2. Statistics and research
Updates and background for this project (Digest)
INTRODUCTION
2.1 This chapter examines the available statistics concerning the numbers of hung juries in NSW, and the factors that could influence the likelihood of jury disagreements. It also looks at statistics in jurisdictions that have introduced majority verdicts to see what inferences may be drawn from the comparison. Finally, the Commission traverses the literature regarding studies that have been conducted into the way in which juries deliberate, with a view to determining key differences between juries deciding under a requirement of unanimity and those operating in systems where a majority verdict will suffice.
INCIDENCE AND NATURE OF HUNG JURIES IN NSW
Figures for NSW
2.2 In 1986, the NSW Law Reform Commission reported that 3.55% of 179 trials studied resulted in jury disagreement.1 That percentage had increased by 1997 when the Bureau of Crime Statistics and Research (“BOCSAR”) conducted a comprehensive post-trial survey of jurors to investigate the appropriateness of majority verdicts for NSW, and to examine the incidence of hung juries in NSW.2 This study drew on evidence from 343 trials held between 1 November 1996 and 31 May 1997, involving 853 charges. It found that, of all trials studied, approximately 10% ended with the jury being hung on at least one of the charges.3 Of this number, 33% percent involved one dissenting juror, with a further 10% involving two.4 In terms of the total number of charges, these figures indicate that juries hung with one or two dissenting voters on 2.7% of all charges on which they deliberated.5 The BOCSAR 1997 study further found that “where the jury was hung, there were about twice as many charges where the majority vote was for conviction than where it was for acquittal”.6
2.3 In 2002, BOCSAR conducted another survey into the prevalence of hung juries and aborted trials.7 This was a file-based survey8 of trials held in the District Court of NSW over a three-year period, that were either aborted, reached a verdict, or in which juries were hung. That study found that between 1998 and 2001 around 8% of juries were unable to reach a unanimous verdict.9 Of the 182 trials studied in which juries hung, 77% were hung on all charges, with 23% hung on only some of the charges.10 The average length of trials involving hung juries was 6.6 days, compared with 4.5 days for trials that reached a verdict. Of the 182 hung jury trials studied, 82% were listed for retrial, with 54% actually proceeding to retrial (some on more than one occasion). Following a hung jury, cases took, on average, an additional 7.3 months to finalise, either by retrial or another method.11 It was estimated that trials with hung juries accounted for 598 court days per year: 176 days more than the trial would have taken had it proceeded to verdict initially.12
2.4 The Commission has also received more recent, unofficial figures, directly from the Supreme Court and District Courts. Those figures indicate that, in the Supreme Court in 2003, 112 cases were listed for trial. The jury was discharged in 8 of those cases,13 with at least 3 of those 8 cases identified as having hung juries.14 In the period from January to October 2004, the Supreme Court informed the Commission that 4 matters had resulted in hung juries.15 In the District Court during 2003, 27 trials resulted in hung juries, representing 3.9% of all cases that proceeded to trial.16
Predictive factors regarding hung juries
2.5 A significant indicator of whether a jury will hang appears to be the length and complexity of a trial. The BOCSAR 1997 study found that the average duration of trials in which juries hung was 33% longer than the duration of trials that delivered a verdict on all charges, suggesting that longer trials are more likely to result in hung juries.17 This result was confirmed in the BOCSAR 2002 study, which found that the odds of hung juries in trials lasting 4-5 days, 6-10 days, and 11 days or more, were, respectively 3.4, 3.0, and 3.9 times higher than trials lasting 1-3 days. This is understandable, given that the greater number of charges, and the amount and complexity of the evidence in longer trials, increase the likelihood that a jury will disagree on the interpretation of the evidence.18
2.6 The BOCSAR 2002 study also found that one of the most significant predictors of hung juries is the location of the court registry. In trials held in a Sydney metropolitan court, the odds of trials ending in a hung jury were 3.8 times higher than trials held in a country court. The odds of juries hanging in trials in metropolitan courts outside Sydney were 2.7 times higher than in trials held in a country court.19 This may be due to the increased diversity of the juror pool in metropolitan areas, and thereby the increased likelihood of juror disagreement.20 Another reason could be that Sydney metropolitan courts are more likely to host longer, more complex trials, which have a higher chance of ending in a hung jury.
2.7 Interestingly, BOCSAR found that a number of factors had little or no influence on whether or not trials resulted in hung juries. Those factors included the type of offence with which the accused was charged, or the number of accused persons and whether or not there were multiple charges.21 This reinforces the 1997 study, which found that trials involving sexual assault charges were no more likely to result in a hung jury than trials involving other offences.22 Also having little or no impact on the likelihood of a hung jury was the bail status of the accused, whether an interpreter was required for the trial, whether a voir dire had occurred, the judge’s years of experience, the number of times a case had been listed for trial, and whether a case had been transferred from another venue.23
Comparison with other jurisdictions
Australian jurisdictions
2.8 The unanimity rule has been preserved so far as offences under Commonwealth law are concerned. In Cheatle v The Queen,24 the High Court determined that the guarantee of trial by jury in section 80 of the Commonwealth Constitution precluded “a verdict of guilty being returned in a trial upon indictment of an offence against a law of the Commonwealth otherwise than by the agreement or consensus of all the jurors”.25 Indictable Commonwealth offences are tried in State and Territory courts,26 either alone, or in conjunction with State and Territory offences.27 Consequently, the decision in Cheatle v The Queen has a direct impact on jury trials in NSW.
