Jurimetrics

A major report - Jury Reasoning in Joint and Separate Trials of Institutional Child Sexual Abuse: An Empirical Study - by Jane Goodman-Delahunty and Annie Cossins for the Royal Commission into Institutional Responses to Child Sexual Abuse considers 

the extent to which joint trials with cross-admissible tendency evidence infringed defendants’ rights, and the extent to which joint trials posed a risk of unfair prejudice to the defendant. In particular, we investigated the reasoning processes of juries in a simulated joint trial of sex offences involving three complainants versus a separate trial involving a single complainant.

The authors state that

Our jury deliberation and reasoning study investigated these issues by presenting 10 different versions of a videotaped trial involving the same core evidence to a total of 1,029 jury-eligible mock jurors. The study tested the impact of evidence strength, the number of charges and the presence of specific judicial directions on jury decision-making in joint versus separate trials.

The five key aims of the project were to:

1. document juries’ interpretation of cross-admissible evidence in a joint child sexual abuse trial, to determine the extent to which juries engage in impermissible reasoning regarding such evidence 

2. compare the above decision-making processes with those of juries in a separate trial involving the same defendant 

3. compare trial outcomes (acquittal, conviction or hung jury) in a joint versus separate trial involving the same defendant 

4. examine the relationship between jurors’ misconceptions about child sexual abuse, jury deliberations and decisions, and trial outcomes 

5. determine the effect of question trail use on juries’ reasoning and decisions.

Previous research

The research identified three types of potentially unfairly prejudicial reasoning in cases of joinder: (a) inter-case conflation of the evidence; (b) accumulation prejudice through accumulation of counts or witnesses; and (c) character prejudice. Past studies yielded a ‘joinder effect’ in the form of increased conviction rates when at least three similar crimes were joined in a single trial compared to when separate trials were held for these offences. No prior experimental studies of joinder examined jury decisions in cases of child sexual abuse. Past research focused almost exclusively on conviction rates and failed to distinguish logically related permissible reasoning on the one hand, from logically unrelated and impermissible uses of the evidence in joint trials on the other. Whether observed joinder effects were due to permissible or impermissible and unfairly prejudicial reasoning remains unknown.

The jury deliberation and reasoning study

To address methodological limitations of previous research, and provide empirical evidence on the issues raised by tendency evidence, we used an experimental jury simulation approach to examine the relationship between jury decision making and trial outcomes in a joint trial of alleged child sexual abuse. The trial type was one of four variations: (a) Separate trial with an adult male complainant with moderately strong evidence (a basic separate trial) (b) Separate trial with an adult male complainant with moderately strong evidence, in which relationship evidence about the defendant’s uncharged sexual acts and grooming behaviours was presented (a relationship evidence trial) (c) Separate trial with an adult male complainant with moderately strong evidence, in which tendency evidence from two prosecution witnesses was admitted (a tendency evidence trial) (d) A joint trial involving the same defendant and three adult male complainants, who gave weak, strong and moderately strong evidence, respectively (a joint trial).

Judicial directions were defined as one of five variations: (a) Standard jury directions (b) Standard jury directions plus a context evidence direction (c) Standard jury directions plus a context evidence direction with a question trail (d) Standard jury directions plus a tendency evidence direction (e) Standard jury directions plus a tendency evidence direction with a question trail.

A total of 1,029 mock jurors – 580 women and 449 men aged between 18 and 82 years – were randomly allocated to one of 90 juries to view an experimental trial and deliberate to a verdict with fellow jurors.

Key findings – Chapter 4

Part 4.1: The influence of mock jurors’ pre-trial expectations and attitudes

The first step in the analysis was to assess individual mock jurors’ pre-trial expectations and attitudes. The aim was to examine the contribution of their individual differences to jury reasoning and decision making, and ensure that observed differences in responses to the trials were the result of changes in the trial information and not due to pre-existing differences in the mock jurors assigned to any particular trial group. The results show that the more mock jurors knew about child sexual abuse, the less likely they were to endorse other types of pre-trial bias. Accurate recall of the case facts was higher among mock jurors with more accurate knowledge about child sexual abuse, and lower among those with high expectations of forensic evidence being presented at trial. Mock jurors with higher educational achievement were less likely to expect forensic evidence at trial. Mock jurors who were more knowledgeable about factors that influence a complainant’s reports of child sexual abuse – and who favoured the prosecution – rated the complainant as more credible.

