Just some brief comments, your Honor. Thank you. I think the narrow issue before the Court frankly is foundation at this point. The Court asked there be a foundation provided. The People had the burden of providing a foundation for the admissibility of the frequencies, that portion of Dr. Weir's work that refers to frequencies, not likelihood ratios.
Uh-huh. But I'm also addressing the 352 issue as to whether or not it would be unduly confusing to the jury to present in this manner or whether or not it would have the possibility of misleading the jury, this manner of presentation. That's--
I think I might address those best by describing the intent with which we had in the manner of offering this type of evidence.
All right. We're talking about 29, 78. Which other mixtures are we talking about? The Bronco console stains?
Exactly. There are several in the Bronco, there are several if not nearly all--they're nearly all on the glove. In other words, nearly all of the results on that board. So those are by far the bulk. So it's largely on the glove board, to a lesser extent, on the Bronco board, and then one item, no. 78, on the Bundy scene board. So that's what we're dealing with in our view as far as the mixtures. It was our intent and using--and we'll just use the example of no. 29, the Bronco stain, to use a range nearly--well, actually identical in terms of its description on the board and to the jury--from the most common number to the rarest number for those mixture frequencies calculated by Dr. Weir. And one of those examples would again be--and I just don't have the page. Does the Court have the report before it?
All right. And, for example, with regard to item 29, we put a 2 colon, meaning assuming two contributors, and the range would be, and just looking at that particular chart, from 1 in 71--I'm sorry--from 1 in 59--that's actually the most common--to 1 in 11,300. That was the most uncommon. And then with the example of 29, put the number 3 colon and do the same thing and making it clear that that's assuming three contributors to the stain. So that's the manner of presentation in terms of how this would be presented to the jury. I don't think in any manner it's more complicated than frankly the methods used to calculate a frequency for an overall stain, for a single person stain. So I don't think the Court should have concern about that because we could be spending days and days on simply the single stain donors or could have that were presented previously to this jury. So I don't think 352 is a valid basis to prevent this type of calculation any different than I don't think 352 by itself, other than the numbers as suggested by the Defense, are simply misleading. In fact, they're terribly misleading as Dr. Weir described earlier. But addressing the issue of admissibility in itself, as far as the decision this Court has to make, it is not a decision of general acceptance in our view. That was waived many, many months ago by this Defendant at a time when these mixtures were clearly known in terms of the samples and results in this case, which we argued to this Court back in December and January. The question is, in this situation--and the Court's evidence at this point is frankly the testimony of two experts and a series--well, a few exhibits, also including some of the peer review scientific literature. The fact that two experts reach contrary conclusions or opinions rather about the use of a particular technique does not render that type of testing result and the method used inadmissible in the context of this hearing.
No. I don't think Dr. Shields and Dr. Weir necessarily disagreed. Dr. Shields agreed that if I accept the appropriateness of making assumptions, then that mathematical calculation is the appropriate way to do it. So they agree on the math. They just don't agree to the nature of the presentation.
Actually the Court anticipated the next item I was going to state, which was that, when one carefully listened to Dr. Shields' testimony, as the Court has already noted, he conceded that if there were two contributors, Dr. Weir's calculations were absolutely correct. The same answer was given by him to the assumption that there are three contributors. This jury has the right to make or reach certain conclusions about what happened in this case. And here's an example. This jury might mind or a jury in any case might find, this case perhaps less than others, that even though it's given frequencies for a matching stain of 1 in a certain number for one database such as Caucasian, 1 in a certain different number for African Americans and so forth, the jury may hear other evidence that satisfies it as to which database is the most appropriate to use. There may have been an eyewitness to a crime. And, therefore, the jury can decide well, that assumption I can now make based on the evidence that's been presented, that this person is from a certain category. Therefore, that number becomes more relevant to that jury or individual juror. This is no different because this jury can reach a decision that as to some of these items of evidence, I'm satisfied that there were two contributors. Therefore, Dr. Weir's number about two contributors, even Dr. Shields concedes is absolutely accurate. And I think that goes hand in hand again with allowing the jury to reach certain assumptions or make assumptions if the evidence supports it. If it doesn't, then they're given the ranges, just as they were with frequency data for single donor stains.