2.9 Like NSW, Queensland and the Australian Capital Territory have retained the unanimity rule for criminal offences.28
2.10 Majority jury verdicts are permitted in criminal trials in Victoria,29 Tasmania,30 South Australia,31 Western Australia32 and the Northern Territory.33 South Australia was the first Australian jurisdiction to adopt majority verdicts (in 1927), with Victoria being the latest to change in 1994. There are differences in the detail of the models adopted by each State and Territory. For example, Victoria permits only one juror to disagree with the majority,34 whereas the other jurisdictions provide for two dissenters.35
2.11 Most States have excluded certain crimes from the application of majority verdicts. Only the Northern Territory permits majority verdicts in all criminal trials, including murder trials. South Australia and Tasmania require verdicts to be unanimous when juries vote to convict a person of murder and treason, but a majority verdict is sufficient to acquit someone of those offences.36 Western Australia has expanded the requirement of unanimity beyond murder and treason to include offences that are punishable by “strict security life imprisonment”. In Victoria, majority verdicts are allowed except in relation to murder, treason, and drug trafficking or cultivation of narcotic plants in large commercial quantities.
2.12 Another variation is the hours of deliberation required before the jury is given a direction that they are allowed to return a majority verdict. In Tasmania, juries must deliberate for a minimum of two hours before a majority verdict may be delivered, except for murder and treason trials, where the period is six hours before a jury may decide on the basis of a majority rule to acquit the accused on a charge of murder or treason. Jurors are required to consider their verdict for at least three hours in Western Australia, at least six hours in Victoria and the Northern Territory, and at least four hours in South Australia.
New Zealand
2.13 Unanimous verdicts are currently required under the common law in New Zealand. In 2001, the New Zealand Law Reform Commission released a report into the role of juries in criminal trials.37 The Report followed more than three years of investigation and research into jury trials, including the issue of majority verdicts, and recommended that verdicts on the basis of an 11:1 majority should be introduced in all criminal trials.38 The significant number of hung juries in High Court cases, together with the rogue juror argument, appear to be the primary reasons for this recommendation.39 As a result, the Criminal Procedure Bill 2005 (NZ) contains a provision to amend the Juries Act 1981 (NZ) to introduce majority verdicts in criminal cases on an 11:1 basis, where the jury has deliberated for at least four hours.40
Canada
2.14 The unanimity requirement for juries in criminal trials has been preserved in Canada, with the Criminal Code making no provision for majority verdicts.41
England and Wales
2.15 Majority verdicts were introduced in England and Wales in 1967 for all criminal trials by the Criminal Justice Act 1967 (UK). A majority verdict of at least 10 jurors will be accepted after the jury has deliberated for at least two hours, or a longer period of time if the court thinks this is reasonable, having regard for the nature and complexity of the case.42 Juries are initially instructed by the judge that they must seek a unanimous verdict, but that “a time may come when a majority verdict will be permissible, at which point they will receive further directions”.43 Interestingly, the role of the jury in the UK has been in gradual decline over many years, with the option for trial by jury having been virtually abolished in civil cases, and seriously curtailed in many criminal cases.44
Scotland
2.16 Unlike other jurisdictions where the requirement of unanimity existed for centuries and has gradually been eroded, majority verdicts appear to have always been a feature of the Scottish legal system, becoming formally established in the mid-sixteenth century. Furthermore, the Scottish system requires only a majority of one, so that the agreement of eight jurors from a panel of fifteen is sufficient to secure either a conviction or an acquittal.45 Before retiring to consider their verdict, juries are made aware that a bare majority is enough to produce a result in the trial. Consequently, hung juries do not exist in Scotland.
2.17 The adequacy of having a bare majority, and the greater likelihood that it may result in wrongful conviction, has been examined in recent years.46 However, there has been little support for its abandonment. This is possibly due to the long history of majority verdicts in Scotland, and to the fact that the Scottish legal system has other unique features to help ensure proof of guilt, such as the requirement that the Crown case be corroborated, and the availability of an additional verdict of “not proven”.47
Incidence of hung juries in other jurisdictions
2.18 In New Zealand, figures show that in 1999-2000, the percentage of District and High Court trials ending with a jury hung on at least one charge ranged from 7.7% to 8.7%, averaging 8.27%.48
2.19 Available statistics on hung juries in other jurisdiction are sketchy and imprecise, since most States and Territories do not systematically track the number of trials with hung juries. For example, the Supreme Court of Victoria does not keep statistics on hung juries. There are also differences in the way that figures on hung juries are counted and recorded. For example, in NSW, if a jury is hung and the retrial also results in a hung jury, then this is counted as two hung juries. Whereas, in other jurisdictions, the final outcome overwrites any prior case history, and so only one hung jury would be recorded.49
2.20 However, some statistics are available. In its 2002 study, BOCSAR noted that, between 1998 and 2001, the prevalence of hung juries in Queensland was 5%, in South Australia it was 3%, while Western Australia reported that 4% of juries failed to reach a verdict. The Commission has also received information from the various court registries. The Registrar of the Supreme Court of the Australian Capital Territory reported that, in 2003-2004, 20 trials were completed, with one jury hung on all charges (representing 5%).50 In South Australia from 1 January 2004 to 2 December 2004, 74 trials were listed in the Supreme Court, with 23 finalised and no recorded hung juries.51 In Western Australia, the District Court reported 38 hung juries out of a total of 515 trials in 2003-2004 (representing 7.3%).52 In the Supreme Court of Western Australia, between 1994 and 2004, there had been 43 hung juries out of a total of 735 cases heard (representing 5.35%).53 These figures have fluctuated quite significantly over that 10 year period, with a high of 9.68% in 1996 to a low of less than 2% in 2004, with the average settling at around 6 %.
2.21 Several court registries also indicated to the Commission that statistics record only the verdicts themselves, and not the composition of the verdicts. Consequently, there are no figures kept concerning the number of trials determined by means of majority verdicts.