Part 4.2: The influence of the trial type on jury reasoning and verdicts

The study compares jury reasoning and decisions across four different types of trials: a basic separate trial, a relationship evidence trial, a tendency evidence trial and a joint trial. These analyses are based on individual mock juror responses to a written questionnaire completed at the conclusion of their jury deliberations. The purpose of these quantitative analyses is to explore whether there was a ‘joinder’ effect and whether the verdicts were motivated by permissible or impermissible reasoning.

Was there a joinder effect?

As more inculpatory evidence against the defendant was added to the trials, conviction rates for both non-penetrative and penetrative offences against the focal complainant increased. Conviction rates in separate trials with relationship evidence and tendency evidence were significantly higher than those in the basic separate trial. However, there were no significant differences between conviction rates in the tendency evidence trial compared to the joint trial. Thus, we did not identify a joinder effect.

The findings demonstrated that increases in the culpability of the defendant and the credibility of the focal complainant were most prominent in response to sources of evidence that were independent of the focal complainant, and did not increase merely when more evidence was added or the claims were presented in a joint trial. In deliberations about the basic separate and relationship evidence trial, mock jurors were more likely to express the view that the evidence was unpersuasive because it was simply one person’s word against that of another.

A major finding was that as more independent sources of evidence were introduced to support the focal complainant’s account, the complainant’s credibility increased, he was perceived as more convincing and his evidence was accorded more weight. Thus, there were no significant differences in the assessed credibility of the focal complainant in the basic separate trial versus the relationship evidence trial, although the addition of the relationship evidence increased the plausibility of the complainant’s account and his evidence was rated as significantly more convincing. In line with this finding, mock jurors were more likely to blame the complainant in the basic separate trial than in any other type of trial. Similarly, ratings of the defendant’s sexual interest in boys, inferences about his criminal intent and the factual culpability of the defendant were lowest in the basic separate trial and increased significantly in the tendency evidence and joint trials; that is, as more inculpatory evidence against the defendant was admitted.

Jury deliberations significantly increased ratings of the defendant’s criminal intent and factual culpability in trials that involved relationship and tendency evidence. The inferred criminal intent of the defendant predicted the verdict at both juror and jury levels, irrespective of the type of offence. In the absence of tendency evidence, juries were more reluctant to convict for the more serious penetrative offences. Jury distinctions between penetrative and non- penetrative offences confirmed that they reasoned separately about the counts, making distinctions between the counts relating to the same complainant. The presence of tendency evidence increased convictions for both the non-penetrative and penetrative offences, in both separate and joint trials.

Were juries in joint trials more susceptible to inter-case conflation of the evidence?

We tested mock jurors’ recall accuracy by asking multiple-choice questions about the case of the focal complainant. Results showed that trial complexity, not trial type, predicted the accuracy of factual recall. Accuracy on these questions was greatest in the less complex trials where only two witnesses appeared for the prosecution (an average of three errors), and decreased as more witnesses appeared for the prosecution, in both the tendency evidence and joint trials (an average of four errors). Mock jurors’ formal education had no effect on their factual recall accuracy. Across all juries, individual mock jurors who made fewer errors on the multiple-choice questions were more prone to acquit, and individual mock jurors who made more errors were more prone to convict, but this was unrelated to the type of trial.

Part 4.3: Jury reasoning by type of trial

To supplement the quantitative analyses reported in Part 2, we conducted a series of additional analyses using other sources of data to gain further insight into jury reasoning and decision making. Results reported in this section are drawn from quantitative and qualitative analyses of the content of the jury deliberations; open-ended responses by individual mock jurors about the main reasons for their verdicts; and a case study of jury reasoning in joint trials. These analyses focused on the prevalence of impermissible reasoning and jury susceptibility to unfair prejudice against the defendant. Contrary to expectations, juries in this study were not prone to impermissible reasoning and made very few factual errors. Most errors were corrected in jury deliberations.

The prevalence of impermissible reasoning in jury deliberations

A quantitative analysis of the content of jury deliberations in which all statements that might indicate unfair prejudice against the defendant were coded revealed that impermissible reasoning was rare, and when it might have occurred, it was more likely in the separate trials without tendency evidence than in the trials with tendency evidence.