And I think that's absolutely admissible under our law. These techniques of calculation are complicated as they're complicated for single stains. So I don't think there's any reason this jury cannot deal appropriately with this evidence in the same manner that they can deal appropriately with the evidence that's already been presented. I think even a simple reading by a layperson such as ourselves involved in the Criminal Justice System of the National Research Council report can see why Dr. Weir's testimony about that report and its single sentence is perfectly compatible with his testimony and the method that he described today. I think Dr. Shields was overstating it for all of us when he said that that language in the national research counsel report is unambiguous.
But, Mr. Clarke, doesn't any trial Judge have to be very cautious about adopting techniques and testimony that is contrary to the NRC report?
Not at all. I think the testimony before this Court--and if I recall, the Court noted some weeks ago the fact that that report in and of itself was not only subject to controversy, but I think the Court used the term "Outdated."
I don't think I ever said that. I don't recollect having said that. But certainly, the case law dealing in this area often comments that the legal system is often years--a year or two behind the scientific community in these matters.
Fortunately, in California, we see our case law going away from the National Research Council report if we can Judge anything by the last half a dozen decisions or so. So if the Court is concerned about Criminal Justice System versus the National Research Council report, I think one only needs to look at California, let alone the number of states across the country that either rejected it in terms of having any effect on admissibility or finding that things have moved away from that, and that frankly that report is now very old in an area that's moving quickly. So no, I don't think courts should be concerned about the National Research Council report playing a role and telling courts what to do. I hope that answers your question. Based on all of these reasons and again considering the nature of this hearing, it's our opinion that this Court should allow and in fact the jury should be given the benefit of conservative, but probative information about what these mixtures mean to this jury. And I think no one could state it better than Dr. Weir has today in terms of the significance of this evidence, not even incorporating the known individuals in this case which ultimately would arrive or lead to a likelihood ratio that we're not even seeking to use at this time. Thank you.
Well, your Honor, I think the reason that Dr. Weir's statistics are likely to be misleading in the context of the trial at this point is that the Prosecution has already framed the issues for the jury. We've already had a great deal of testimony about DNA evidence and the testimony has all been framed in terms of here's a sample and here's an individual. Is this individual consistent with this stain? Again and again, they framed it in that--using the terms "Match" or "Consistency." And when that kind of evidence is presented, it is always logically relevant to ask what percentage of the population would also be included or be excluded, which is exactly the number that the National Research Council computes. Now, Dr. Weir has asked a different question which he thinks is the appropriate question which has to do with what's the probability of selecting a specified number of individuals who could completely account for the alleles in this stain under various assumptions. And we have to make assumptions about the number of people and whether all alleles can be seen and so on. So--so what--what's happened is that the Prosecution is changing horses in mid-stream. Having conditioned the jury to expect one kind of numbers, they are now presenting a different kind of numbers that--that in fact are--are much less favorable to the Defendant. I think that the difference in the numbers is going to be very difficult to explain and I think all you have to do is read through Dr. Weir's report to see why that's so. It's very complex. It's riddled with assumptions, and the assumptions are going to be controversial. If Dr. Weir is allowed to testify under these assumptions, then the Defense will have to challenge those assumptions. For example, on stain 78, Dr. Weir is assuming for some--for some of the locuses, that all alleles from all contributors can be seen. For others, that not all alleles of all contributors can be seen. We're going to have to get into the scientific basis for that, talk about whether that assumption is valid or not valid based on the scientific evidence as opposed to other evidence in the case.
But aren't we sort of in this situation because the Prosecution initially wanted to say that all we want to do is say that these persons are not excluded and that's all we want to present, it's a mixture, and then the Defense says no, you've got to put on a number on it if you're going to use it?
