Director of Public Prosecutions v Folau (Ruling No 5)
[2022] VSC 133
•17 March 2022
| IN THE SUPREME COURT OF VICTORIA | Not Restricted |
AT MELBOURNE
CRIMINAL DIVISION
S CR 2020 0311
| DIRECTOR OF PUBLIC PROSECUTIONS | Crown |
| v | |
| TORRIS FOLAU | Accused |
---
JUDGE: | LASRY J |
WHERE HELD: | Melbourne |
DATES OF HEARING: | 24-25 November 2021, 16 December 2021, 11 March 2022 |
DATE OF RULING: | 17 March 2022 |
CASE MAY BE CITED AS: | DPP v Folau (Ruling No 5) |
MEDIUM NEUTRAL CITATION: | [2022] VSC 133 |
---
CRIMINAL LAW - Trial - Charges of murder, intentionally causing serious injury and home invasion – Application for the exclusion of DNA evidence – Whether DNA evidence ought to be excluded pursuant to s 137 of the Evidence Act 2008 (Vic) – Whether probative value of evidence outweighed by danger of unfair prejudice – Reliability of likelihood ratios - Probability the accused is a contributor - Evidence relevant – Probative value not outweighed by danger of unfair prejudice – Identified issues with the evidence a question for properly instructed jury – Application refused – Cases cited: R v Tuite [2014] VSC 662; Tuite v The Queen [2020] VSCA 318.
---
APPEARANCES: | Counsel | Solicitors |
| For the Crown | Mr Patrick Bourke QC with | Office of Public Prosecutions |
| For the Accused | Mr John Desmond | Matthew White & Associates |
HIS HONOUR:
Torris Folau (‘the accused’) is charged with one charge of home invasion, one charge of causing serious injury intentionally and one charge of murder. The circumstances from which these charges arise all occurred on the 16th of May 2020.
On that day, at about 8:30pm, the accused is alleged to have gone to premises at Nolan Drive, Epping with a person named Chris Segio. Earlier in the day, sometime after 4:00 am, an occupant of those premises, Nadim Ghattas had been the victim of an armed robbery. He suspected a person named Christian Movondo had been involved in that event. Later when the accused and Segio attended the property a number of people involved in the drugs milieu were present. They were not known to those occupants. So, at about 8:30 pm on that day, the accused and Segio are alleged to have forced their way into the property. From one room they are alleged to have taken possession of a knife and then went to another room where the accused is alleged to have stabbed Hashmat Nawabi and assaulted Ahad Mohamad. Nawabi died of his injuries and Mohamed received a fatal brain injury.
The accused has indicated he will plead not guilty to all three charges and the issue in the trial, which is to commence on 28 March 2022, is whether or not he was even present at the scene when these incidents occurred. The defence response seems to suggest that beyond that the accused proposes to put the prosecution to its proof on all elements of the offences charged.
There is a procedural history in this matter which has, among other things, involved preliminary hearings pursuant to s 198B of the Criminal Procedure Act (2009) and has led to four separate rulings on pre-trial issues. The evidence from the relevant witness regarding DNA evidence was given in November and December 2021 but due to the commitments of counsel for the accused, the final hearing of submissions on this pre-trial issue concerning the admissibility of DNA evidence was adjourned to 11 March 2022. The trial before a jury will start on 21 March 2022.
Part of the evidence on which the prosecution seeks to make good its case against the accused is DNA evidence relating to the analysis of four crime scene samples. Under that heading in the summary of prosecution opening,[1] the following description appears:
[1]Dated 19 February 2021.
76.The interior of the Commodore’s cabin was tested for possible blood. A number of areas that were identified as positive to a screening test for blood were swabbed, and the swabs tested for DNA. Relevantly:
(a)A sample taken from the driver’s door armrest revealed four contributors to the DNA. The DNA evidence was 75,000 times more likely if the deceased was a contributor and 100 billion times more likely if he (sic) accused was a contributor;
(b)A sample taken from the driver’s seat back revealed four contributors to the DNA. The DNA evidence was 2.9 billion times more likely if the deceased was a contributor;
(c)A sample taken from the rear of the front passenger seat revealed three contributors. The DNA evidence was 160 thousand times more likely if the deceased was one of the contributors and 100 billion times more likely if the accused was a contributor.