We found a low rate of factual errors in jury deliberations; only 7.7 per cent of juries made more than two factual errors. Two or more factual errors were more likely to occur in jury deliberations about the joint trial; that is, the trial with the most complex evidence. When errors were made, the vast majority were corrected in the course of deliberations, demonstrating the ability of jury groups to self-correct. None of the 90 jury verdicts were based on inter-case conflation of the evidence.

We found no evidence of emotional or illogical reasoning by juries in any of the trials in which tendency evidence was admitted. We found only two jurors who appeared to use a lower standard of proof than the criminal standard, and only two jurors whose verdicts were driven by emotion. None of the juries featured a juror who reasoned illogically about the evidence.

A qualitative analysis of individual mock jurors’ main reasons for their verdict revealed that 90 per cent of the decisions to convict were based on the consistency of evidence from multiple witnesses, the credibility of the witnesses and the pattern of grooming behaviour engaged in by the defendant. Reasons that might indicate character prejudice as a reason for conviction were less than 3 per cent. These findings were supplemented by a qualitative thematic analysis of jury deliberations about the focal complainant in 33 joint trials, which did not uncover any conviction based on character prejudice.

Overall, our analyses of the reasons for decisions to convict provided negligible support for the notion that joint trials produce verdicts based on inter-case conflation of the evidence, character prejudice or accumulation prejudice. As instructed by the trial judge, mock jurors used their common knowledge and experience of the world in understanding the behaviours of the complainants and the defendant. Together, these findings provided no support for the hypothesis that joint trials lead to impermissible reasoning.

Part 4.4: Were juries in joint trials susceptible to accumulation prejudice?

This section tests the hypothesis that juries in a joint trial use the overall number of charges or witnesses to determine the guilt of the defendant. The results provid no support for the hypothesis that impermissible reasoning was triggered by accumulation of the counts or witnesses against the defendant. This conclusion was based on separate statistical analyses conducted on the accumulation of counts and accumulation of witnesses. Together, convergent results of quantitative and qualitative analyses on each issue confirmed that jurors and juries made logical and appropriate distinctions between the same types of offence alleged by different complainants, based on the strength of the evidence.

The findings demonstrate that the culpability of the defendant was predicted by mock jurors’ assessments of the credibility of the complainants, not the overall number of counts and witnesses. We found no reliance on reasoning by accumulation in a joint trial, as there was no significant increase in conviction rates or in the defendant’s factual culpability for allegations by the focal complainant in trials with six counts versus those with two counts.

Similarly, the addition of two prosecution witnesses in a joint trial did not increase conviction rates and, most notably, did not elevate the conviction rate for the complainant with the weak claim. In addition, mock juror ratings of victim blame did not vary in response to increases in the number of Crown witnesses, as might be expected if jurors were improperly accumulating the evidence. Rather, victim blame was predicted by individual mock jurors’ misconceptions about child sexual abuse. Results of coding the content of the jury deliberations revealed no impermissible reasoning or reduction in the onus of proof in trials with more counts. As juries were exposed to more witnesses and their cognitive load increased, they made more factual errors, but there were no observed differences in uncorrected or persistent errors across trials; that is, the results in separate and joint trials were similar. A case study of deliberations in a joint trial showed that juries in trials with six counts devoted most available deliberation time to the weak claim where the disparities in evidence were greatest, controverting the view that juries would gloss over these differences in a joint trial. A further case study of deliberations in a joint trial confirmed that no jury decision to convict or acquit was based on impermissible reasoning about the tendency evidence.

Part 4.5: The influence of jury directions on jury reasoning and decision making

In this section we examine whether there was support for judicial assumptions about the effectiveness of jury directions in reducing impermissible reasoning. As was noted above, we did not find that mock jury verdicts were based on impermissible reasoning. Nonetheless, we compared jury reasoning with and without specific jury directions provided to jurors in the relationship evidence trial regarding the uses of context evidence, and with and without specific jury directions on the uses of tendency evidence, provided to juries in the tendency evidence and joint trials. In addition, following the jury deliberations, we asked mock jurors a series of questions about how helpful the directions were.