Yeah. And both experts acknowledge that it would be misleading to simply say this person is not excluded and here's the frequency of this person's genotypes. Both experts acknowledge that was misleading and inappropriate. For example, in the case of Nicole Brown Simpson, who has a genotype frequency of 1 in 22,000 or something like that, but the likelihood of her being--of her being consistent with the stain is either 50 percent or so under our calculations or 1 in 77 under Dr. Weir's calculation, but much higher than the frequency. So it's--it's absolutely clear that we have to have numbers in connection with the mixtures, and the issue I think is which numbers. And I think it's very clear that the National Research Council numbers are the most appropriate ones. These are--they're the easiest to understand because they involve fewer assumptions, fewer of any assumptions. They can be presented with much simpler charts. They are consistent with the nature of the frequency data that the jury has already been conditioned to expect and so on. Now, in addition to being difficult for jurors to understand, our position is that Dr. Weir's numbers are also argumentative in that, as he acknowledges, certain assumptions need to be made about number of contributors and particularly this tricky one about whether all alleles are accounted for. Now, Mr. Clarke said well, that's fine. We'll just ask him to make a number of different assumptions. But it's clear that Dr. Weir is not going to be able to make all possible assumptions. There are an endless variation of assumptions. So necessarily the Prosecution is going to be picking and choosing among the assumptions that he comments on.
But I think there's only going to be two assumptions, the number of contributors and whether or not you can completely see the alleles present.
Well, and which--which--which databases are involved, whether--what--the racial background of the contributors.
Yeah. But this assumption, for example, about whether all alleles can be--can be observed, that assumption--if--he's making different assumptions at different locuses saying--
It's consistent with his blood. It's his shoe. He's lying there in his own blood. Four feet away is Nicole Brown Simpson in a pool of her own blood. After we all agree a relatively savage attack, it's not unexpected that blood from one victim will be on the other given the proximity of the two. I mean, isn't that a pretty valid assumption that it's probably Nicole Brown Simpson's blood and Ronald Goldman's blood? Isn't that a pretty good assumption?
KEY QUOTEWell, the problem is, what's the likelihood of some sort of coincidental match. Excuse me.
What's--what's truly misleading about this and really pernicious is, Cellmark's report itself--and this is Cellmark's report of December 5th--says that the additional contributor to that stain other than Nicole can't be identified. Ron Goldman cannot either be included or excluded according to Cellmark's standards. But by computing these frequency statistics in this--in the way that Dr. Weir is doing it, it's a way of roping Goldman back into the stain capitalizing on the rarity of Nicole Simpson's genotype. So that--
KEY QUOTEThere's--there's--there's little controversy that item 78 contains DNA from Nicole Brown Simpson and the--the pattern, her pattern is extremely rare. The real question is whether Ron Goldman's included too. Under Dr. Weir's method, when you ask what's the probability of both of them being there, you get an extremely low, rare number because both of them includes Nicole. So the--the inference with regard to Goldman is being bootstrapped against the rarity of Nicole Brown Simpson's pattern.
You know, our position is that since Cellmark's report says it's inconclusive as to--as to Goldman, it shouldn't--we shouldn't be engaged in this game at all. But we think this is just a way of rescuing an inconclusive match and trying to make the evidence appear more powerful than it is.
Well, the fact that Ronald Goldman's blood is on Ronald Goldman's shoe, you know, the probative value of that as far as your client is concerned is not compelling. I mean, it would be expected that Ronald Goldman's blood would be on Ronald Goldman's shoe; wouldn't you say?
Given the nature of the manner in which he died. All right. Do you have any other bootstrapping concerns with any of the other samples?
--when Gary Sims testified, he was speculating that--that--that maybe Ronald Goldman was part of stain 25 and one of his alleles was simply not detected. Now, it was speculative testimony, but if--if Prosecution experts are going to engage in that form of speculation with respect to 29, then I think that we would have to--and if Dr. Weir testifies to his numbers, then we'll have to explore the implications for Dr. Weir's numbers--
My recollection though is on 29, I did not allow the inconclusive on the board as to Goldman is my recollection. Is that correct, Mr. Clarke?
Well, my recollection of the testimony is, he testified that he could not tell whether Goldman was there. Either could be or--
All right. But I did not allow it on the board, on the results board is my recollection.
We didn't try to put it up there. It was a double negative. We didn't try to put it up there.