77.The DNA evidence in relation to these samples either excluded Segio as a contributor to the DNA, or the evidence was more likely if he was not a contributor.
78.The cigarette butt found outside 19 Nolan Drive was also tested for DNA. That testing revealed two contributors to DNA located on the end of the butt. The DNA evidence was 100 billion times more likely if the accused and Segio were the contributors to that DNA.
An application has been made by Mr Desmond on behalf of the accused for all of that evidence to be excluded, pursuant to s 137 of the Evidence Act 2008 (Vic) (‘the Act’). As is well known, the section provides that:
In a criminal proceeding, the court must refuse to admit evidence adduced by the prosecutor if its probative value is outweighed by the danger of unfair prejudice to the accused.
It is obvious that since an issue in the trial is whether the prosecution can prove the accused was present at the scene where these crimes are alleged to have occurred, the DNA evidence will be significant.
It is asserted on behalf of the accused that there is unfair prejudice to him in admitting the evidence, although the exact nature of that prejudice is somewhat opaque. At the basis of his submission is the claim that the likelihood ratios to be relied upon in the DNA evidence are unreliable because they are inconsistent with the individual parts of the DNA analysis. It is therefore argued that there is a prejudice to the accused.
I note that the issues concerning the likelihood ratios and their reliability have been the subject of careful analysis by this Court and the Court of Appeal. I also note that the only evidence about the DNA in this case is from the prosecution witness, Kate Outteridge. In evidence on the voir dire she rejected Mr Desmond’s thesis about the reliability of the likelihood ratios.
The application that the DNA evidence be excluded pursuant to s 137 of the Act is put on the following basis:
It is first conceded by Mr Desmond that the evidence is relevant since it goes to a fact in issue as to whether the accused was at the scene when the offences he is charged with were committed.
Next, however the argument is put that when ‘critically examined/analysed’ the likelihood ratios that will be the subject of evidence ‘overstate the weight or strength of the evidence’. [2] That is because, in Mr Desmond’s submission, the individual low weightings which he points to are inconsistent with high likelihood ratio in the overall analysis of the particular samples. As I said, Ms Kate Outteridge, gave evidence on the voir dire in November and December 2021 and made it clear in that evidence that “You can't use the weightings themselves to judge what the likelihood ratio is going to be”. [3] That is where the dispute seems to lie. I will return to it.
[2]Defence submissions on the admissibility of the DNA evidence dated 24 January 2022, [17].
[3]Evidence of Kate Outteridge during preliminary argument on 25 November 2021, 84.
The next step in Mr Desmond’s submission is that the probative value of the likelihood ratio evidence taken at its highest is not particularly high or strong. However, that argument assumes the correctness of his analysis that low weightings must inevitably be reflected in the likelihood ratio. The only evidence about that is from Ms Outteridge and her evidence makes it clear that because the STRmix process is comparative between more than one proposition, counsel’s proposition misunderstands the process.
A consequential point that Mr Desmond also seeks to make is that Ms Outteridge is only authorised to report using the likelihood ratio method and results. In his submission, the absence of a separate forensic biologist solely reporting on the probability the accused or the deceased (as the case maybe) is a contributor to the crime scene item without comparing the same with the probability of person unknown were probably contributors to the given sample creates an inherent unfairness on or to the accused. This submission appears to demonstrate an unwillingness to confront the endorsed manner in which DNA evidence is given. There is no need for a separate forensic biologist performing the role suggested. In my opinion Ms Outteridge’s evidence is all that is required.
