said, or of obtaining from them further information that is accurate….(and) it is potentially abusive” (Spencer, 2012: 178). Giving evidence in criminal trials has been shown repeatedly to risk adverse mental, physical and psychological effects on child witnesses (Goodman et al, 1992; Hamlyn et al, 2004). Moreover, previous research on memory has shown that while all witnesses forget information over time, younger children are more susceptible to forgetting than older children and adults (Flin et al, 1992; Gudjonsson & Henry, 2003), and may also be more likely to guess answers to questions when they are unsure (Waterman & Blades, 2013; O’Neill & Zajac, 2013).
For many commentators and practitioners, such findings point in the direction of the routine use of special measures in respect of child witnesses. Enabling the child to give testimony without entering the intimidating environment of the courtroom or coming into direct contact with, or view of, the accused has been said to reduce witness stress and, consequently, to increase the ability to obtain ‘best evidence’. In addition, the use of pre-recorded evidence in this context has been thought to have an added advantage in terms of being able to capture an account that is more contemporaneous (in many cases) to the alleged incident, and thus more likely to be complete than one given some months later. But concerns have been expressed about whether the use of such modes of testimony places the accused at a disadvantage in trial proceedings by positioning the witness as particularly frail and affected by events in the minds of the jury or, conversely, reduces the prospects of a child witness being believed and empathised with by a jury who are not able to observe them ‘in the flesh’.
This chapter will first outline and evaluate the findings of existing research in relation to any impact, as a result of the use of pre-recorded and live-linked testimony by child witnesses, on jurors’ ability to detect witnesses’ deception and accuracy. Having done so, it will then turn to explore any evident effects in relation to jurors’ perceptions of witnesses’ credibility. It will consider any evidence pointing to an impact on verdict outcomes as a consequence of these factors, and reflect on any specific conditions associated with children’s testimony that may be particularly relevant to an evaluation of this evidence.
3.2 Jurors’ Ability to Detect Deception
In his evaluation of arguments for and against the use of video technology in relation to child witness testimony in England and Wales, Davies notes that “advocates of conventional cross-examination argue that much useful non-verbal information is lost to the jury by the restricted view of the witness afforded by the conventional television shot” (1999: 244). This concern is grounded in an assumption, of course, that jurors are capable of accurately decoding, and drawing appropriate inferences from, the nonverbal cues they discern from seeing the witness ‘in the flesh’. A number of studies have set out to investigate the reliability of this assumption and, broadly speaking, they have consistently established that observers often make mistakes in their decoding and interpreting of nonverbal cues, and that the likelihood of an untrained third party being able to detect truth on the basis of a witness’ nonverbal behaviour is no different to chance (DePaulo et al, 1985; Ekman, 1985).
In a relatively early study, conducted by Clifford et al (1992), children were observed being questioned, either live from behind a one-way screen or via CCTV, in respect of a classroom incident they had witnessed eight months previously. Observers’ ratings of witness credibility and accuracy showed little relationship to actual accuracy and there was no systematic impact on those ratings or their accuracy depending on the medium of observation in 9 of the 10 areas of testimony probed. The one exception to this was the “attractiveness” of the witness (meaning likeability rather than physical attractiveness), where evidence given in the courtroom by the child witness did lead to higher ratings from jurors than CCTV testimony. This is a consideration that will be discussed further below.
Similarly, in a study by Tobey at al (1995), 1,201 jurors were asked to judge the overall accuracy of children’s testimony (aged 6 and 8 years old), regarding a play session with an adult where stickers were placed either on exposed areas of the children’s bodies or on their clothes. Juror assessments were found to be poorly correlated to actual accuracy with no discernible differences between judgments made in regard to testimony given in or out of the courtroom (see also Goodman et al, 1992).