Yeah. So in any case. And the assumptions regarding number of contributors also involved difficulties because there's no scientific basis for determining whether a particular stain has two or three if it has more than one, for example, unless you--unless you draw conclusions from other evidence in the case. Now, the Prosecutors are--for some of the stains, are presenting assumptions of two and three. They don't present any assumption that there could be four. But look at stain 29. Under their theory of the case, there could well be four. They seem to be assuming it's--it's--it's--so--you know, so their approach requires this plethora of numbers. If you'll look at Dr. Weir's report on page 38 when we're talking about item 29, the jury with respect to the assumption of two contributors, assuming all the alleles of all contributors can be seen, which we may dispute--but under that assumption, I see one, two, three, four, five, six, seven, eight, nine, ten different numbers which range from 1 in 71 to 1 in 11,306, which is--seems a broad range and confusing. But that's not the end of it. Then under the assumption of trios or three contributors, we have a list of--a very long list, it looks about 15 numbers, ranging from 1 in 9 to 1 in 3500. And of course, then we get into the confusion factor of how the jurors are going to intergrade these different estimates together if they don't--if they're not sure how many contributors it would be. So we have--we have the 11 different numbers under the assumption of two contributors and the 15 different numbers under the assumption of three contributors, and maybe if we had to compute it under four contributors, there would even be more numbers and they would be all over the place, and the jurors have to put it together and try to understand it. It's very confusing. And I think the National Research Council report is much simpler and clearer and easily to--more easily understood intuitively. Now, if the Prosecutors want to make the point that the probability of getting a match with two people would be higher than the probability--than the percentage of people who would match, that's fine. But that's an argument, your Honor. That should not be presented to the jury--that kind of argument should not be presented to the jury in the guise of expert testimony.
But then you would--then Mr. Clarke would be testifying during the course of his argument as to the math behind it. That's the alternative if I accept your argument.
He simply makes the argument that--he simply makes the argument that--that the percentage of the population that would be included in the mixture is 50 percent. But take into account, we've got two people included in here, for example, who both--you know, who both would have to be included. And that--that's an argument. It's--it's--to--to allow the expert to come in and state that as a conclusion based upon factual assumptions that haven't been demonstrated by the evidence is--is--is argumentative in its form and it would present to the jury the impression that these conclusions have been proved scientifically when in fact they're premised on assumptions being made about the nature of the evidence. And, you know, and of course the Prosecution in its presentation controls which assumptions are chosen and which assumptions are not chosen, giving the Prosecutors the option of influencing the nature of the scientific testimony by the way they frame the questions.
None of that can happen under the National Research Council's because it makes no assumptions about these--these crucial factors. So it's simple, it's clear, it's conservative, and that's the way we--we should go. There's one other point about--about the bootstrapping issue, and that comes from Dr. Weir's tendency to slip into the language of likelihood ratios. And although the Prosecutors are--say they're not offering likelihood ratios, I heard Dr. Weir repeatedly talk about the frequencies as if they were likelihood ratios and even make the statement that these statements can be used as likelihood ratios and likelihood ratios for testing the hypotheses that he frames as contact versus no contact. The problem with that formulation--and we certainly would object to any such testimony. But to the extent Dr. Weir testifies in that mode, he is--he is again invading the province of the jury and engaging in argumentation because the implication is that the genetic evidence and the frequencies derived from it are in themselves a characterization of the likelihood of the two hypotheses, contact, no contact, when of course, as Dr. Shields pointed out, those two hypotheses, the relative likelihood depends on a number of other factors.
But my understanding of Mr. Clarke's offer was that it was merely going to be what is the frequency of these patterns being in pairs or in trios without any contact, no contact assumption. Just what is the frequency, what is the likelihood that we will--not likelihood. Bad choice of words in this context. But the frequency of the pattern that is shown here without any reference to the known types, without any reference to the other known facts of the case. That's my understanding of how it's going to be presented.
Yeah. Well, my point is only that as we heard when Dr. Weir testified, there's some slippage of language into that mode, and I think that needs to be made clear.
One other point that Mr. Scheck brings to my attention; is that even doing it under the approach you've just outlined without no assumption, there's still a very broad range of numbers here the jury has to--
Oh, I agree. There's huge numbers. But then we've got huge numbers already, don't we?
Yeah. Yeah. But--yeah. When we have 25 to 35 different numbers which range from 1 in--1 in 9 to 1 in 11,000, what is the jury going to make of that? That's--that's difficult to understand.
But don't we have to assume that the jurors have some basic fundamental intelligence that they can evaluate what this is worth? If you have a statistic that's 1 in 9 to 1 in 3500 and it's a mixture and given the facts of this case, aren't they able to say yeah, it's probably a mixture?
KEY QUOTEWell--but the National Research Council's method gives a way of characterizing it statistically. Doesn't depend on assumptions and so on.