Inferences that can be drawn from the statistics
2.22 It would appear from the available statistics that the rate of hung juries in NSW has increased over the last 20 years from around 3.5% in 1985 to somewhere between 8-10% today. While this increase is significant, it can be explained to some extent by methodological differences between the Commission’s 1986 report and the later BOCSAR surveys. The focus of the 1986 report was not specifically targeted at the issue of hung juries, but was a more general review of the role of the jury. That report also excluded evidence from long running trials and consequently did not capture the more complex cases heard in NSW in that year. This is significant because, as the both BOCSAR studies show, the complexity and length of a case are apparently causal to the incidence of hung juries.54
2.23 Another interesting inference that could be drawn from the statistics is the apparently lower incidence of hung trials in jurisdictions with majority verdicts. While this can be attributed in part to the differences in tracking and recording figures on hung juries, other Australian jurisdictions do appear to have fewer jury disagreements than NSW. However, whether or not this can be attributed to majority verdicts alone is questionable. BOCSAR noted that if these statistics reflect a real difference, it is worth considering that South Australia and Western Australia (which have majority verdicts) have only marginally lower rates of hung juries that Queensland (where unanimity is required).55
2.24 The figures also indicate that, while the percentage rate of hung juries may be quite small, the impact on court delays due to trials with hung juries taking longer, and possibly needing to be retried, is troubling. The number and length of trials with hung juries, and the court time taken up as a result, may seem to be a compelling argument in favour of majority verdicts. However, as BOCSAR noted in its 1997 study, the figures should be treated with caution and read in conjunction with other statistics and considerations. For example, since only a minority of juries hang with one or two dissenters, the introduction of a majority decision based on 11:1 or 10:2, would affect the outcome of less than half of the 8% of trials with hung juries in NSW. When this is taken into account with the fact that not all matters proceed to retrial, and that some cases with hung juries also involve Commonwealth offences which require a unanimous verdict, BOCSAR estimates that the introduction of majority verdicts would result in a potential net saving in criminal court time of only 1.7% for 10:2 verdicts, and 1.1% for 11:1 verdicts.56
STUDIES INTO JURY DELIBERATIONS
2.25 The way in which juries reach decisions, or, in some cases, fail to arrive at a verdict, has been of interest to researchers for several decades. While jurors are instructed not to form any clear view until they retire to consider their verdict, research has suggested that most jurors begin deliberations with an opinion as to the accused’s guilt or innocence, formed on the basis of listening to the evidence, the judge’s instructions, and the juror’s own knowledge and life experience.57 The manner in which group deliberations are conducted has a significant impact on whether that initial view is confirmed or changes. A recent study in New Zealand revealed the importance of jury deliberations, reporting that 22% of jurors changed their mind from their initial view during deliberations, with a further 20.5% forming a view after being undecided at first.58
2.26 A significant amount of research has been conducted over many years, mostly in the United States of America, examining the impact on jury deliberations of requiring unanimous decisions as opposed to majority verdicts. These studies have adopted four primary methodologies. First, mock jury experiments have been conducted involving simulated trials; secondly, researchers have conducted interviews with ex-jurors; thirdly, file-based analyses of jury verdicts have been obtained from court records; and fourthly, field studies or experiments involving real juries.
2.27 While each of these methodologies has inherent drawbacks,59 the following paragraphs represent fairly consistent findings from a number of studies over the last four decades.
Early majority view tends to prevail
2.28 The first systematic research into the way in which juries deliberate began in the United States in the 1950s with the Chicago Jury Project. As part of that project, Harry Kalven and Hans Zeisel conducted post-trial interviews with judges, attorneys and jurors. The study looked at, among other things, the verdict preference of jurors based on the first ballot cast during deliberations, as compared with the final verdict in each case. Their report, entitled The American Jury,60 published in 1966, found that the verdict preferred by the majority of jurors on the first ballot was the jury’s final verdict in over 90% of cases. This has become one of the most “robust and widely replicated findings in jury research”,61 both in studies with real juries and in mock jury studies.
2.29 Mock trial research has indicated that the size of the majority at the first ballot has an influence in whether juries convict, acquit or hang.62 The studies suggest that, where there is a strong majority of two-thirds or more, that view will usually constitute the final verdict. Where there is a weak majority, or juries are evenly divided, the final result will usually be an acquittal or a hung jury.63 Research has also supported the existence of an asymmetrical leniency effect towards acquittal, so that if “7 or fewer jurors favour conviction at the beginning of the deliberation, the jury will probably acquit, and if 10 or more jurors believe the defendant is guilty, the jury will probably convict. With 8 or 9 jurors initially favouring conviction, the final verdict is basically a toss-up”.64
2.30 While studies with real juries have also noted the tendency for the early majority vote to prevail, a “small but not insignificant” number of “Twelve Angry Men” cases has also been observed. A recent United States study has reported that, in 89 cases where there was a strong initial majority favouring conviction, the jury ultimately voted to acquit in 11 (or about 12%) of those cases. The reverse scenario occurs less often, with 3 out of 71 juries voting to convict following an overwhelming initial majority supporting acquittal.65 A NSW study reported one case in which a minority of jurors persuaded the majority to accept a lesser verdict. When interviewed, the judge considered the lesser verdict to be preferable to the verdict initially favoured by the majority.66
Deliberation styles
2.31 Deliberation style refers to way in which juries reach their decisions. The bulk of research on jury decision-making has suggested that there are two basic forms of deliberation. The first is evidence-driven deliberation, where jurors identify and discuss the evidence and issues in the case at significant length before any vote is taken. The second deliberation style is known as poll-driven or verdict-driven, since jurors take an initial vote to see where they stand, and then work on eliminating the difference of opinion among them.67 Other studies have identified a mixed style, where jurors briefly discuss the evidence or the relevant law, and then quickly take a vote.68 Given the tendency for the early majority vote to determine the verdict, the type of deliberation style adopted by a jury may have a significant effect on the eventual outcome. It has generally been assumed that juries required to be unanimous favour the evidence-driven approach, while juries operating under a majority decision rule more commonly use the poll-driven approach.69