Likewise, in a study conducted by Orcutt et al (2001), 70 children aged between 7 to 9 years old played games with an adult male, during the course of which a video was made, in some cases involving the children showing their bare upper arms, bare toes and bare belly-buttons to the camera and in others showing only their shirt sleeves, shoes and belt buckles. Three weeks later, the children were asked to come to a courthouse and testify in open court or via CCTV about the play session (with some children being told to deliberately lie about their experiences). 987 jurors recruited from the local community were separated into groups (with an average of 12 members) to observe the children’s testimony within the broader context of a live mock trial reconstruction. Thereafter, they completed individual questionnaires and were allowed up to 30 minutes within their groups to discuss and arrive at a verdict. Across this fairly elaborate study, it was found that jurors were no more likely to detect deceit when the child’s testimony was delivered in open court than when it was delivered via CCTV.
This finding in respect of children’s live-linked CCTV testimony was replicated in a subsequent study by Goodman et al (2006) which explored the impact of pre-recorded video evidence. Here, 370 mock jurors observed a live reconstruction of a trial involving an alleged touching of a child witness (aged between 5 and 7) in the course of a ‘Simon Says’ play session with an adult male. Half of the children were telling the truth about having been touched on their noses, necks and bare stomachs during this game, while the other half were not. Across the study, half of witnesses gave live testimony and the other half used a pre-recorded video deposition with a social worker as evidence-in-chief. Having observed the trial reconstruction, participants completed questionnaires and then deliberated in groups ranging from 7 to 12 members towards a collective verdict for up to 30 minutes. On the basis of resultant data, the researchers in this study concluded that “adults, when faced with the task of determining whether unauthorised touching of a child occurred, were poor at distinguishing whether a particular child was lying or telling the truth. Furthermore, adults’ abilities were neither helped nor hindered, for the most part, by seeing the child live or on videotape” (390).
Most recently, Landström and Granhag (2010) have explored the impact of live, CCTV and videotaped evidence together. 108 child witnesses (aged 10-11 years old) were divided into groups: 54 children participated in an actual event where they had an encounter with an adult male as they posted a letter near the gates of their school, while the remainder were read an account of this event and asked to imagine that it had happened to them. 3 weeks later, the children were asked to testify, and those who had not actually experienced the event were asked to report that they had. Witnesses were then divided across live, CCTV and pre-recorded evidence conditions. In total, 240 jurors took part in the study, 86% of whom were undergraduate students. They were asked to complete questionnaires but not to deliberate as a group towards a verdict. In respect of the adult observers’ ability to accurately distinguish truth from falsehoods, the researchers concluded – in line with previous research - that no significant differences could be discerned across these modes of testimony.
Taken together, this provides a fairly compelling evidence base for rejecting any suggestion that the removal of children from the courtroom, and the presentation of their evidence on a TV screen, in and of itself, reduces the capacity of (mock) jurors to accurately decipher behavioural cues and discern truth from falsehood. Moreover, as Davies (1999) has argued, to the extent that the factual content of the testimony is likely to be a better guide to veracity than nonverbal behaviour, the suggestion in some previous research that jurors will pay more attention to that content when it is provided via TV link (Miller & Fontes, 1979; Clifford et al, 1992) may make use of this medium of testimony positively advantageous in terms of the process of truth-seeking.
3.3 Jurors’ Empathy and Assessments of Credibility
A separate concern frequently raised by criminal justice practitioners relates to the impact of the use of live-links or pre-recorded evidence on jurors’ assessments of the trial parties’ credibility, and, relatedly, on the level of empathy that jurors can feel for a witness who is not physically present in the courtroom but instead displayed in ‘unreal’ form on a TV screen. From as early as the mid 1980s, commentators have suggested that the use of TV and video medium may create an ‘emotional distance’ that will make it harder for the jury to connect with the witness and believe his or her testimony (MacFarlane & Krebs, 1986; Sharp, 1989: 96; for discussion see Davies, 1999: 246). This is a concern that continues to be expressed, for example, in the comment of one legal expert, cited in the Stern Review, that “juries prefer theatre to film” (2010:90).
