Scottish jury research: findings from a mock jury study

3. What impact do the unique features of the Scottish jury system have on verdict choice?

Key findings

• There were a number of statistically significant differences at juror level.
  • There was an overall tendency for jurors to change from favouring guilty to favouring acquittal over the course of deliberations. However, most juries (51 of 64) in this study started their deliberations with a majority of jurors favouring acquittal. It is therefore unclear whether this reflects a tendency (noted in other research) for jurors to become more lenient as a result of deliberating, or a tendency for jurors to move towards the majority position (whatever that might be).
  • When the not proven verdict was available, more individual jurors favoured acquittal, both before and after deliberating.
  • Jurors in 15-person juries were less likely to change their minds about the verdict than jurors in 12-person juries.
  • Jurors asked to reach a simple majority were more likely to think the verdict should be guilty (after deliberating) than were those asked to reach a unanimous verdict.
• As expected, due to the small number of juries in each condition, there were no significant differences at jury level in the proportion of guilty verdicts returned between juries that had the not proven verdict available and juries that did not, between 12 and 15-person juries, or between juries asked to reach a unanimous verdict and those asked to reach a simple majority.
• Separate analysis of jurors in the rape and assault trials shows no significant difference in the proportion of rape trial jurors favouring guilty or acquittal verdicts by either the number of verdicts available, or by jury size. The finding that requiring unanimity increases the proportion of jurors favouring acquittal holds across both types of trial, however.
• The combination of features most strongly associated with individual jurors favouring a guilty verdict at the end of deliberations was 15-person, simple majority, and two verdicts. The combination in which jurors were least likely to favour guilty by the end of deliberations was 12-person, unanimity, and three verdicts. Neither of these jury system combinations are currently used in Scotland or in England and Wales, although the first would represent the system in Scotland if the not proven verdict were abolished without any other reforms.

3.1 Introduction

A central aim of this research was to understand what effect, if any, the three unique features of the Scottish jury system - three verdicts, 15 jurors, and the requirement that juries reach a simple majority (eight out of 15) before returning a verdict - have on jury decision-making. This chapter discusses findings on the effects of these unique features on verdict choice, at both jury level (verdicts returned by the 64 juries) and individual juror level (individual jurors' views on what the verdict should be, which may differ from the verdict returned by their jury collectively).

3.2 Jury versus juror-level findings

As discussed in Chapter 1, this study is the largest mock jury study conducted in the UK to date. However, the total number of juries (64) is still relatively small in statistical terms. This means that, at the jury level, anything other than large differences in the likelihood of a guilty verdict being returned would not be picked up (as differences would need to be large to be statistically significant).

However, we can also examine the individual verdict preferences of all 863 jurors who took part in the jury deliberations, based on their responses to questionnaires completed before and after deliberation. This juror-level analysis has the advantage of a considerably larger sample size, meaning there is a greater likelihood of identifying significant differences by the number of verdicts available, jury size, and majority required.

One disadvantage of this approach is that in real trials jurors do not return verdicts individually - they do so collectively, after deliberation. At the same time, juries' verdicts are the product of individual jurors' views, mediated through the deliberation process and the rules which the jury must apply to reach their collective verdict. In this sample, 30 out of 32 simple majority juries (94%) returned verdicts which were the same as the majority preferences of individual jurors (as indicated in their pre-deliberation questionnaires). Similarly, of the unanimity juries who were able to return a verdict rather than hanging (25 of 32 juries), the choice to convict or acquit could have been accurately predicted from the balance of individual jurors' initial verdict preferences in almost all cases (though not necessarily whether not guilty or not proven would be used as the means of acquittal).^[46] In other words, the majority of jury-level verdicts could have been predicted from the pre-deliberation responses of individual jurors.^[47] Examining the relationship between individual verdict preferences and the different features of the jury system in this study can therefore shed light on how these features might tilt jurors - and by extension, juries - in one direction or another.