Now, let me finally address this point about the general acceptance of the technique. I think it's very clear the National Research Council produced this report after a two-year study by a very distinguished scientific panel appointed for that purpose. The panel included a wide variety of scientific opinion. It was viewed by some as stacked in favor of the Prosecution because it had four or five people who had testified for the Prosecution on these issues, only one person that testified for the Defense. Nevertheless, the National Research Council, after study, issued this document. It was peer reviewed both within the National Research Council and the national academy itself and by outside reviewers. So that's an authoritative statement on the--on the issue. And I think that it's--it's--it's apparent from the testimony that the National Research Council's position is clear as to what they're recommending. I think a good--and although Dr. Weir claims it's not clear, this is a recent opinion he arrived at. In his earlier reports, he was criticizing--he was criticizing the National Research Council for taking the very position that in his most recent report, he claims that it's unclear that they're--that they're taking. So I think he--Dr. Weir knows what the National Research Council means. I think it's clear that it's unambiguous.
Now, in contrast to the National Research Council's report, which is clear and--clear and unambiguous recommendation from a distinguished panel, the Prosecution is offering a--a novel approach which, as far as we can tell, hasn't been used in any other case. It's based on an article published in England by E.N. Evett, a distinguished man. But Evett's article itself is an article describing the computation of likelihood ratios, which isn't what's being offered here. And Evett's article describes--describes how likelihood ratios could be computed so that forensic scientists can talk about the value of the evidence. Evett's article assumes that the number of contributors is known with certainty from other data in the case, an assumption that we cannot make in this case. And furthermore--I think this is the most telling--the--the lead author of the article, E.N. Evett himself, although he's using likelihood ratios to communicate with scientific colleagues, does not himself advocate that statistics be used to characterize DNA evidence. He thinks they're too confusing and that experts should simply say it's strong or weak and so on.
So--so I don't see that the Evett article provides any kind of powerful or authoritative support for the statistical methodology that's being offered here. It's a novel methodology. There is no peer-reviewed publication that supports its--its use and, you know, and I think it's--it's dangerous and has the potential to be misleading. Now, there's one final point that I'd like to make, you know, concerns--concerns what would be required in the event that you decide to admit Dr. Weir's statistics which are conditioned on all of these assumptions. If that happens, I think it's very clear under California law that the Defense would be entitled to have a--a 403 style instruction to the jury that they are to disregard any computations based on assumptions that they view as unproven. And as authority for that proposition, I'd like to cite to the case of Everett v. Everett, it's the Chad Everett paternity case. It's at 150 Cal. App. 1053.
Well, must be Cal. App. 3D, 1984 case. Okay. And this case includes the following language. It says: "When the relevance of evidence before the jury depends on the validity of an underlying assumption, i.e., a 50 percent chance the alleged father and mother had intercourse--" this was a paternity case of course.
Yeah. So it says: "The jury must be instructed to determine whether the assumption is valid and to disregard the evidence if it finds such assumption is invalid. The validity of such an assumption is akin to a finding of the existence of a preliminary fact," citing section 403 of the evidence code--"Where statistical evidence is derived from a formula which relies on certain factual assumptions, the accuracy of those assumptions must be determined by the jury as a preliminary fact before the statistical evidence may be accorded any weight."
Now--so this suggests that in the event that Dr. Weir's computations are--are allowed, we are going to have to have a very complex, structured set of assumptions to the jurors as to what--what they can and can't consider based on what they--they conclude adding another--adding another layer of confusion to the issue, which again would not be required under the National Research Council report because that report makes no assumptions about controversial matters.
Yes. The Everett case does include an instruction although it's not directly applicable because the Everett case had to do with assumptions in a paternity case. But, yes, if you look at--if you look in that case, 150 Cal. App. 1069 at note 15, I believe you'll see text of the instructions that were offered by the court in that case.
All right. Just to conclude, the National Research Council's recommendations are simple, straightforward and safe. They don't--they don't incorporate any assumptions that are argumentative and that might invade the province of the jury and potentially be misleading. That's the approach we should use here.
All right. Thank you. Mr. Clarke, let me just invite one comment on an instruction, limiting instruction. Do you have any comment on that issue?