There is not, however, a consistent evidence base to support this. Amongst the earliest studies to be conducted on this topic, Swim et al (1993) used a videotaped trial simulation in which the key witness, an 8 year old child, gave evidence in relation to an alleged sexual assault by her stepfather either live in court or via a videotaped deposition. 143 volunteering undergraduate students were asked to watch the trial simulation and then, alongside engaging in collective deliberations for up to 45 minutes, were asked to provide individual ratings of the parties’ guilt and credibility at various stages. While live testimony was viewed somewhat more favourably by mock jurors (reflected in their scores for witness accuracy, consistency and confidence), the difference was not statistically significant, and the researchers uncovered little discernible impact in terms of overall attitudes towards the accused or trial outcomes.
Similarly, in a study by Ross et al (1994), a videotaped trial simulation involving a 10 year old child giving evidence in relation to an alleged sexual assault by her father, either in front of the accused at court, in court behind a screen, or via a live-link, was utilised. Here, the medium of evidence presentation was found to have had no overall impact on the pattern of guilty verdicts returned by the 300 undergraduate students who participated as mock jurors, and nor did it influence the perceived credibility of the parties in any significant way. Interestingly, in this study however, when the simulation was re-run, with a different cohort of 300 student participants being asked to provide ratings immediately after the child had given testimony, rather than at the end of the trial simulation, the researchers found that they were significantly more likely to convict the accused in cases where the child gave testimony in open court.
Ross et al concluded that while there may be some transient impact in the aftermath of live testimony that predisposes jurors towards conviction, when the evidence is situated in the context of the whole trial, this impact was of little significance, and mode of evidence delivery was ultimately largely neutral in its effects. Though there was no deliberation component included in this study, this finding about the significance of the time at which influence is captured also speaks to the importance of including a deliberation requirement, which has been suggested in other research to similarly contextualise and mediate any juror preference for live testimony.
In both of these studies, researchers evaluated the impact of TV-linked or video-recorded testimony by showing jurors a videotaped trial simulation. While this can test the influence associated with the fact that non-standard modes of evidence were used, the situation of a video medium within a video prevents researchers from evaluating the impact of the screen per se, in contrast to ‘in the flesh’ testimony (Swim et al, 1993: 606). These difficulties are not, however, associated with a study conducted by Tobey et al (1995), in which a live trial re-enactment was used, with similar results to those uncovered by Swim et al (1993) and Ross et al (1994). Here, as discussed above, a trial was staged before 1,201 mock jurors (across the entirety of the study), with children aged 6 or 8 years old giving testimony about a play session with an adult in which stickers had been placed either on exposed parts of their bodies or their clothing. While jurors in this study rated the evidence given by children in the open court condition as significantly more accurate, believable and honest than that given via CCTV, in line with previous research, the medium of presentation of evidence was found to have had no overall impact upon the proportion of guilty verdicts returned.
This lack of transferability between mock jurors’ assessments of credibility and their final verdicts is supported, moreover, by other research. In a study by Goodman et al (1998), for example, children aged either 5-6 or 8-9 years old participated in a play session with an adult male and then, about 2 weeks later, were asked to testify about it in a courthouse setting in front of mock juries comprised of volunteer members of the public (1,201 over the course of the study). Compared with children who testified live in court, children who testified via CCTV tended to be viewed as less distressed and more consistent in their testimony, but they also tended to be viewed as less credible. Nonetheless, the researchers found no significant relationship between mode of evidence delivery and verdict outcome.
Similarly, in the study by Orcutt et al (2001) mentioned above, in which children aged 7-9 years old testified about an incident in which they made a video with an adult male, researchers found that while witnesses were perceived to be less accurate, less attractive, less intelligent, less honest and more likely to have made up their story when they testified via CCTV rather than in the courtroom, and this did provoke a reduced proclivity to convict amongst mock jurors drawn from the local community at the predeliberation stage, after a 30 minute deliberation period, the researchers reported that “there was no significant main effect of trial condition on postdeliberation verdicts” (2001: 366). This suggests, then, that any such negative effects – though initially potentially significant - may be amenable to reappraisal and rebalancing as a result of a broader group discussion.
