..Looks like a quiet night. I need to get something off my chest. I have just received a notice from the Juries Commission in Victoria that I am wanted for jury service. It’s one of the letters a busy person dreads. You cannot get out of it, even by paying a fine. And they are relentless in their pursuit if you try to delay – like a dog with a bone. While jury service may be our civic duty, you would hope that the the courts would give some weight to the level of inconvenience they impose. They might also check their understanding of sampling principles. Let me tell you why.
Around two hundred potential jurors will have to attend the supreme court on the appointed day, even though they ultimately need only a fraction of them. I know this because the person on the phone said I have a small chance of actually being chosen for the jury. But I have to hang around for two days while they decide. There will be a random ballot and some selection procedure based on visual appearance rather than interview to obtain the jurors required.
Do they really need two hundred people for two days? How about allowing challenges before we turn up? Many challenges are based on demographic information that is available to the court and does not require the potential jurors attendance. You would just receive a letter that you had been selected as a potential juror, and then challenged, and told to get on with your life.
One also wonders why N=200 are required. How many should they call? Well, if p is the historical proportion of selected jurors that survive the challenge procedure, and J jurors are required, then N=J/p should just about do it. And if you are worried that this number might be insufficient then you can add in a bit of padding and/or bring in some reserve jurors the next day. Once you assign cost to a potential juror’s day and a cost to a day of court procedure, this is a pretty easy optimisation problem. As it is, we have 200 people lose at least one day which must be valued at around $50,000 or so. But this cost is born by the jurors and their employers, not the court.
So by receiving the letter it looks like I will lose two days at the very least. Is this fair? Well, we might say it is fair if people are chosen randomly. According to the court:
Potential jurors are randomly selected from the Victorian Electoral Roll. A number of random processes are used throughout the jury selection process in order to produce a representative cross-section of the community. Every person over the age of 18, from all sections of the community have an equal chance of being called upon to do jury service.
I am suspicious about the effectiveness of their sampling. During the first six months of this year, there have been 5 academics (out of about 40) from Melbourne Business School summoned! Unless they are calling tens of thousands of people, there is no way that is “random” in the sense that a pollster would understand it.
The outcome has the appearance of some kind of sequential sampling from a list that has been sorted by geography – like an electoral roll. An extreme example would be to randomly sample phone numbers in consecutive order – just work your way through the phone book and toss a coin for each person to see if you summon them. This is random and it also gives everybody (with a phone) the same chance of being a juror. But the sampling is not independent – you can end up with sets of similar people on the jury. Imagine that you get to the 9349xxxx section of the phone book (which is what MBS phone numbers start with). I wonder how much respect the public will have in the notion of “a jury of your peers” if these peers end up being a bunch of middle aged business professors!
While the court’s understanding of sampling seems dodgy, some readers might be thinking the same thing about me. I have chosen to report the statistic “5 MBS academics out of 40 in a half year” after the fact. If there were only 1 person summoned then I would not have reported it. This is the same kind of reporting bias that we face in assessing clustering of disease. It would actually be surprising if clustering of some disease did not occur somewhere. But when you see the cluster post hoc, it looks significant.
In my case, such selection/reporting bias would be most problematic if everybody in Australia had a blog and could report such an occurrence. You might then expect such a concentration somewhere in a given (half) year. But only a small proportion of people are bloggers. From the point of view of you gentle readers, the relevant blogging population is those who post at this site. The chance of one of us experiencing such a summons concentration if sampling is truly random is, I would have thought, pretty small. But to answer for sure I would need to know the total number of summons in a year and the numbers on the electoral roll. It may turn out that “5 MBS academics out of 40 in a half year” is consistent with a sensible sampling scheme, but I doubt it.
But there is a much less debatable point about the sampling that I want to make. If I am unlucky, I will be one of those chosen to actually sit in court. I will then be exempt for three years only….after which my name is added to the pool to be randomly sampled. But hold on…the purpose of random sampling is to give everybody an equal chance of being selected. This is desirable to spread the burden evenly and also to make the jury representative. No problem with the aims. But if I have already served once why do I get another chance? Surely I should wait until every eligible person in Victoria has served, before I am punished summoned again.
It seems that the courts have never heard of sampling without replacement.
What is the chance of being called twice I hear you cry? Well, yours truly already served in 2002 as a juror on a murder trial. While it provided me with some interesting stories for dinner parties and probability lectures, I did lose three weeks of my life, never to be regained. The case was in NSW which, under their rules, exempts me for 8 years. But since the case was in NSW, it has no effect on my eligibility in Victoria. It is a separate jurisdiction. Another example of the our federal system at work.
Now a wise but cynical person once defined a jury to be
a group of twelve people of average ignorance, chosen to decide who has the best lawyer.
Harsh perhaps, but not indefensibe. Perhaps this might explain why jurors are paid only $36 per day, and why no lunch is provided. That’s correct. The supreme court can not find the spare cash to provide the jury a decent meal. No wonder those 12 guys look so angry.