2.32 As pointed out in the New Zealand study, it has sometimes been said that evidence-driven deliberations promote more effective decision-making, as they are less divisive, keep jurors working together, and promote thoughtful discussion.70 The New Zealand study found that some juries that took an initial poll floundered in their task as their deliberations were unstructured, with the focus being on the difference of opinions amongst jurors, and how to persuade the minority to change their minds. However, other “poll-driven” juries methodically applied the facts to the law and functioned efficiently. Interestingly, some juries that did not take an initial poll were “disorganised, inefficient and essentially lacking in focus or direction”.71 The study concluded that the “most important factor in determining the effectiveness of jury decision-making … was not how they started but the extent to which they adopted a systematic structure for assessing the evidence and applying the law”.72 Methods used by juries to structure their deliberations included flow charts and diagrams of the evidence and the law constructed from the judge’s summing up, a written list of questions supplied by the judge at the request of the jury, and a summary of the evidence compiled by a juror prior to the commencement of the deliberations.73 A skilful foreperson was also found to be crucial in structuring productive deliberations.74
2.33 Whether or not juries adopt a “poll-driven” approach to deliberating, all juries must eventually take a vote on whether to convict or acquit the accused on each charge. Studies which examined the way opinions are expressed in the jury room, found that individual preference change during deliberation is influenced by polling regularity, polling format (public versus secret), poll timing (early versus late) and the prior sequence of votes.75 However, the actual effect of these influencing factors is often unclear. For example, one study noted that juries polled at regular intervals were somewhat less likely to hang than juries that were not, whereas another found that mandated polling at regular intervals produced longer deliberations and more hung juries.76
Unanimous decisions versus majority rulings
2.34 Studies have reported various findings concerning the effect of requiring consensus as opposed to majority rule on jury decision-making. Some studies found that there was little or no difference in decision-making between groups required to decide under unanimity and those deciding under a majority rule.77 However, it has been pointed out that these studies have methodological weaknesses, such as severely curtailed deliberation times and small samples.78 In most studies, it has been found that juries operating under a majority verdict system generally tend to reach a verdict in less time, as they can stop deliberating when the requisite number of jurors reach agreement.79 Also, juries who are not required to reach unanimity were found to take fewer polls,80 debate the evidence more quickly and less thoroughly,81 recall less evidence,82 and hang less often.83 Consistent with these findings are studies reporting that decisions made under the unanimity rule took longer, involved more “robust” argument, and required more rounds of voting.84
2.35 On the other hand, the unanimity rule gives every group member the power of veto, and each member’s opinion must be taken into account in the decision, enabling minority group members to participate more in discussions.85 Research has found that juries under the unanimity rule conduct a more thorough assessment of the evidence and the law, with jurors more likely to participate equally and be more satisfied with the outcome.86 In unanimous verdict jurisdictions, research has shown that discussion and debate over the evidence continues even after a substantial majority of 10 or so has been reached. It is often at this time that error correction occurs and references are made to the standard of proof, or questions are put to the judge.87
2.36 It has also been found that marked differences between jury deliberations requiring unanimity and those made under a majority verdict scheme are likely to be contingent on factors other than the type of verdict required, such as the strength of the evidence. So, for example, where the prosecution case is not very strong, or the evidence lends itself to a number of interpretations, deliberations are likely to be lengthier and more difficult where unanimity is required than under a majority verdict scheme.88
Little difference as to verdict
2.37 Many researchers in the 1970s and early 1980s conducted studies to examine the effects of group decision rules on jury deliberations. These studies repeatedly demonstrated that requiring a unanimous decision as opposed to a majority ruling has little effect on the direction of final verdicts.89 That is, regardless of whether unanimity or some variant of the majority rule is prescribed, juries tend to return the verdict preferred by a majority of jurors at the beginning of the deliberation. However, it has been pointed out that the failure to find any significant differences in verdict choices could be explained by the fact that most jury studies involve a choice between two alternative decisions (that is, guilty or not guilty). Differences in verdicts have been recorded in studies where a wider range of alternative choices along a continuum have presented.90
2.38 Furthermore, post-trial interviews with jurors in systems requiring unanimous verdicts found that juries did not always decide between guilty or not guilty on each charge in isolation. The studies found that jurors sometimes compromised on verdicts. For example, a jury may agree to a conviction of manslaughter rather than convicting or acquitting for murder, or may “trade off” acquittals on some charges for convictions on others.91
Juror satisfaction
2.39 The New Zealand study involving post-trial interviews with jurors in 48 trials revealed that many jurors felt pressure to reach uniformity.92 That pressure was due mainly to the juror’s own feelings of obligation to reach a verdict and not cause a hung jury; the directions given by the judge to juries who may be having difficulty reaching an agreement; pressure from other jurors to accord with the majority view; and other factors such as time constraints, poor facilities and late sittings. In some cases, the pressure from other jurors involved deliberately presenting the hold out juror with false information about the evidence, or refusing to allow that juror to take a break in order to coerce him or her into agreement.93
2.40 However, jurors have also reported that unanimity leads to a fuller exploration of relevant arguments and to more thorough deliberation and participation by all members of the panel.94 This results in jurors feeling that their opinions had been listened to, and led to more satisfaction and confidence in the final verdict.95 Some studies also found the converse to be true regarding juries deliberating under a majority verdict rule, with jurors less likely to feel that justice had been administered.96
Research into hung jury deliberations
2.41 The classic study conducted by Kalven and Zeisel in 1966 contained findings relating to hung juries. Although the sample was small and the methodology may be somewhat questionable today, those findings set the benchmark for decades regarding assumptions about hung juries, and have been relied on in subsequent studies. Kalven and Zeisel found that approximately 5.6% of juries in the United States required to reach a unanimous verdict were unable to do so, while those juries allowed to reach a majority verdict hung at a rate of 3.1%.97