In respect of the use of pre-recorded video deposition specifically, results are similarly complicated. In a study by Eaton et al (2001), 108 volunteers, of whom 92 were undergraduate students, observed a 20 minute videotape of a simulated sexual abuse trial involving an 11 year old victim and then recorded their perceptions of the witness’ credibility and the accused’s guilt on a 5 point scale before deliberating for up to 20 minutes in groups to reach a verdict. The researchers found significantly lower ratings of overall credibility when the child gave testimony via CCTV live-link as compared to either giving evidence in person in the courtroom or via a video deposition conducted by a female psychologist. Again, however, there is some suggestion that the strength of this impact upon credibility had diminished by the close of deliberations.
Similarly, in Goodman et al’s 2006 study in which 370 jurors observed a live trial reconstruction relating to the alleged touching of a child (aged 5-7 years old), with trials varying between live evidence and the use of a video deposition, it was found that “trial condition [negatively] affected the perceived credibility of the child…and influenced jurors’ sympathy for the child which, in turn, contributed to guilt judgments.” But despite this, it was also acknowledged by researchers that, by the end of jury deliberations, “no significant effects involving trial condition or child condition emerged” (2006: 387).
Overall, then, while some previous studies have indicated that there is no significant impact associated with the use of live-link or pre-recorded evidence modes of delivery vis-à-vis live testimony, others have suggested that there is a discernible impact – typically, but not always entirely consistently, in the direction of rendering the child witness less attractive and less credible.However, these studies have typically failed to find any reliable or significant impact on verdict outcomes, particularly when a deliberative component is built into the study design. This finding is supported by a study conducted in England and Wales for the Home Office in the wake of the introduction of video-evidence under the 1991 Criminal Justice Act , in which 93 real trials involving child witnesses (with an average age of 12 years) were observed to assess the content and style of questioning, and the apparent competence and demeanour of the witness, across live examination and videotaped evidence conditions. It found no significant differences as between the proportion of guilty verdicts delivered (Davies et al, 1995).
3.4 Jurors’ Responses to Emotion, Age and Memory
The interaction between mode of delivery and the (real or expected) emotionality of witnesses’ testimony poses further difficulties. Previous research has established an ‘Emotional Victim Effect’ whereby victim-witnesses who display (an acceptable level of) negative emotion, such as sadness or agitation, when testifying tend to be perceived by observers as more credible than those who display a more neutral or controlled manner (Kaufmann et al, 2003; Ask & Landström, 2010; Ellison & Munro, 2009a, 2009b). Much of this research has focused on adult female rape complainers. Indeed, a recent study which failed to replicate this finding in relation to an adult male testifying about physical assault has raised the question of whether the ‘Emotional Victim Effect’ correlates specifically to gender or to offences that are sexual in nature, which third parties may believe to be “particularly emotional events” (Landström et al, 2015a: 103). In respect of child witnesses, the evidence is more limited, but support for an ‘Emotional Victim Effect’ has been found in relation to allegations both of bullying (Ask & Landström, 2010) and physical assault (Wessel et al, 2013).
The existence of this research poses a dilemma for simulation studies. On the one hand, keeping the emotional demeanour of the trial parties constant permits a more precise isolation of the effects of changes in the mode of evidence delivery for experimental purposes. On the other hand, incorporating some differentiation in demeanour across conditions may provide a more accurate reflection of how they would unfold in the courtroom. The complexity and diversity of individual reactions to trauma further complicate this picture: a witness may, for example, be more emotional in pre-recorded evidence given the proximity in time to the incident, or calmer because she is not in the courtroom in the presence of the accused. While the majority of studies have opted for a consistent demeanour across evidence conditions, this may limit their applicability to ‘real life’ cases where the tone of pre-recorded testimony is likely to be (indeed, is designed to be) quite different from its live in-court equivalent.