The findings in this study are in line with a large body of empirical research demonstrating the ineffectiveness of most jury directions. Systematic statistical comparisons of jury reasoning and decisions in the relationship evidence trial, tendency evidence trial and joint trial accompanied by standard directions (on the one hand) and specific directions on the uses of relationship evidence and tendency evidence (on the other) yielded few differences. Overall, the relationship evidence direction was more effective than the tendency evidence direction, which produced no apparent benefits, irrespective of whether it was provided in a separate or a joint trial. Analyses of the content of jury deliberations revealed that error rates in using the context evidence and the tendency evidence were unaffected by the presence of these directions. More deliberation time was devoted to discussing ‘beyond reasonable doubt’ when standard jury directions were given than when juries received tendency directions.

The context evidence direction helped juries overcome their reluctance to convict for the penetrative offence, but the rate of conviction by juries for the non-penetrative offence was unaffected, although factual culpability ratings on both counts increased significantly in the presence of the context evidence direction. Consistent with the findings reported in Part 4.2, conviction rates were predicted by the convincingness of the complainant, irrespective of the presence of the direction on context evidence. In addition, convictions for penetrative offences in trials with tendency evidence were predicted by higher child sexual abuse knowledge on the part of individual jurors, not the jury directions. In both separate and joint trials with tendency evidence, the judge’s tendency evidence direction had no significant influence on the verdict, inferred criminal intent or the factual culpability of the defendant.

Self-report measures provided by mock jurors following their deliberations revealed that mock jurors who received context directions as opposed to the standard directions perceived the judge’s instructions as more confusing; found it more difficult to assess witness credibility and apply the law; reported a higher cognitive load; and felt that the judicial instructions made it harder to understand the charges, recall the facts, weigh the evidence and assess the case for the prosecution. Similarly, compared to the standard directions, mock jurors rated tendency evidence directions as more difficult to understand, and perceived that these directions increased their cognitive load. However, mock jurors rated the charges as easier to understand when they were given tendency evidence directions in a joint trial than when they were not given these directions.

Part 4.6: The influence of question trails on jury reasoning and decision making

In this section, we examine the influence of a question trail on jury reasoning and decision making, to discern whether this assisted the juries in their deliberations. Overall, using a question trail appeared to increase the efficiency of jury decision making. The main finding was that using a question trail reduced the overall duration of deliberation in relationship evidence trials, where deliberations persisted far longer in the absence of a question trail. Mock jurors who used a question trail reported that they required significantly less cognitive effort to reach a unanimous verdict than was the case among those who deliberated without this aid.

The question trail had no influence on mock jurors’ memory of the case facts. Separate analyses conducted on the relationship evidence trial showed that with the aid of a question trail, the defendant was rated significantly less factually culpable, and accordingly, the conviction rate for both the penetrative and the non-penetrative offences declined. Content analysis of deliberations in those trials revealed that with a question trail, a significantly greater proportion of deliberation time was devoted to discussing the counts and the judge’s instructions. When given a question trail, the mock jurors perceived that they required less cognitive effort to evaluate the defence case. This difference may account for the observed verdict shift from hung juries to acquittals.

Separate analyses conducted on the joint trial revealed that a question trail had no significant influence on the defendant’s factual culpability or on conviction rates, regardless of the evidence strength or offence type. However, mock juries reported significantly more difficulty in understanding the charges in a joint trial when given a question trail than when deliberating without one.

Part 4.7: Self-reported cognitive effort by type of trial

In this project, the complexity of the four types of trials varied. Accordingly, the cognitive load imposed on the mock jurors varied by trial type, and was greater in the tendency evidence and joint trials than in the basic separate and relationship evidence trials. Following their deliberations, all mock jurors responded to a series of questions about the extent of effort they had expended in reasoning about the case and coming to a verdict. To gain further insight into mock jury reasoning and decision making, in this section we present the results of mock jurors’ self-reports about the difficulty of their tasks, by trial type. As might be expected, mock jurors perceived that recalling the case facts was significantly more demanding in the tendency evidence trial than the basic separate trial, and that understanding jury instructions was more difficult in the joint trial than in the basic separate trial. Mock jurors reported that it required more effort to understand the charges as more inculpatory evidence was admitted.