However, it is important to keep in mind that the precise effects that any impact on juror-level views will have on jury-level outcomes in any given case cannot be reliably predicted - it will depend on other factors, including the balance of evidence, the initial balance of opinion of the jury, and individual jury dynamics. It is not, therefore, possible to estimate the likely scale of any impact that differences at juror-level would have on jury-level outcomes across a larger number of differently balanced trials. As explained in section 2.3, the trials used in this research were designed to be finely balanced in order to encourage discussion of the not proven verdict. In practice, not all trials will be so finely balanced.

3.3 Does jury size, the number of verdicts available, or the majority required make any difference to jury-level verdicts?

Table 3.1 shows the verdicts returned by all 64 mock juries, broken down by trial type and by each system feature being examined (number of verdicts, jury size, and majority required).^[48] Overall, seven juries returned a guilty verdict, 26 returned not guilty, 24 returned not proven, and seven failed to reach a verdict within the 90 minutes allowed for deliberation and were thus recorded as hung.

All seven hung juries had been asked to reach a unanimous verdict, on which all (or almost all)^[49] jurors were agreed.^[50] The relatively high number of hung juries in this study is likely to reflect both the deliberately finely balanced nature of the trial videos (reasons for which are discussed in section 2.3), and the fact that deliberations were time-limited (to a maximum of 90 minutes). The findings do not, therefore, indicate that juries asked to reach unanimity would be expected to hang with this frequency in reality. Indeed, in England and Wales, where juries are required to try to reach unanimity,^[51] the proportion of juries unable to return a verdict is typically under 1%.^[52]

The pattern of verdicts reached in the two different trials shown to mock jurors was very similar. Across rape trial juries, there were four guilty verdicts, 12 not guilty verdicts, 12 not proven verdicts, and four hung juries. Across assault trial juries, there were three guilty verdicts, 14 not guilty verdicts, 12 not proven verdicts and three hung juries.

Table 3.1 - Verdicts returned by jury features

	Guilty	Not guilty	Not proven	Hung	Number of juries
All 64 juries	7	26	24	7	64
Trial type
Rape trial (All)	4	12	12	4	32
Rape trial, two verdicts	1	12	NA	3	16
Rape trial, three verdicts	3	0	12	1	16
Assault trial (All)	3	14	12	3	32
Assault trial, two verdicts	2	12	NA	2	16
Assault trial, three verdicts	1	2	12	1	16
Number of verdicts
Two verdicts (All)	3	24	NA	5	32
Three verdicts (All)	4	2	24	2	32
Number of jurors
12-person juries (All)	2	13	15	2	32
12-person, two verdicts	1	13	NA	2	16
12-person, three verdicts	1	0	15	0	16
15-person juries (All)	5	13	9	5	32
15-person, two verdicts	2	11	NA	3	16
15-person, three verdicts	3	2	9	2	16
Majority type
Simple majority (All)	6	14	12	NA	32
Simple majority, two verdicts	3	13	NA	NA	16
Simple majority, three verdicts verdicts	3	1	12	NA	16
Unanimous (All)	1	12	12	7	32
Unanimous, two verdicts	0	11	NA	5	16
Unanimous, three verdicts	1	1	12	2	16

While juries asked to choose between three verdicts were no more or less likely to return a guilty verdict than those with two verdicts available, when the not proven verdict was an option it was almost always chosen as the means of acquittal. Of the 32 juries asked to choose between three verdicts, 26 acquitted, and 24 of those 26 returned a verdict of not proven.

This pattern of acquittal verdicts is quite different to the actual pattern of acquittals returned in Scottish courts. In 2017-18, only 17% of all acquittal verdicts in the Scottish courts were not proven verdicts. However, the proportion of not proven verdicts returned was higher in more serious cases: for example, not proven accounted for 26% of all acquittals in homicide cases and 35% of all acquittals in rape and attempted rape cases.^[53] The 17% figure above includes both jury (solemn) and non-jury (summary) trials, while homicide and rape cases would all have been determined by a jury.^[54] The relatively low level of guilty verdicts returned by mock juries in this study, and the high proportion of acquittals that were not proven (compared with statistics for actual criminal trials in Scotland), is again likely to reflect the fact that the evidence in both trials was deliberately finely balanced.