Actually, I'd like to see Everett. My recollection is, it's a prior probability bayesian approach, frankly a likelihood ratio case, which is interesting. But I would like to see it if I could before making comment.
I think it basically says before you consider this evidence, you have to make a finding that the assumptions upon which it is based is true.
I can make one comment initially. The same is true with all the reported frequencies and frankly with much of the evidence in this case. It's always true in a criminal case. But if we may.
It's interesting, your Honor, of course that relies on the 50 percent prior probability that is not a part of the calculations but--
You have to find first that the assumptions upon which this calculation is based are in fact supported by the evidence before you should consider the evidence presented on the frequencies.
In concept, there's no disagreement with that as far as certainly the jury has to be satisfied of a preliminary fact such as that. So I don't think in concept, we have any great difficulty with that.
All right. All right. This debate regarding how mixtures are to be interpreted and presented to a jury or other trier of fact is based upon the NRC 1992 report, specifically the one sentence that appears at the end of paragraph 4 on page 59. This is 185-page report, and it devotes all of one sentence to the analysis of mixtures, although that particular sentence is very clear in what it says I think. I don't find it to be ambiguous. The problem is, I don't find it to be a full discussion of the issue. The NRC report at the time it was issued was a conservative compromise regarding DNA testing and its forensic applications. And although I found it of great guidance in framing the issue here, having listened to the testimony of Dr. Weir and Dr. Shields, I agree that the manner proposed by the Prosecution to present the mixture evidence is appropriate as long as it is made clear that it is based upon the two basic assumptions and that there's no reference to the known samples, that it be presented merely that Dr. Weir has the certain qualifications, that he's analyzed the data with regards to these mixtures based upon the evaluation of the databases, the ranges of frequencies are from A to Z, period, thank you very much. That should take half an hour.
Three contributors, four contributors. I make this finding based upon the testimony both of Dr. Weir and Dr. Shields. Looking at which is the most accurate or the more accurate analysis of the evidence, I find that the analysis offered by Dr. Weir is the more accurate and closest to what the evidence truly presents. And under those restrictions, I'll allow the testimony. All right. We'll take a break for 15--
You said he's compelled to give frequencies on the assumption of two contributors and then other frequencies on the assumption of three contributors and finally on the assumption of four contributors and you're going to require they do all three of those; is that correct?
Oh, I know. But four I think is reasonable, especially on the Bronco. And we need to take 15 minutes to shuttle the jury down.
Also, I'm likely to give an instruction at the conclusion based upon what I told you.
Or some appropriate time when I have time to sit down and think about it some more because I had not thought of the--we had thought about instruction before.
Just as to the four, Dr. Weir would need overnight to do those. And if you could narrow--if we can narrow down as to which samples, that would make it easier also.
All the Bronco--the problem is the Bronco samples because you've clearly got at least three people arguably. So you ought to include a fourth. Let me suggest we do this. Let me suggest we start with the non-Bronco evidence where the range ought to be three, get his qualifications. Let's at least get some time in front of the jury today I would suggest. All right. We'll take 15 and we'll start with the jury. Thank you, counsel.
What's truly misleading about this and really pernicious is, Cellmark's report itself—says that the additional contributor to that stain other than Nicole can't be identified. Ron Goldman cannot either be included or excluded according to Cellmark's standards. But by computing these frequency statistics in the way that Dr. Weir is doing it, it's a way of roping Goldman back into the stain capitalizing on the rarity of Nicole Simpson's genotype.
It's consistent with his blood. It's his shoe. He's lying there in his own blood. Four feet away is Nicole Brown Simpson in a pool of her own blood. After we all agree a relatively savage attack, it's not unexpected that blood from one victim will be on the other given the proximity of the two.
Looking at which is the most accurate or the more accurate analysis of the evidence, I find that the analysis offered by Dr. Weir is the more accurate and closest to what the evidence truly presents. And under those restrictions, I'll allow the testimony.
The Prosecution is changing horses in mid-stream. Having conditioned the jury to expect one kind of numbers, they are now presenting a different kind of numbers that in fact are much less favorable to the Defendant.
But don't we have to assume that the jurors have some basic fundamental intelligence that they can evaluate what this is worth? If you have a statistic that's 1 in 9 to 1 in 3500 and it's a mixture and given the facts of this case, aren't they able to say yeah, it's probably a mixture?