Unexpectedly, mock jurors perceived that significantly more cognitive effort was required in the separate trial with relationship evidence and the tendency evidence trial than in a basic separate trial, while the joint and basic separate trials were perceived as requiring equivalent effort. The same pattern held for the tasks of assessing witness credibility, weighing the evidence, and evaluating the case for the prosecution and defence. Mock jurors rated reaching a unanimous verdict as significantly more difficult in the relationship evidence trial than in the basic separate trial. They rated deliberation in the relationship evidence and joint trials as more useful in understanding the case than in basic separate trials and in tendency evidence trials. Finally, mock jurors reported that deliberation significantly increased their confidence in the verdict reached.

Part 4.8: Fairness of the trial

A primary concern when considering the use of separate versus joint trials in child sexual abuse cases is the fairness of the trial to the defendant. In this section, we present a series of analyses that assess the perceptions of juries of the fairness of the trial, by type of trial. These analyses draw on mock jurors’ post-trial responses to a range of questions about the fairness of the trial, their expectation that they would be informed of any prior offending by the defendant, and the threshold they applied in interpreting the standard ‘beyond reasonable doubt’. In addition, we draw on some coding of the content of jury deliberations.

The main outcome of these analyses is a series of convergent findings showing that mock jurors rated the joint trial as more fair to the defendant than the basic separate trial. As we expected, mock jurors inferred more criminal intent on the part of the defendant as more inculpatory evidence was admitted in the different types of trials, and intent was rated as equivalent in the tendency evidence and joint trials. These ratings show that mock jurors made a logical analysis of the inculpatory evidence presented in the different types of trials. Other results are unexpected. First among them is the finding that mock jurors viewed the basic separate trial as significantly less fair to the defendant than trials that included more inculpatory evidence. Secondly, the defendant was rated as significantly less convincing in the separate trial with relationship evidence than in the joint trial with tendency evidence. Similarly, the mock jurors perceived the instructions from the judge as significantly less fair to the defendant in the basic separate trial than in the joint trial. A fourth measure that reflected an unexpected difference compared to what might be anticipated was the mock jurors’ interpretation of the threshold ‘beyond reasonable doubt’ applied to convict the defendant. In the basic separate trial, the threshold applied was significantly lower (85.2 per cent) than that in the joint trial (92.1 per cent).

Finally, with respect to information about prior offending by the defendant, a substantial proportion (three-fifths) of the mock jurors expected that they would have been informed at trial of any prior child sexual abuse incidents, charges or convictions involving the defendant. Significantly more mock jurors who attended a separate trial believed that if other charges had been made against the defendant, they would have been informed. In the course of jury deliberation, our content analysis reveals that concern about prior allegations against the defendant were rarely expressed, and no significant relationship existed between trial type and comments made by mock jurors that they would or would not have been informed of prior allegations of sexual misconduct.

Discussion – Chapter 5

Part 5.1: Was there a joinder effect?

As expected, we found that conviction rates varied according to the strength of the inculpatory evidence presented at each type of trial. Conviction rates increased with the admission of more inculpatory tendency evidence. Since these increases in the conviction rate occurred in both the tendency evidence trial and the joint trial, these findings do not support the hypothesised joinder effect. Although the conviction rates by juries and individual jurors in the joint trial were, on average, higher than those in the tendency evidence trial, these increases were not statistically significant, and were not due to the type of trial; that is, they were not due to the joinder of counts in the joint trial. In other words, we did not find a significant joinder effect.

Importantly, we did not find that the verdicts rendered were based on impermissible or prejudicial jury reasoning. Our analysis of credibility ratings confirmed that juries were sensitive to the source of additional prosecution evidence in assessing witness credibility. We can attribute increases in credibility ratings to systematic and permissible reasoning based on the probative value of the tendency evidence. Multiple convergent findings showed that jury decision making in the tendency evidence trial was similar to that in a joint trial, indicating that the juries were not reasoning in an illogical and superficial manner in the joint trial when given cross-admissible tendency evidence, compared to the tendency evidence trial which involved one complainant and two witnesses who gave similar accounts of sexual abuse by the defendant. The admission of the tendency evidence, whether in the context of a separate or a joint trial, did not lead to impermissible reasoning.

Part 5.2: Were convictions in joint trials the result of impermissible reasoning?

Inter-case conflation of the evidence To test the hypothesis that jurors would confuse or conflate the evidence tendered in support of different counts in joint trials, we compared the accuracy of jurors’ factual recall and their factual culpability ratings of the defendant.