This relates, of course, not only to observable emotional demeanour, but also to the levels of accuracy and confidence that observers expect witnesses to be able to provide in recollecting and recounting the alleged incident using different testimony conditions. Certainly in respect of child witnesses, the interaction between these factors remains unclear. In a study by McAuliff and Kovera (2012), 261 jury service eligible volunteers from the local community completed questionnaires regarding their expectations of the emotionality that would be displayed by child witnesses (aged variously 5, 10 and 15 years old) who gave evidence via different modes of delivery. The researchers reported that participants typically expected a child using a live-link to be more confident and cooperative, provide longer responses and be less ‘fidgety’ than a child giving testimony live in court. But these expectations sit at odds with the bulk of previous research exploring the impact on children’s observable behaviour across these conditions in actual trials, where no reliable differences emerged in terms of children’s confidence, cooperation, ability to answer questions, amount of detail provided or levels of concentration, whether they testified in court or via a live-link (Murray, 1995; Davies et al, 1995). The researchers hypothesised that this may cause observers to judge children in live-link conditions more harshly when they fail to evidence the increased levels of confidence and comfort expected; and that it is this mistaken assumption, rather than the medium of delivery per se, that is central.
Closely bound up with these expectations regarding emotionality, behaviour and reliability of memory is the question of the relevance of the age of the child witness. Once again, however, there is little consistency in existing research on the impact of age on a third party’s evaluations of credibility. A number of studies indicate that as age increases so too do witnesses’ assessment of credibility (Goodman et al, 1987; Leippe & Romanczyk, 1987, 1989). However, in other studies, differences in age have been shown to have no, or even the reverse, effect on the verdict given (Castelli et al, 2005; McCauley & Parker, 2001). Indeed, Nikonova and Ogloff (2005) have illustrated that witness age can have positive, negative and neutral effects – in their work, children were regarded as more ‘trustworthy’ than young adults, and children under the age of 7 were deemed to have higher suggestibility and lower cognitive competence but higher accuracy than children over the age of 10. Despite this, age had no reliable or consistent impact on verdict outcome.
Much here may depend on the nature of the alleged violation, or the delay in time between its incident and its recounting. In cases where children have alleged physical assault (Pozzulo & Dempsey, 2009) or sexual assault (Ross et al, 2003), they have been rated in some previous studies as more credible than adults, but the opposite has been found in cases involving road traffic associated manslaughter (Goodman et al, 1987) (see also Krahenbuhl, 2012). Meanwhile, Antrobus et al (2016) have found that, across a sample of 102 student participants presented with a written sexual assault trial vignette involving a child witness, those who rated evidence presented via CCTV as less credible (c.f. pre-recorded evidence) also tended to hold the strongest pre-existing views about children’s memory being unreliable over time, mapping on to a preference for pre-recorded and more contemporaneous forms of testimony. Trial outcomes may thus be influenced by jurors’ assumptions regarding the reliability of children’s memories, which interact with mode of testimony.
3.5 Jurors’ Perceptions in Jurisdictional Context
It is important to bear in mind that this research exploring the impact on jurors of children’s pre-recorded or live-link testimony has been designed and conducted across a range of legal jurisdictions, each with their own distinctive procedural rules for capturing and including such evidence. This makes reading across the studies to create a reliable evidence base difficult and it means that any findings which are based on the procedures of one jurisdiction may not translate reliably to another.
The CCTV condition used in both the Goodman et al (1998) and Orcutt et al (2001) studies mentioned above, for example, involved a one-way link in which the witness sat in a separate room within the court building, and counsel for the prosecution and defence individually entered the room to examine and cross-examine, whilst the judge remained in the courtroom with the jury and defendant. In the study by Ross et al (1994), however, the judge also relocated along with the attorneys, whilst the defendant and jury remained in the courtroom to watch the testimony via a one-way link. Both of these models differ, of course, from the two-way CCTV system that is utilised in other jurisdictions, including Scotland, where the witness gives testimony from a remote location but with questions being directed by counsel in the courtroom.
Considerably less research has been conducted to date on the impact upon jurors of this latter mode of live-link delivery by child witnesses. Indeed, Landström & Granhag, in 2010, were amongst the first to do so. The study involved a mix of boy and girl witnesses, aged 10-11 years old, who testified as part of a broader trial re-enactment regarding an encounter they had had with an adult when posting a letter outside their school. The researchers found – in line with some previous research utilising a one-way CCTV model – that across the 240 mock jurors involved in the study, the majority of whom were undergraduate students, participants tended to perceive children in the live evidence condition more positively than children in the live-link condition, who in turn were perceived more positively than those who testified via pre-recorded video.