2.42 The reasons for hung juries, they concluded, were mainly evidence factors, which accounted for 71% of hung juries. Juries were most likely to hang in cases where the evidence was evenly balanced, in that it did not strongly favour the accused or the defendant.98 In 63% of hung juries, the majority favoured conviction, compared with 24% where the majority favoured acquittal.99 Less than half (42%) of all juries failing to reach a verdict hung because of one or two hold out jurors. Finally, Kalven and Zeisel found that most hung juries were characterised by an evenly split vote on the initial ballot, involving significant disagreement amongst jurors. Consequently, even in cases where the jury eventually hung on the basis of only one or two votes, the views of the hold out jurors had initial support from a number of other jurors, signifying that debate about the evidence, rather than the recalcitrance of a single juror, accounted for the failure to decide in most hung juries.100
2.43 These findings of nearly 40 years ago are remarkably similar to the results of recent studies. A four-year study of hung juries in the United States,101 completed in 2002, found that the average hung jury rate across 30 large urban courts was 6.2%, slightly higher than in 1966.102 That study also found evidentiary factors to be major contributors to whether or not a jury hung, with complex, weak and/or ambiguous evidence, and a poor performance presenting the evidence by either the prosecutor or defence attorney, more likely to result in a deadlocked jury. Also relevant was the strength and credibility of police evidence and the testimony of the accused, with hung juries having greater differences of opinion as to credibility.103 Juror’s views concerning the fairness of the law were also found to have an impact on the outcome of a trial (playing some role in 27% of deadlocked cases), with some jurors deciding to hang the jury as they believed the “legally correct” outcome to be unfair.104 Hung juries also reported more trouble recalling the evidence, understanding and interpreting the judge’s instructions, and a greater degree of interpersonal conflict amongst the jury panel.105 However, this study revealed that, in general, juror characteristics and demographics, such as age, ethnicity, race or sex, were not significant predictors of whether or not a jury would hang.106
2.44 Like the 1966 study, the 2002 study found that juries with a fairly even split, or with only a slight majority on the first ballot, tended to hang.107 Interestingly, the 2002 study also found that 42% of hung juries were on the basis of one or two votes.108 This also mirrors the findings of the BOCSAR 1997 study mentioned above.109
2.45 In the 2001 New Zealand study, 5 out of the 48 trials studies ended with a hung jury. In three of those trials, the minority jurors provided an articulated and reasoned basis for their dissent. In fact, regarding one of those three trials, the researchers were of the opinion that the minority view was the correct one, and had the trial not hung, the verdict of the majority would have been “questionable, if not…perverse”. In the second case, the evidence was finely balanced, and a post-trial interview revealed that the judge agreed with the minority view.110 The lack of a verdict in the third case stemmed from a misunderstanding of the evidence, which the researchers suggest could have been rectified if the evidence had been presented more clearly.111
2.46 The New Zealand study found that the remaining two hung juries were the result of a single juror who steadfastly refused to consider a guilty verdict, explain the reasons for adhering to the contrary view, or take part in deliberations.112 However, the study also noted that it could not be determined conclusively that a guilty verdict would have been recorded had a majority verdict system been operating, since other jurors in those trials had misgivings about the evidence which were not resolved or addressed, since it was clear that the jury was going to hang anyway.113
The Commission’s view
Conclusions from research
2.47 It would appear from the research to date that juries required to make unanimous decisions consider the evidence more carefully and thoroughly, and report higher levels of juror confidence in the ultimate decision, than juries operating a majority verdict system. Where a verdict must be unanimous, the views of each juror must be considered, allowing those in the minority to be included in the decision making process, and encouraging groups to expend more effort on problem solving. However, it would also appear that juries required to reach consensus take longer to deliberate and hang more often than juries requiring only a majority of votes. Jurors also reported feeling pressure to achieve uniformity, not only stemming from other jurors, but also from their own desire not to cause the jury to hang.
2.48 Studies have apparently shown the majority decision rule to be less time consuming, less subject to simple compromise, with jurors able to be more “honest” in their views. It would also appear that juries not required to attain unanimity also hang less often. On the other hand, decisions under majority rule often exclude the desires of some group participants from the final decision, and the use of majority rule is more likely to reduce the cohesiveness or unity of a group. Interestingly, studies do not appear to show a significant difference in verdict preferences between juries deciding on the basis of unanimous, as opposed to majority, verdicts.
2.49 Studies also show the importance of how jury deliberations are structured, and the crucial nature of the first ballot in determining the eventual verdict, with juries more likely to hang where the discussions are disorganised and the initial poll is evenly split. They also highlight the significance of evidentiary factors, with hung juries more likely in trials involving complex and ambiguous evidence that does not strongly favour one side over the other. It has also been consistently found that less than half of all juries that hang do so on the basis of one or two votes, and, even in those cases, there was initial support from other jurors for the minority view.
Need for caution
2.50 The results of the various studies described above are interesting in terms of the light they shed on jury deliberations, particularly given the sacrosanct nature of those deliberations and the difficulty of knowing exactly what occurs beyond the jury room door. However, caution should be adopted when drawing conclusions from the studies, especially concerning the benefits of one verdict system over another.
2.51 Many of the findings need to be considered in light of the methodology used in the studies. A number of studies, particularly those involving mock juries, used groups much smaller than the twelve that generally constitute a criminal jury, and placed stringent limits on deliberation times. Mock jury studies have the benefit of being able to control and manipulate evidence in order to isolate and identify behavioural patterns and test juror reaction to specific stimuli. However, this works against them replicating the environment of real jury deliberations. Further, mock jury studies cannot mirror courtroom conditions, such as the judge’s instruction, the demeanour of the witnesses, and way in which prosecutors and defence counsel present the evidence.