Overall, then, the verdicts returned by the 64 mock juries in this study do not show that any of the three unique features of the Scottish system have an effect on the likelihood of juries convicting an accused person. As discussed, given the sample size, this is not surprising. However, the findings do suggest that - at least where the evidence is finely balanced - juries prefer to acquit via not proven rather than not guilty.

As discussed in section 3.2, the fact that there were no significant differences by any of the features examined at jury level is not surprising given that the number of juries (64) is relatively small in statistical terms. In the next section, we examine whether there were any differences across the much bigger sample of 863 individual jurors who deliberated in those juries.

3.4 Does jury size, the number of verdicts available, or the majority required make any difference to juror-level verdicts?

3.4.1 Overall patterns in the verdicts favoured by individual jurors

Before deliberating with their peers, 33% of jurors thought the jury should return a verdict of guilty, 42% favoured not guilty and 25% not proven. After deliberating, the proportion favouring a guilty verdict fell to 26%, not guilty remained the same at 42%, and the proportion favouring not proven increased to 32%. This indicates an overall tendency for individual jurors to shift from favouring guilty before deliberating to favouring acquittal afterwards.

The explanation for this shift to acquittal is not straightforward. As 51 of the 64 juries in this research started out with a majority favouring acquittal, this pattern might reflect a tendency for jurors to move towards the majority position (in this case, acquittal) as a result of deliberation. However, there is a body of existing research that suggests that deliberating with other jurors tends to create a 'leniency bias', whereby jurors are more likely to shift position from guilty to acquittal than from acquittal to guilty.^[55] The results might reflect this tendency.

3.4.2 Does the number of verdicts available have any impact on the verdicts favoured by individual jurors?

Individual jurors were more likely to support an acquittal and less likely to support a guilty verdict when the not proven verdict was available. 28% of jurors who had three verdicts available to them thought the verdict should be guilty before deliberating, compared with 38% of jurors asked to choose between guilty and not guilty only. Post-deliberation, 22% of three-verdict jurors thought the verdict should be guilty, compared with 31% of two-verdict jurors. The fact that three-verdict jurors were more likely to favour acquittal both before and after deliberating suggests that the availability of the not proven verdict has an effect on jurors' verdict preferences independent of any impact of deliberation.

This finding broadly reflects existing research (summarised in Chapter 1), which suggests - albeit not emphatically given the methodological limitations and relative paucity of such studies to date - that the availability of a not proven verdict may be associated with individual jurors being less likely to favour conviction.

The pattern of acquittal verdicts favoured by jurors in three-verdict juries also reinforces the jury-level finding, above (section 3.3), that for these trials jurors had a clear preference for acquitting via not proven. 79% of jurors asked to choose between three verdicts favoured acquittal after deliberating. This 79% splits into 65% who thought the verdict should be not proven, and 14% who thought it should be not guilty.

3.4.3 Does jury size have any impact on the verdicts favoured by individual jurors?

Unsurprisingly (since they had yet to discuss the trial with each other), jury size did not have any impact on the verdicts favoured by individual jurors before deliberating. However, after deliberating as a group, jurors in 12-person juries were less likely to think the verdict should be guilty than were jurors in 15-person juries (21% vs 30%). This is a result of jurors in 12-person juries being more likely to shift towards acquittal during deliberation than those in 15-person juries.^[56]

However, this does not necessarily indicate that 15-person juries would be more likely to return guilty verdicts across a larger number of differently balanced trials. Rather, it may reflect the fact that, where a 15-person jury is split, on average more people would need to shift their position to change the outcome. In 15-person juries, there may therefore be less motivation for those in a substantial minority (which, in this study, was usually those who favoured guilty) to change their individual position in order to bring deliberations to a close.