Unfortunately, however, no deliberation component was included. This meant that the impact on collective verdict outcomes was not explored, despite its centrality in a ‘real’ case. This is a significant limitation, particularly in light of previous research which often indicates a difference between individual and collective decision-making. Indeed, as Landström et al (2007) point out in relation to a prior study, which identified a similar effect in a comparison of evidence given by 10-11 year old child witnesses live in the courtroom versus via video testimony, “showing that the influence of one presentation mode is undue requires more in terms of experimental sophistication than showing enhanced influence. It is for future research to investigate whether one presentation mode leads to too much (or too little) influence on, for example, how jurors weigh oral testimony, which might result in biased verdicts” (2007: 345). In other words, the fact that a degree of impact – generally against finding the child witness credible – has been found in some studies that use both 1-way and 2-way CCTV models does not mean that this has an ultimately prejudicial influence upon trial outcomes, and research exploring how this translates during and after deliberation in 2-way live-links has yet to be adequately undertaken.
Similarly, in respect of video evidence, in some studies (e.g. Eaton et al, 2001) this took the form of an advance statement provided by the witness to a professional, such as a social worker, whilst in others (e.g. Swim et al, 1993), it took the form of an advance trial deposition held in the judge’s chambers and attended by the witness, judge and counsel on both sides. In both cases, this is some way removed from the sort of video testimony that is routinely utilised in many jurisdictions which rely primarily on video-recorded forensic interviews conducted by police officers as part of the investigation process. Again, there has been insufficient research conducted involving this particular mode of pre-recorded evidence for us to be confident of its impact upon jurors. The findings of Davies et al (1999) in respect of real trials in England and Wales suggest that, however its use may impact upon jurors’ evaluations of credibility and honesty, it does not translate into a consistent shift in conviction proclivity.
3.6 England and Wales: Section 28 Pilot
England and Wales have recently moved towards more expansive use of pre-recorded evidence, implementing section 28 of the Youth Justice and Criminal Evidence Act 1999 to combine the established use of recorded police forensic interviews as evidence-in-chief with a videotaped cross-examination conducted at a later date by the defence in the presence of the judge and prosecution counsel. Child witnesses were a key constituency included in this section 28 pilot in 2014. An evaluation of its implementation, which involved interviews with 40 practitioners and 16 witnesses (along with their parents or carers as appropriate), concluded that while the process of being cross-examined continued to be stressful and difficult for witnesses, section 28 reduced levels of distress and trauma, and encouraged a more focused set of questions from defence counsel.
Importantly, the introduction of pre-recorded cross-examination was bound up with a broader shift towards greater case management, and with a requirement to commit to an expedited timescale for ensuring that parties were ‘trial ready’ and to hold a ‘ground rules’ hearing so that questions to be put to the witness by counsel could be appraised in advance. It was also influenced by the growing currency of advocacy toolkits for responding to vulnerability, and by a decision of the Lord Chief Justice in relation to child witnesses, which sought to adjust the tone of cross-examination prior to, or at, court ( R v B  EWCA Crim 4 at ).
In order to assess the impact of section 28 upon jurors an evaluation would need to consider how these factors – in tandem – produced, or failed to produce, an effect. However, the influence of the use of pre-recorded examination and cross-examination upon jurors was not included in the Ministry of Justice Report. The only mention of this influence comes in a brief remark that four of the practitioners interviewed suggested that the complainant felt too remote and that this was apt to damage the prosecution case, whilst six observed that there had been a high level of (in their view ‘right’) convictions, arguably undermining these concerns (2016: 38). Beyond this, there was no detailed consideration of jurors’ perceptions. Nevertheless, on the basis of this pilot, England and Wales have opted to roll out section 28 to all vulnerable witnesses. Further research is clearly needed to explore in more detail the impact upon jurors of this type of model of evidence delivery.