2.52 Also, many of the mock juries were directed to reach a verdict of either guilty or not guilty: that is, participants were not given the option to hang the jury. It is also difficult to simulate long trials, which is crucial to obtaining information about hung juries since it is lengthy trials that are less likely to result in a verdict.114 Perhaps the biggest barrier to the general applicability of mock jury studies is that it is “impossible to recreate the responsibility of deciding the fate of one’s fellow man”.115 The real life consequences are lacking in simulated jury experiments, and jurors who acquiesce with the majority in a mock trial may hang a jury when confronted with the prospect of a wrongful conviction or acquittal.116 As a result, “although mock jury research can provide some valuable insights, it is limited in what it can explain about the actual frequency of hung juries or their underlying causes”.117
2.53 Interviews with ex jurors and file studies are useful techniques as they involve actual juries, but also have their limitations. For example, while interviews obtain the feelings and perceptions of actual jurors, they rely on ex post facto recall of events in trials that may have lasted weeks. Further, they can only occur with the consent and cooperation of jurors who choose to participate, and so risk obtaining a skewed picture of events. File studies are limited by the types of records kept by courts, which vary between jurisdictions, and often lack the sort of information that would help to determine why and how juries hang.118
2.54 These methodological deficiencies do not negate the worth of the studies into jury deliberations, and weight is certainly added to the findings when they occur repeatedly across different studies using various methods. The studies are useful indicators of when and why juries may hang, and the merits of unanimity over majority verdicts, and vice versa. However, the Commission is of the view that they are indicative only, and do not offer definitive conclusions.
FOOTNOTES
1. Report 48 at para 9.19. The study was of trials between 30 September 1985 to 13 December 1985.
2. P Salmelainen, R Bonney and D Weatherburn, “Hung juries and majority verdicts” (1997) 36 Crime and Justice Bulletin 1(hereafter “BOCSAR 1997 study”).
3. BOCSAR 1997 study at 2.
4. BOCSAR 1997 study at 2-3.
5. BOCSAR 1997 study at 3.
6. BOCSAR 1997 study at 3.
7. J Baker, A Allen, and D Weatherburn, “Hung Juries and Aborted Trials: An analysis of their prevalence, predictors and effects” (2002) 66 Crime and Justice Bulletin 1 (hereafter “BOCSAR 2002 study”).
8. In contrast to the 1997 study which involved post-trial interviews with jurors.
9. BOCSAR 2002 study at 5.
10. BOCSAR 2002 study at 6.
11. BOCSAR 2002 study at 6.
12. BOCSAR 2002 study at 7.
13. Which could be for a number of reasons, including failure to agree.
14. Information supplied in an email from the Supreme Court to the Commission dated 5 November 2004. Assuming that the actual number of hung juries is somewhere between 3 and 8, this puts the percentage rate at between 2.6% and 7.1% of trials listed.
15. Information supplied in an email from the Supreme Court to the Commission dated 20 October 2004.
16. Information supplied to the Commission from the District Court in October 2004.
17. BOCSAR 1997 study at 2. Trials involving hung juries took an average of 7.3 days to complete, whereas other trials delivered a verdict in 5.5 days on average.
18. BOCSAR 2002 study at 8.
19. BOCSAR 2002 study at 8.
20. This finding is supported by research in the United States: see V Hans, P Hannaford-Agor, N Mott and T Munsterman, “The Hung Jury: The American Jury’s Insights and Contemporary Understanding” (2000) Criminal Law Bulletin 1 at 9.
21. BOCSAR 2002 study at 7-8.
22. BOCSAR 1997 study at 2. However, trials involving fraud, sex, or other violent offences, were more likely than other trials to be aborted: BOCSAR 2002 study at 8-9.
23. BOCSAR 2002 study at 7-8.
24. (1993) 177 CLR 541.
25. (1993) 177 CLR 541 at 562.
26. Indictable Commonwealth offences are those punishable by imprisonment for more than 12 months, unless the contrary intention appears: Crimes Act 1914 (Cth) s 4G. While some indictable offences may be punished summarily, offences with a penalty of more than 10 years imprisonment must be punished on indictment: Crimes Act 1914 (Cth) s 4J.
27. See Judiciary Act 1903 (Cth) s 68(2). See also para 1.33.
28. See Jury Act 1995 (Qld) s 59. The common law requirement of unanimity has not been abrogated by the Juries Act 1967 (ACT).
29. Juries Act 2000 (Vic) s 46.
30. Juries Act 1899 (Tas) s 48.
31. Juries Act 1927 (SA) s 57.
32. Juries Act 1957 (WA) s 41.
33. Criminal Code Act (NT) s 368.
34. Where the jury consists of 12, 11 must agree. Where the number of jurors is reduced, 10 out of 11, or 9 out of 10 must agree: Juries Act 2000 (Vic) s 46.
35. In South Australia, at least 10 must agree where the jury panel consists of 12 jurors. If the jury has been reduced, 10 out of 11, or 9 out of 10 must agree: Juries Act 1927 (SA) s 57. In the Northern Territory, where the jury panel consists of 11 or 12, at least 10 must agree, but if the panel is reduced to 10, then 9 must agree: Criminal Code (NT) s 368.
36. Juries Act 1927 (SA) s 57. In Tasmania, unanimity is also required in relation to a conviction for a crime punishable by death, or any special finding on which the accused would be convicted of such a crime: Jury Act 1899 (Tas) s 48(3).