In Mr Desmond’s further submission, Ms Outteridge is unable to explain the apparent inconsistency between individual genotype and component weightings expressed in percentages/probabilities with the likelihood ratios. However, as I understand it, whilst she would not claim to understand the mathematical mechanics of STRmix, she does understand the process and has explained it, an example of which is as follows:[4]
And so when the likelihood ratio is calculated that's the - the first proposition is if we just consider one person being compared. So, if the first proposition is assuming that that person is a contributor, so the likelihood - likelihood ration will be that, yes, that person is a contributor, plus those three other people. But it incorporates all of the possible combinations that those other three people could have. So, for example, um, in this particular profile Mr Folau is a 15 18, so if we look at that as contributor 2.
So, in the likelihood ratio that 15 and 18 would be um, taken to be from him, and then the other DNA types would be ah, seen as from contributors 2, 3 and 4, in all of their varying combinations, and they will be used for that ah, the first hypothesis which is the (indistinct) hypothesis. Then when the - we look at the reverse, which is that Mr Folau is not a contributor and that there's ah, someone else from the population.
STRmix then will calculate the ah, defence hypothesis which is the second one, and then it can use any combination of the DNA types that are seen at that locus. So, it looks at any particular combination and the likelihood of those occurring in the general population. And then it divides the value obtained from the prosecutor's hypothesis by the number obtained from the defence hypothesis, but also incorporated into that calculation is the frequencies of the DNA types within the population and all of the weightings that it has determined during that convolution.
So all of those are incorporated into the likelihood ratio which then results in the final number which could be above one, it could be below one, it could be zero excluding a person so the final likelihood ratio is the best indication of how well the person aligns with the evidential sample that it's been compared to. And for - the example that you were just going through previously where the likelihood ratio was 2500, it's not a high value therefore it's not giving a huge amount of support for the proposition that Mr Folau is a contributor, and that is reflected in the fact that the likelihood ratios are low that - in the weightings might not be in the top combinations, so all of that is all adding information to why that final likelihood ratio is lower than if he was well represented in an evidentiary sample.
[4]Ibid, 82.
As this evidence demonstrates, the witness has a clear and logical understanding of STRmix and the way in which it analyses the information, including the division of the value obtained from the prosecutor's hypothesis by the number obtained from the defence hypothesis, and the frequencies of the DNA types within the population and can give that explanation before a jury.
On 16 December 2021, in a further voir dire, the witness returned to Court with a view to giving a further explanation of the manner in which STRmix works, and by way of PowerPoint presentation expanded on her understanding and explanation of the manner in which it calculates the likelihood ratios in a particular case.
In brief summary she described how STRmix does a deconvolution looking at DNA types that are seen at particular sites and prepares combinations that could be best explained by at least two people contributing. In the deconvolution the STRmix program incorporates all of the combinations that can best explain each of the contributors and discount those that are not logical with regard to the types seen. This is based on information seen on the peak heights.
Coming to the significant issue as to whether the likelihood ratio misrepresents the strength of the evidence, the witness made clear that she did not agree with that and in her evidence explained that what the likelihood ratio does is compare two propositions into which the weightings are incorporated, producing a number that indicates support for one proposition or the other. There needs to be two propositions to give the analysis meaning.
I agree with the prosecution submission that Mr Desmond’s submission erroneously isolates from the STRmix data the prosecution hypothesis at each individual marker and then attempts to argue that those figures are miniscule. In my view, Ms Outteridge has made it clear this is not a valid approach because, for among other reasons, the likelihood ratio depends on the prosecution hypothesis being divided by the defence hypothesis.
Also, as the prosecutor has pointed out in his written submissions, a concise summary of how STRmix calculates a likelihood ratio was given by Emerton J (as she then was) in R v Tuite.[5] For whatever reason the written submissions filed on behalf of the accused make no reference to the rulings of this Court or the Court of Appeal in that case. He sought to distinguish that case but the issue being litigated several times was the reliability of the likelihood ratios.