3.4.4 Does the majority required have any impact on the verdicts favoured by individual jurors?

The majority required did not have any impact on the verdicts preferred by individual jurors prior to deliberating. Again, this is unsurprising, since at that point jurors were unaware of how close or far away their jury might be from reaching the majority required to return a verdict. However, after deliberating as a group, jurors asked to reach a simple majority decision were more likely to think the verdict should be guilty than were those asked to reach a unanimous decision (32% vs 20%). Again, this occurred as a result of jurors in the unanimity condition changing their view to the majority opinion (which was, in most juries, to acquit) over the course of deliberating. Unsurprisingly, this shift to the majority view was more likely to happen when jurors had to agree on a unanimous verdict rather than returning a simple majority verdict. ^[57]

In order to return a verdict, juries in a unanimity condition have to persuade jurors who support a minority position to shift their view, while a simple majority jury can return a verdict even if those in the minority do not shift at all. However, further analysis indicates that jurors may be less willing to move from acquittal to guilty when the initial balance of opinion is the other way around. In the unanimity condition, six juries started out with a majority for guilty, but only one of these actually went on to return a guilty verdict - the other five hung. In contrast, of the 25 juries that started with a majority in favour of acquittal, 24 went on to acquit, and only one hung.^[58]

These findings also raise questions over the extent to which jurors asked to reach unanimity shift in order to facilitate a verdict, rather than because they have genuinely been convinced by the opposing arguments. Of course, it is important to remember that these findings are based on simulations rather than real juries. It is possible that, in a real trial, the inclination of jurors to shift their personal view to facilitate a verdict would differ.

Table 3.2 - Percentage of individual jurors personally favouring each verdict, Pre-deliberation by experimental condition (all jurors who deliberated)

	Conviction	Acquittal			Base (jurors)
	Guilty	Acquittal (all forms)	Not guilty	Not proven
All jurors who deliberated	33%	67%	42%	25%	863
Trial type
Rape trial (All)	40%	59%	36%	23%	431
Rape trial, two verdicts	44%	55%	55%	NA	216
Rape trial, three verdicts	35%	65%	18%	47%	215
Assault trial (All)	26%	74%	48%	26%	432
Assault trial, two verdicts	32%	68%	68%	NA	216
Assault trial, three verdicts	20%	81%	29%	52%	216
Number of verdicts
Two verdicts (All)	38%	62%	62%	NA	432
Three verdicts (All)	28%	73%	23%	50%	431
Number of jurors
12-person juries (All)	31%	69%	43%	26%	383
12-person, two verdicts	36%	64%	64%	NA	192
12-person, three verdicts	26%	74%	23%	51%	191
15-person juries (All)	34%	66%	42%	24%	480
15-person, two verdicts	40%	60%	60%	NA	240
15-person, three verdicts	29%	71%	23%	48%	240
Majority type
Simple majority (All)	32%	67%	43%	24%	432
Simple majority, two verdicts	37%	63%	63%	NA	216
Simple majority, three verdicts	28%	72%	23%	49%	216
Unanimous (All)	33%	66%	41%	25%	431
Unanimous, two verdicts	40%	60%	60%	NA	216
Unanimous, three verdicts	27%	73%	22%	51%	215

Table 3.3 - Percentage of individual jurors personally favouring each verdict, post-deliberation by experimental condition (all jurors who deliberated)

	Conviction	Acquittal			Base (jurors)
	Guilty	Acquittal (all forms)	Not guilty	Not proven
All jurors who deliberated	26%	74%	42%	32%	863
Trial type
Rape trial (All)	33%	67%	36%	31%	431
Rape trial, two verdicts	37%	63%	63%	NA	216
Rape trial, three verdicts	30%	70%	9%	61%	215
Assault trial (All)	19%	81%	47%	34%	432
Assault trial, two verdicts	25%	75%	75%	NA	216
Assault trial, three verdicts	13%	87%	19%	68%	216
Number of verdicts
Two verdicts (All)	31%	69%	69%	NA	432
Three verdicts (All)	22%	79%	14%	65%	431
Number of jurors
12-person juries (All)	21%	79%	40%	39%	383
12-person, two verdicts	28%	72%	72%	NA	192
12-person, three verdicts	15%	85%	7%	78%	191
15-person juries (All)	30%	70%	43%	27%	480
15-person, two verdicts	33%	67%	67%	NA	240
15-person, three verdicts	27%	73%	19%	54%	240
Majority type
Simple majority (All)	32%	67%	39%	28%	432
Simple majority, two verdicts	36%	64%	64%	NA	216
Simple majority, three verdicts	29%	71%	14%	57%	216
Unanimous (All)	20%	80%	44%	36%	431
Unanimous, two verdicts	25%	75%	75%	NA	216
Unanimous, three verdicts	14%	85%	13%	72%	215