37. New Zealand Law Commission, Juries in Criminal Trials (Report 69, 2001) (hereafter “NZLC Report 69”).
38. NZLC Report 69, Recommendation after para 441.
39. NZLC Report 69 at para 419 and 435.
40. Criminal Procedure Bill 2005 (NZ) s 82A. That Bill also provides for majority verdicts in civil cases in circumstances where the jury has deliberated for four hours, and the verdict has been agreed upon by at least three-quarters of the jury panel
41. N Vidmar, “The Canadian Criminal Jury: Searching for a Middle Ground” in N Vidmar (ed) World Jury Systems (Oxford University Press, New York, 2000) at 219.
42. The agreement of 10 jurors is sufficient where the panel consists of 11 or 12. Where the panel has been reduced to 10, 9 jurors must agree: Juries Act 1974 (UK) s 17(1).
43. S Lloyd-Bostock and C Thomas, “The Continuing Decline of the English Jury” in Vidmar (ed) at 86.
44. S Lloyd-Bostock and C Thomas, “Decline of the ‘Little Parliament’: Juries and Jury Reform in England and Wales” (1999) 62 Law and Contemporary Problems 7 at 39-40.
45. P Duff, “The Scottish Criminal Jury: a very peculiar institution” in Vidmar at 270.
46. P Duff in Vidmar at 270-271.
47. Juries in Scotland have three verdict options: guilty, not guilty or not proven. Verdicts of not guilty and not proven result in acquittal. The not proven verdict implies that a jury is not convinced of the innocence of the accused, but is equally unconvinced that the evidence in the Crown case is strong enough to justify a guilty verdict. It is estimated that about one-third of all jury acquittals are the product of the not proven verdict: see P Duff in Vidmar at 272-277.
48. NZLC Report 69 at para 4.17-4.18.
49. See BOCSAR 2002 study at 6 and note 8 at 12.
50. This compares with previous years: in 2002-2003, two trials out of total of 33 involved a jury hung on some charges; and in 2001-2002, juries in 8 out of a total of 33 trials hung on all charges: letter from the Registrar, Supreme Court of the Australian Capital Territory to the Executive Director of the Law Reform Commission dated 13 December 2004.
51. Letter from the Deputy Registrar, South Australian Criminal Courts Registry, to the Executive Director of the Law Reform Commission dated 20 December 2004.
52. Letter from the Principal Registrar, District Court of Western Australia, to the Executive Director of the Law Reform Commission dated 24 December 2004. The Principal Registrar noted that in some, but not all trials in which there was a hung jury, the judge would have given a direction that the jury could determine the matter by majority verdict.
53. Letter from the Business Services Project Manager, Supreme Court of Western Australia, to the Executive Director of the Law Reform Commission dated 15 December 2004.
54. See para 2.5.
55. BOCSAR 2002 study at 6.
56. BOCSAR 1997 study at 4.
57. See P Darbyshire, A Maughan and A Stewart, “What Can the English Legal System Learn From Jury Research Published up to 2001?” (2001) (http://www.kingston.ac.uk/~ku00596/elsres01.pdf) at 22 and 29.
58. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.53.
59. The Commission discusses the strengths and weaknesses of each approach at para 2.51-2.54.
60. H Kalven and H Zeisel, The American Jury (Chicago University Press, 1966).
61. See DJ Devine, LD Clayton, BB Dunford, R Seying and J Pryce, “Jury Decision Making” (2001) 7(3) Psychology, Public Policy and Law 622 (hereafter referred to as “Devine et al”) at 623.
62. Studies conducted by J H Davis (1973) and R J MacCoun and N L Kerr (1988) that reported findings on the majority effect are discussed in Devine et al at 690-691.
63. See Devine et al at 690. See also V P Hans, P L Hannaford-Agor, N L Mott and G T Munsterman, “The Hung Jury The American Jury’s Insights and Contemporary Understanding” (2003) Criminal Law Bulletin at 14 and 16.
64. See MacCoun and L Kerr (1988) reported in Devine et al at 691-692.
65. See also V P Hans, P L Hannaford-Agor, N L Mott and G T Munsterman (2003) at 15.
66. M Chesterman, J Chan, S Hampton, Managing Prejudicial Publicity (February 2001, Justice Research Centre, Law and Justice Foundation of NSW) at para 414.
67. See Devine et al at 693. See also Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.1, and Y Tinsley, “Juror Decision-Making: A Look Inside the Jury Room”, British Criminology Conference 2000: Selected Proceedings (Volume 4, 2001) (www.britsoccrim.org/bccsp/vol04/tinsley1.htm).
68. See Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.2; and R Hastie, S D Penrod and N Pennington, Inside the Jury (Harvard University Press, Cambridge Massachusetts, 1983) reported in Devine et al at 692.
69. P L Hannaford-Agor, V P Hans, N L Mott and G T Munsterman, Are Hung Juries a Problem? (National Centre for State Courts, 2002) at 14.
70. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.1.
71. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.6.
72. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.7.
73. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.7.
74. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.24.
75. See Devine et al at 694.
76. See Devine et al at 694.
77. See T Kameda, “Procedural influence in small group decision-making: Deliberations style and assigned decision rule (1991) 61 Journal of Personality and Social Psychology 245-256; and P D P Velasco, “The influence of style and decision rule in jury decision-making” in G Davies, S Lloyd-Bostock, M McMurran and C Wilson (eds), Psychology, law and criminal justice: International developments in research and practice (Berlin, Germany, 1995) at 344-348.
