False beliefs on crime rates influencing juries?

Jurors often have to make decisions about whether they believe a complainant’s or defendant’s account of an event. However, research has shown laypeople are not very good at determining whether another person is telling the truth or not, or whether they are accurate or inaccurate more generally. This reality is unsurprising. There are at best few reliable objective cues in descriptions of events that indicate whether an account is true or false. Even inconsistencies in testimony, which have traditionally been thought of as clear indicators of deception, are now thought to provide little to no insight into the overall accuracy of testimony (aside from of the specific inconsistent statement). When jurors do pick up on cues in testimony, these cues can reasonably be interpreted in different ways. For example, inconsistencies might be perceived as the result of stress or natural memory decline, or as indicating a witness is not being honest. In light of this subjectivity, context is likely to heavily influence how cues in testimony are interpreted, the perceived credibility of a witness and accuracy of their testimony – and, ultimately, the legal verdict that will follow. It is therefore important to understand the context surrounding juror assessments of testimony, and how this context is likely to influence decision-making.

New and original research

In a current project, funded by UK Research and Innovation, my post-doctoral research assistant Dr Bethany Growns and I are examining the context surrounding assessments of testimony. We are looking at how that context is likely to be influencing juror decisions, and whether that influence is justifiable legally, or requires correction. In a recent study as part of that project, we examined how juror estimates of the underlying commonality of relevant events can influence their assessments of the testimony of witnesses. Psychological theory relating to juror decision-making suggests that this context is likely to be relevant to evaluations ­since it is relevant to the plausibility of witness accounts. We also examined how differences in estimates of commonality could explain differences in evaluations of testimony and differences in legal verdicts between jurors of different genders – and with differing political opinions. In the study, we surveyed about 550 adults asking them to give us their estimates of the commonality of a range of legally relevant events – and to provide us with their evaluations of two short sets of case materials – one from a case involving an alleged child sexual assault, and another from a case involving a domestic violence related defence to homicide.

The results

The responses provided by our participants offered important insight into the perceived commonality of a set of legally important events – and hinted at both potential inaccuracy and bias in these perceptions. For example, one of the questions we asked our participants was how often they believed women were abused by their partners. Participants were asked to estimate the number of women who have been abused by their partners (on a scale from more than one in every ten women to less than one in every million women).

Responses ranged from more than one in every ten women to about one in every million, indicating huge variability in estimates. On average, participants rated commonality at about one in every 55 women. This represents a significant underestimate of commonality compared to many current evidence-based estimates. The charity Refuge suggested that almost one in three women aged 16-59 will experience domestic abuse in her lifetime. Responses were also polarised based on gender, with male participants rating domestic abuse as significantly less common, on average, than female participants.

A correlation?

Importantly, participant’s prevalence estimates (including estimates of the commonality of child sexual assault and false allegations of child sexual assault, and psychological harm resulting from domestic abuse and fabrication of psychological harm) were significant predictors of their evaluations of case materials. For example, in the child sexual assault case, participants who believed child sexual assault was more common, and/or false allegations of child sexual assault were less common, generally rated a complainant alleging child sexual assault as being more honest, accurate, and believable, rated a defendant alleging a false allegation as less honest, accurate, and believable, and were more likely to indicate that they believed the defendant was guilty. Differences in prevalence estimates between male and female participants also partially explained differences in evidence evaluation and verdicts based on gender. For example, women were more likely to indicate that they believed the defendant was guilty in the child sexual assault case, which was partly explained by the fact that women rated child sexual assault as being more prevalent.

What does this mean?

These findings suggest background context of perceived commonality of alleged events is likely to feed into juror evaluations of testimony. It is important, therefore, to understand this background context, and also to be alert to when it might have a biasing impact on a jury – and how it might be accounted for. The influence of perceived commonality has the potential to be problematic for at least two reasons. First, there is the issue some estimates of commonality are clearly inaccurate, such beliefs that domestic abuse or sexual assault are extremely uncommon. Inaccuracy may also be introduced or enhanced by high-profile events and media coverage, which can sway people’s opinions, often by making comparatively rare events (e.g. false allegations) appear common. Second, the influence might be seen as problematic – even where estimates of commonality are reasonable. On the one hand, taking account of commonality appears a reasonable and logical thing for a juror to do – after all, more common events are more probable – and therefore are a more plausible explanation of underlying facts. On the other hand, accounting for commonality has the potential to lead to the discounting of testimony concerning relatively rare events – and relatedly, to create discrimination against people more frequently associated with wrongdoing, and people less frequently associated with being harmed. This discrimination may be based on characteristics protected by the Equality Act. 2010 For example, the fact that domestic violence against women is more common could lead to discrimination against male complainants. As practitioners, it might be helpful for us to look out for cases in which the impact of prevalence estimates might have problematic effects in either of these two regards.

Is there a gender effect?

One clear implication of our work is it highlights the importance of gender balance on juries, particularly in cases where perceptions of commonality are likely to differ along gender lines (e.g., in cases involving sexual offences). This balance minimises the risk of bias, including bias resulting from polarisation in perceptions of commonality. More broadly, our work in this field is contributing to a wider body of literature undermining the presumption that laypeople are necessarily well-placed to determine whether others are lying or telling the truth, particularly to the standard required in criminal litigation. Careful analysis of how juries should – and do – function in their role as lie detectors is necessary, to ensure they are functioning in a way that is normatively desirable or at least acceptable. Where juries are not functioning as required, evidence-based changes in practice and procedure can help to bring decisions back in line with normative goals.

Overall

Gaining a good understanding of how juries function as lie detectors and how procedure can help them function better is particularly important as governments seek to deal more effectively with cases involving sexual offences, which can be highly reliant on defendant and complainant testimony due to the lack of other witnesses and corroborating evidence. Understanding underlying estimates of commonality is one piece of this puzzle.

Dr Rebecca Helm is director of the Evidence-Based Justice Lab and a Law Clinic solicitor at the University of Exeter. Her research draws on techniques and insight from psychology and data science to understand how the law and legal procedures operate in practice: evidencebasedjustice.exeter.ac.uk