3.4.5 Was the impact of the Scottish system's unique features the same in the rape and assault trial?

This research included two types of trials - a rape trial and an assault trial.^[59] Jurors in the 32 rape trial juries were more likely to think the verdict should be guilty than were jurors in the 32 assault trial juries, both before deliberation (40% vs 26%) and after (33% vs 19%). Considering these differences, the analysis above of juror views on the verdict was repeated separately for each type of trial.

Differences between jurors asked to reach a unanimous verdict and those asked to reach a simple majority were apparent across both trial types. In each case, those asked to reach a simple majority were more likely to favour a guilty verdict after deliberating than those asked to reach unanimity.

Both rape and assault trial jurors were less likely to favour a guilty verdict after deliberating when the not proven verdict was available than when it was not,^[60] but this difference between the two-verdict and three-verdict conditions was only statistically significant for the assault trial. Similarly, in both trials, jurors in 15-person juries were more likely than those in 12-person juries to favour a guilty verdict after deliberating,^[61] but again this difference was only statistically significant for the assault trial.

Findings for each individual trial type are based on around 430 jurors, compared with the full sample of 863 jurors. This means that bigger differences are required to reach statistical significance when analysing findings for each trial type separately.

On the basis of this data it is not possible to conclude that availability of the not proven verdict or jury size made a significant difference to juror verdict preferences in the rape trial. However, the finding that requiring unanimity increases the proportion of jurors favouring acquittal holds across both types of trial.

3.4.6 How do jury size, number of verdicts, and the majority required interact with each other?

Our analysis indicates that each unique feature of the Scottish jury system was independently and significantly related to the likelihood of individual jurors favouring a guilty verdict, in the ways described above.^[62] However, these features are also likely to interact with each other. Table 3.4 shows differences in the proportions of jurors favouring a guilty verdict before and after deliberating, within each of the eight possible combinations of the three features. Condition F in Table 3.4 is the jury system presently used in Scotland (15-person, simple majority, three verdicts) and condition B is the jury system used in England and Wales (12-person, unanimity, two verdicts).

Of the three features, the majority required has the biggest impact on the likelihood of individual jurors shifting towards acquittal - jurors were most likely to change from guilty to acquittal when required to reach a unanimous verdict. Jurors were less likely to think the verdict should be guilty before deliberating when the not proven verdict was available. However, the impact of the number of verdicts available was moderated during deliberations by the majority required. Regardless of the number of verdicts available, if the jury had to reach a unanimous verdict, jurors were more likely to shift to support acquittal.

The combination of features most likely to result in individual jurors favouring a guilty verdict after deliberating was 15-jurors, simple majority, and two verdicts (condition H in Table 3.4, where 37% favoured a guilty verdict). The combination least likely to result in individual jurors favouring a guilty verdict after deliberating was 12-jurors, unanimity, and three verdicts (condition A in Table 3.4, where just 3% favoured a guilty verdict). It should again be emphasised that these figures are based on two deliberately finely-balanced trials which are not representative of the full range of cases which come before the courts. Neither of these combinations represents the jury system currently used in Scotland or in England and Wales, although the first would be the system in Scotland if the not proven verdict were removed with no other changes being made.

Table 3.4 - Proportion of individual jurors favouring a guilty verdict pre- and post-deliberation by combination of features

Condition	Jury size (15 or 12)	Majority required (simple majority or unanimous)	Number of verdicts (3 or 2)	Pre-deliberation % guilty	Post-deliberation % guilty	Difference (post minus pre)	Base (jurors)
A	12	U	3	23%	3%	-20	95
B	12	U	2	39%	21%	-18	96
C	15	U	3	30%	23%	-7	120
D	12	SM	3	29%	27%	-2	96
E	15	U	2	41%	29%	-12	120
F	15	SM	3	28%	30%	3	120
G	12	SM	2	34%	34%	0	96
H	15	SM	2	38%	37%	-2	120