78. See Devine et al at 669.
79. See studies by J H Davis, N L Kerr, R S Atkin, R Holt, D Meek, “The decision processes of 6 and 12 person mock juries assigned unanimous and two-thirds majority rules” (1975) 57 Journal of Personality and Social Psychology 1-14; R D Foss, “Structural effects in simulated jury decision making” (1981) 40 Journal of Personality and Social Psychology 1055-1062; R Hastie, S D Penrod, and N Pennington, Inside the Jury (Harvard University Press, Cambridge Massachusetts, 1983); and C Nemeth, “Interactions between jurors as a function of majority versus unanimity decision rules” (1977) 7 Journal of Applied Social Psychology 38-56 at 55; R J MacCoun, “Experimental Research on Jury Decision-Making” (June 1989) 244 Science 1046-1050: all cited in Devine et al at 669.
80. J H Davis, N L Kerr, R S Atkin, R Holt, D Meek, “The decision processes of 6 and 12 person mock juries assigned unanimous and two-thirds majority rules” (1975) 57 Journal of Personality and Social Psychology 1-14; N L Kerr, R S Atkin, G Stasser, D Meek, R W Holt and J H Davis, “Guilt beyond a reasonable doubt: Effects of concept definition and assigned decision rule on the judgments of mock jurors” (1976) 34 Journal of Personality and Social Psychology 282-294: cited in Devine et al at 669.
81. See Hans, Hannaford-Agor, Mott and Munsterman, (2003) at 4-5; R J MacCoun, (June 1989).
82. MacCoun, (June 1989).
83. N L Kerr, R S Atkin, G Stasser, D Meek, R W Holt and J H Davis, “Guilt beyond a reasonable doubt: Effects of concept definition and assigned decision rule on the judgments of mock jurors” (1976) 34 Journal of Personality and Social Psychology 282-294; C Nemeth, “Interactions between jurors as a function of majority versus unanimity decision rules” (1977) 7 Journal of Applied Social Psychology 38-56; and M J Saks, “What do jury experiments tell us about how juries (should) make decisions? (1997) 6 Southern California Interdisciplinary Law Journal 1-53: all cited in Devine et al at 669.
84. C E Miller, Group Decision Making Under Majority and Unanimous Decision Rules (1985) 48(1) Social Psychology Quarterly 51-61 at 59; M F Kaplan and C E Miller, “Group Decision Making and Normative Versus Informational Influence: Effects of Type of Issue and Assigned Decision Rule” (1987) 53(2) Journal of Personality and Social Psychology 306-313 at 307; C Nemeth, “Interactions between jurors as a function of majority versus unanimity decision rules” (1977) 7 Journal of Applied Social Psychology 38-56 at 55.
85. Nemeth (1977) at 53 and 55.
86. R Hastie, S D Penrod, and N Pennington (1983); Nemeth (1977) at 55.
87. Darbyshire, Maughan and Stewart (2001) at 30.
88. See Devine et al at 669.
89. See, eg, B Grofman, “Not necessarily twelve and not necessarily unanimous: Evaluating the impact of Williams v Florida and Johnson v Louisiana” in G Bermant, C Nemeth, and N Vidmar (eds) Psychology and the Law (Lexington, MA, 1976); Nemeth (1977) at 55; and MacCoun (June 1989).
90. C E Miller, “Group Decision Making Under Majority and Unanimous Decision Rules” (1985) 48(1) Social Psychology Quarterly 51-61 at 51 and 58.
91. Chesterman, Chan, and Hampton at para 408-416. Compromise verdicts are discussed further at para 3.23-3.26.
92. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 8.2-8.15.
93. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 6.47
94. Nemeth (1977) at 55; Hastie, Penrod, and Pennington (1983); L L Thompson, E A Mannix, M H Bazerman, “Group Negotiation: Effects of Decision Rule, Agenda and Aspiration (1988) 54(1) Journal of Personality and Psychology 86-95 at 92; Hannaford-Agor, Hans, Mott and Munsterman (2002) at 14; Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 9.16.
95. M J Saks, “What do jury experiments tell us about how juries (should) make decisions? (1997) 6 Southern California Interdisciplinary Law Journal 1-53; Nemeth (1977) at 55; Hannaford-Agor, Hans, Mott and Munsterman (2002) at 14.
96. Nemeth (1977) at 53; M F Kaplan and C E Miller, “Group Decision Making and Normative Versus Informational Influence: Effects of Type of Issue and Assigned Decision Rule” (1987) 53(2) Journal of Personality and Social Psychology 306-313 at 312.
97. Kalven and Zeisel (1966) at 461.
98. Kalven and Zeisel (1966) at 456.
99. Kalven and Zeisel (1966) at 460.
100. Kalven and Zeisel (1966) at 463.
101. This research consisted of a file-based study of hung jury rates in state and federal courts, together with an in-depth study involving interviews with judges, jurors and attorneys from courts across four jurisdictions. Due to difficulty securing co-operation from some jurisdictions, the study covers only jurisdictions which require unanimous verdicts: see Hannaford-Agor, Hans, Mott and Munsterman (2002) at 2-3.
102. This is an averaged rate for those 30 state courts, with individual rates varying significantly: Hannaford-Agor, Hans, Mott and Munsterman (2002) at 83.
103. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 75-78 and 81.
104. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 78 and 81.
105. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 78 and 81.
106. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 57.
107. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 66.
108. Hannaford-Agor, Hans, Mott and Munsterman (2002) at 67.
109. See para 2.2.
110. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 9.13 and 9.16. Both of these cases would have resulted in a verdict under a majority verdicts system of 11:1 or 10:2.
111. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 9.13.
112. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 9.12.
113. Young, Cameron and Tinsley, Preliminary Paper 37(2) at para 9.12.
114. See Chesterman, Chan and Hampton at para 82.
115. See Darbyshire, Maughan and Stewart (2001) at 21.
116. Hans, Hannaford-Agor, Mott and Munsterman (2003) at 8.
117. Hans, Hannaford-Agor, Mott and Munsterman (2003) at 8.
118. Devine et al at 626-627.
