As it stands, juries are free to select their preferred voting method. For instance, they might choose for it to be cast publicly (such as through a show of hands or a roll-call vote) or in the form of a secret, written ballot. In a 2006 edition of Lawyers Weekly, trial consultant Edward Schwartz advised lawyers on how best to subtly bias (or “nudge”) the jury in favour of their client. Schwartz recommended that lawyers facing a polarised jury ought to include a phrase in their closing speech to the jurors that mentions the tearing up of slips of paper (mentally priming a secret ballot) whereas a lawyer expecting a majority faction of jurors to side with their argument ought to mention the raising of hands or speaking in turn (implicitly encouraging a public vote). Given the importance of secrecy in democratic elections, it is nothing short of remarkable that a secret voting mechanism has not yet been mandated on the jury.
Considering the global prevalence of the jury trial (they are used in Australia, Canada, France, Greece, Ireland, Hong Kong, Japan, New Zealand, the UK and the US, among others) and their irrevocable consequences for societal welfare (a third of Americans will serve on a jury in their lifetime, and the 150,000 annual jury trials in the US cost its government $200bn each year) the claim that jury trial outcomes are not invariant to manipulation of the jury’s voting procedure would be completely unacceptable, if proven to be true.
The notion that the jury’s decision-making could be subject to socially normative biases under a public vote is not surprising. Students’ studying decisions are affected by whether participation in study schemes is publicised. Men donate more money when in the presence of a female audience. Employees put more money into an honesty box, when taking coffee and a cake, if the menu of prices displays a pair of human eyes. Introducing a postal vote has decreased voter turnout in some small communities where participation can no longer be visually verified by the community.
In the legal context, judges deciding whether to grant asylum have been found to artificially oscillate between a “yes” and “no” verdict in order to appear fair and impartial to onlookers. Given the systematic human proclivity to care about what others think of our decisions, it is important that policymakers take this into consideration when conducting institutional design in the courts and elsewhere.
While there have been several intuitive arguments in favour of the secret ballot, there has been a notable lack of formal theoretical arguments or presentations of empirical evidence in favour of either a secret or public vote on the jury. This presented me with a natural and unique opportunity to compare whether members of the jury are more likely to vote in accordance with their own private beliefs over the innocence or guilt of the criminal defendant, under either voting mechanism.
Using the tools of game theory, I began by modelling the jury’s voting procedure and subsequent deliberations as a “game” and the jurors as strategic “players” who must vote optimally given their private information (that is, both their verdict preferences and how certain they feel about their choice) either in accordance with their own beliefs or in opposition to them. Unanimity in the vote ends the game, whereas non-unanimity triggers a second game in which jurors must optimally balance the benefits of persuading others of their viewpoint against the time and effort required to do so (the jury deliberation process).
Grinding through the algebra unearths a simple finding. The probability that jurors vote truthfully is greater under a secret ballot if the jury trial is ambiguous, in which several jurors disagree with each other over whether to convict or acquit. I then verified that these predictions hold in the real world, using a data set on hundreds of criminal jury trials in the US in 2000-01.
Using advanced statistical techniques, I discovered that the taking of a secret ballot caused truthful voting to increase by 8.4 per cent in ambiguous jury trials, which is equivalent to one incremental vote in a standard twelve-person jury. If one reasons that jurors are most prone to convicting an innocent defendant or vice versa in such ambiguous cases, then the secret ballot can be seen to increase truthfulness in the court cases where the choice of voting mechanism matters most.
Many countries have abolished jury trials on account of juror bias. They include India, Singapore and South Africa. With judges exhibiting similar biases, such as the fact that they are more likely to grant parole just after lunch, and the controversial use of artificial intelligence on juries still in its infancy, it is imperative that the jury system is reformed to promote truthful voting. From my own research, it is clear that mandating a secret ballot is a necessary step in achieving this.