Accuracy of factual recall

The deliberations revealed that more jurors made factual errors in trials with tendency evidence than they did in trials without tendency evidence, driven in part by the higher number of witnesses in those trials. Jurors’ mean recall accuracy scores decreased as trial complexity increased, indicating that the complexity of the trial evidence rather than joinder, per se, significantly predicted factual recall accuracy. Because juries promptly corrected their inaccuracies, we found no support for the hypothesis that persistent uncorrected errors were a feature of jury decision making in trials with tendency evidence, so no evidence emerged that errors of this nature had any causal effect on jury verdicts.

Compared to a real trial, where there is considerable repetition and more opportunity for juries to discuss the evidence and deliberate, our experimental simulations may have fostered the potential for more confusion than would arise in a real trial. This element of the trial simulation is likely to have contributed to the higher factual error rate observed in cases with more complex evidence. Nonetheless, in real child sexual abuse trials, whether separate or joint, a similar potential for confusion cannot be discounted as result of, for example, trial length, juror fatigue, juror disinterest, changing levels of concentration and trial complexity.

Factual culpability

The observed pattern of factual culpability ratings showed that juries relied on more systematic reasoning, rather than susceptibility to evidentiary conflation. This evidence of jury reasoning in response to additional evidence of the defendant’s other criminal misconduct controverts the hypothesis that juries in joint trials or in trials with complex tendency evidence engaged in impermissible prejudicial reasoning because of inter-case conflation of the evidence.

Accumulation prejudice

To examine the extent to which elevated conviction rates in joint trials with tendency evidence were attributable to impermissible reasoning, we tested whether juries were prone to convict based on the overall number of charges against the defendant or the overall number of witnesses called by the prosecution.

Multiple counts

Courts have hypothesised that a defendant will be unfairly prejudiced in joint trials because juries are prone to reasoning that the defendant is guilty simply because of the number of charges brought by the prosecution. To test whether juries were affected by the number of counts in the joint trial, we compared two trials in which the same evidence was presented by four witnesses called by the prosecution. The only salient difference between the trials was the number of charges against the defendant. Factual culpability ratings differed by count and by complainant according to evidence strength, independently of the type of offence, so we did not find any accumulation prejudice as a result of multiple counts. As with verdicts, the factual culpability ratings reflected that juries were able to evaluate the culpability of the defendant for each separate count according to different evidentiary strength. Thus, juries displayed the ability to distinguish between the evidence of different complainants.

Multiple witnesses

Courts have hypothesised that juries are susceptible to the cumulative effects of multiple witnesses, which are expected to increase in joint trials. This version of the accumulation prejudice hypothesis holds that a defendant will be unfairly prejudiced because juries are prone to reasoning that the defendant is guilty simply because of the number of witnesses appearing for the prosecution. To test jury susceptibility to the accumulative effect of multiple witnesses, we examined conviction rates when either four or six witnesses appeared for the prosecution in a joint trial. The addition of two prosecution witnesses did not significantly increase conviction rates, or ratings of the defendant’s factual culpability, providing no support for the accumulation hypotheses. Most importantly, the presence of these witnesses did not increase the conviction rate for the count with the weakest evidence. These findings directly controverted the accumulation prejudice hypotheses in relation to multiple witnesses, by indicating that both jurors and juries evaluated the evidence of multiple witnesses based on its probative value, not simply the number of witnesses.

Character prejudice

Character prejudice arises when a juror uses the severity or number of allegations of criminal misconduct by the defendant to reason that the defendant is a person of bad character, and is therefore probably guilty of the current charges. Encompassed within this concept is the hypothesis that juries will be less concerned about convicting because the defendant deserves punishment for the prior misconduct, charged or uncharged.

The admission of inculpatory evidence about four other acts of sexual abuse from two additional independent witnesses, irrespective of whether they were witnesses or complainants, did not diminish the ratings of how convincing the defendant was, suggesting that jurors were not engaging in impermissible reasoning on the basis of character prejudice. If they had, these ratings would have differed significantly between the separate trial and the trials with tendency evidence, in which juries were exposed to evidence of the defendant’s other acts of sexual abuse. Thematic evaluation of the jury deliberations revealed that no juries in either the tendency evidence or joint trials impermissibly used the tendency evidence to conclude that the defendant was guilty because of the number of allegations of prior misconduct made. Furthermore, there was no evidence of verdicts motivated by emotional reactions to the severity of the allegations, such as a sense of horror regarding the allegations, or a desire to punish the defendant. Ratings of the credibility and convincingness of the focal complainant showed that the complainant’s credibility was enhanced by the evidence from independent witnesses or complainants who reported similar criminal conduct by the defendant, irrespective of whether the defendant was charged with counts pertaining to those individuals. Similarly, jurors’ ratings of the convincingness of the focal complainant were significantly higher when tendency evidence was admitted, compared to the separate and relationship evidence trials that had no tendency evidence.