[5][2014] VSC 662, [22]-[27].
Mr Desmond concluded his submissions by arguing that any explanation by Outteridge to the jury would be incomprehensible and confusing, and the admission of the DNA evidence would require the jury to listen to, absorb and understand all the arguments which would leave them ‘hopelessly confused’.[6] I disagree.
[6]Defence submissions on the admissibility of the DNA evidence dated 24 January 2022, [17] (i).
In this trial, as in many others, the process by which the likelihood ratios are reached in the operation of STRmix is explicable and a jury is quite capable of understanding it. The argument that the likelihood ratios overstate the strength of the DNA evidence can be argued at trial, within limits. There is no unfairness.
As the Court of Appeal said in Tuite v The Queen, this very issue has been ‘exhaustively investigated’ [7] and, in my view, the basis on which this evidence is to be given is settled and not a matter of scientific controversy. No witness in this proceeding gave evidence to contrary and no alternative expert scientific thesis is foreshadowed on behalf of the accused. As the Court went on to say later in the judgement:
[113]In an ordinary case, the DNA evidence would have taken up far less time, because the expert witness would only have needed to inform the jury of the relevant likelihood ratios, and to provide a basic explanation of what they meant. Even then, as this Court pointed out earlier in the year in Vyater v The Queen, care is required to ensure that the language used is as simple and comprehensible as possible. All a jury needs to know is what the statistical results are, and what they do (and do not) tend to prove. They do not need to understand the underlying science of DNA profiling.
[7][2020] VSCA 318, [39].
In my opinion, the evidence is probative and I do not accept there is any danger of unfair prejudice in the form of misuse by the jury. Obviously, evidence will not be prejudicial simply because it increases the likelihood of conviction. The jury will be given clear directions about the use of this evidence. I therefore have concluded that the application to exclude this evidence should be rejected. This conclusion applies to all of the DNA evidence in this trial, whether explicitly referred to in the prosecution opening or not. In my view the basis of the application to exclude the evidence rests on an unwillingness to come to grips with the way DNA evidence is analysed and presented. It proceeds on the basis that the way in which STRmix produces likelihood ratios should, in some way, be anchored to the genotypes at individual sites. I accept that such an argument is simply incorrect and in order provide a proper analysis the result must be comparative in relation to at least two competing propositions, taking into account all aspects of the DNA profile.
Finally, I would add that in the course of their judgement the Court of Appeal gave a clear direction about the conduct of these issues in the following terms:
[116]For a case concerning a single incident of offending, with a very limited factual compass, these figures are simply astonishing. It is, of course, critical to the integrity of the criminal justice system that forensic evidence be tested according to rigorous reliability standards. It is not in doubt that successive defence counsel were acting conscientiously in pursuit of what they assessed to be the applicant’s best interests. It is, however, incumbent on the parties to a reliability challenge such as was mounted in this case to impose appropriate limits on the resources devoted to the task.
[117]Such limits are essential not least because of the need for the scarce resources of the criminal justice system to be available for similar challenges in other fields of forensic science, where the scrutiny of evidentiary reliability has been much less intense than it has been in relation to DNA evidence. As the Court of Criminal Appeal said in R v Wilson and Grimwade:
Whether the cost of legal representation be privately or publicly borne, counsel are to understand that they are exercising a privilege as well as fulfilling a duty in appearing in a court of law.
As best I understand it, the primary attack on the reliability of the likelihood ratios is based on the claimed discrepancy between that ratio and individual genotype and component weightings. At this stage, I simply indicate that during the trial there will be significant limits on the extent to which, and the mode by which, those matters are canvassed in cross examination, particularly given no contradicting witness to Ms Outteridge is proposed to be called on the behalf of the accused. I will not permit this to become a contest between the witness and the thesis created by counsel for the accused unsupported by expert evidence.
The basis of which counsel suggests the likelihood ratios overstate the evidence can be raised with the witness but not exhaustively pursued.
0