In sum, the low frequency and isolated examples of reasoning in deliberations involving inter- case conflation of the evidence, accumulation prejudice, or character prejudice suggests that the likelihood of impermissible reasoning, whether in joint or separate trials, is exceedingly low. This low probability suggests that there was negligible unfair prejudice to the defendant in joint trials or trials where tendency evidence was admitted.

Part 5.3: Legal safeguards against unfair prejudice

The law attempts to curtail the perceived unfairly prejudicial effect of joint trials and evidence of a defendant’s other misconduct via judicial instructions or directions. We examine the extent to which judicial directions and/or fact-based question trails reduce juries’ reliance on any impermissible reasoning, thereby mitigating unfair prejudice to the defendant.

Did judicial directions reduce any impermissible reasoning?

We found no differences in the ratings of the perceived criminal intent of the defendant, nor in ratings of his factual culpability for each of the counts when standard directions versus tendency directions were given in either the tendency evidence trials or joint trials. These results mirror findings from our deliberation analysis: that many juries appeared to either ignore or misunderstand the tendency direction and, consequently, failed to apply it or misapplied it. Nonetheless, juries perceived that more cognitive effort was required when they were given tendency evidence directions versus standard judicial directions in the tendency evidence trial. Overall, it appears that the tendency evidence directions were not only difficult to understand, but were also difficult to apply. The directions, based on accepted legal practice, were not written in plain English; they were comprised of dense, legal language that, anecdotally, appears to pose comprehension problems for lawyers as well. The outcome may be an effect that favours the defence, rather than the prosecution.

Do question trails reduce impermissible reasoning?

The question trail helped juries in the relationship evidence trial reach a verdict more rapidly – on average 25 minutes faster than in the absence of a question trail. From a content analysis of jury deliberations, we found a second feature of the question trails: they increased the proportion of time that juries devoted to discussing the counts of child sexual abuse.

Part 5.4: General conclusions about unfair prejudice in joint trials

Although the expectation was that more complex trials with tendency evidence would result in more unfair prejudice to the defendant, we found more evidence of impermissible reasoning in the basic separate trial and in the relationship evidence trial than in the more complex trials. For example, in the separate trials, juries were more likely to believe that there was an onus on the defendant to prove his innocence. This finding is a crucial outcome of this study. Overall, the results show that it is unlikely that a defendant will be unfairly prejudiced in the form of impermissible reasoning as a consequence of joinder of counts or the admission of tendency evidence. Given the low probability, we found there is negligible risk to the defendant of a conviction based on reasoning logically unrelated to the evidence. We recommend further empirical research on jury directions and fact--based question trails.

Conclusion

Major outcomes

This project produced two particularly significant findings: 1. There was little indication that mock juries were susceptible to any joinder effect. 2. Even if there was a joinder effect, there was no evidence that jury conviction rates were the result of impermissible propensity reasoning resulting in unfair prejudice to the defendant.

We specifically looked for instances of verdicts driven by inter-case conflation of the evidence, reasoning by accumulation prejudice and character prejudice. Across four different types of trials, no convictions were made on those bases, and very few mock juror comments reflected emotionally motivated, superficial or impressionistic considerations. Overall, jury reasoning and verdicts were logically related to the probative nature of the admitted evidence.

Implications for the criminal justice system

While some individual mock jurors made errors, and others were susceptible to attitudinal biases and decision-making prejudice, these instances were infrequent. As a group, the juries monitored and corrected individual jurors’ errors. The key to the mock juror and jury verdicts was their assessment of the credibility of the complainants, based on the source of the evidence in support of the charges.

In this study, we found that verdicts were not based on impermissible reasoning or unfair prejudice to the defendant. These outcomes suggest that any fears or perceptions that tendency evidence – whether presented in a separate trial or a joint trial – is unfairly prejudicial to the defendant are unfounded.