Direct examination of Dr. Bruce Weir

📄 Direct examination of Dr. Bruce Weir — Friday, December 6, 1996

Address:

C:\DEPT103\CIVIL\1996\DEC\6\DIRECT-EXAMINATION-OF-DR-BRUCE.DOC

TRIAL

▲ Day 27 of 57

◀ Opening remarks Cross-examination of Dr. Bruce Weir ▶

Witness: Dr. Bruce Weir
Examiner: Tom Lambert
Called by: Plaintiff • Date: Friday, December 6, 1996 • Utterances: 93

Plaintiff's expert Dr. Bruce Weir, a statistics and genetics professor at North Carolina State University, was introduced to the jury and qualified as an expert. He then presented frequency calculations for mixed DNA stains from the case, explaining that items 31, 303, 304, and 305 contained DNA from more than one person, with probabilities ranging from 1-in-6,800 (PCR) to 1-in-4-billion (RFLP) that two random unknown people would produce those profiles.

# 1 MR. LAMBERT:

Plaintiff calls Dr. Bruce Weir, Your Honor.

BRUCE WEIR, called as a witness on behalf of Plaintiffs, was duly sworn and testified as follows:

# 2 THE CLERK:

You do solemnly swear that the testimony you may give in the cause now pending before this court shall be the truth, the whole truth, and nothing but the truth, so help you God?

# 3 DR. BRUCE WEIR:

I do.

# 4 THE CLERK:

Sir, if you would, please state and spell your name.

# 5 DR. BRUCE WEIR:

It's Bruce Weir, W-e-i-r.

DIRECT EXAMINATION BY MR. LAMBERT:

# 6 Q:

Good morning, Dr. Weir.

# 7 A:

Good morning.

# 8 Q:

Would you please tell the jury what your occupation is, sir.

# 9 A:

I'm a -- a professor of statistics at North Carolina State University.

# 10 Q:

Could you briefly state for the jury, your formal educational background, starting with your graduation from college.

# 11 A:

I have an undergraduate degree taken in New Zealand in the subject of mathematics. I did a Ph.D. in this country at -- at North Carolina State University in statistics and genetics.

# 12 Q:

What year did you receive your Ph.D.?

# 13 A:

1968.

# 14 Q:

And after obtaining your Ph.D. in statistics and genetics, did you do any postdoctoral study?

# 15 A:

I had a one-year postdoctoral year out of U. C. Davis in the field of genetics.

# 16 Q:

And since that time, have you been affiliated with a university in a teaching or study capacity?

# 17 A:

Yes. I think I -- since 1970, I've been on the faculty of two different universities, teaching and conducting research in the area of statistics, as applied to genetics.

# 18 Q:

What was what the first university, doctor?

# 19 A:

It was a university that was back up in New Zealand.

# 20 Q:

At some point, you joined the faculty at North Carolina State?

# 21 A:

Yes. That was in 1976.

# 22 Q:

What was your position there at North Carolina State, initially?

# 23 A:

At this time, I was an associate professor of statistics.

# 24 Q:

And what is your position now?

# 25 A:

I -- I'm now a full professor of both statistics and genetics.

# 26 Q:

Is that a tenured position at North Carolina State?

# 27 A:

Yes, it is.

# 28 Q:

Do you also hold any chairs at North Carolina State University?

# 29 A:

I -- I have a title; it's called the William Neil Reynolds Professorship.

# 30 Q:

What is the duration of that professorship?

# 31 A:

It's a lifetime appointment.

# 32 Q:

Let me show you, Dr. Weir, a copy of your curriculum vitae, which is Exhibit 435, Your Honor.

# 33 (The instrument herein referred to as Curriculum Vitae of Dr. Bruce Weir was marked for identification as Plaintiffs' Exhibit No. 435.)

# 34 A:

Thank you.

# 35 Q:

(BY MR. LAMBERT) Dr. Weir, is this a current copy of your curriculum vitae?

# 36 A:

Yes, it is.

# 37 Q:

Does this list, at least, the preponderance of the articles that you have -- you have had published, Dr. Weir?

# 38 A:

Yes, I believe so. It goes through earlier to 1996.

# 39 Q:

It lists at least 106 articles that you've either authored or coauthored?

# 40 A:

That's correct.

# 41 Q:

And a number of commentaries and several books, as well?

# 42 A:

That's right, yes.

# 43 Q:

And what is the field in which all of those articles and books have been written?

# 44 A:

They are all in the same field, and they talk about applying statistical methods to analyze genetic data.

# 45 Q:

Are some of those books and articles also dealing with the subject of the forensic uses of DNA and statistical data?

# 46 A:

Yes. For about the past four years, a lot my publications have been in the area specifically to deal with the interpretation of a DNA profile.

# 47 Q:

And have you ever testified on that subject in court?

# 48 A:

Yes, I have. About 20 times, I think.

# 49 Q:

Have you been qualified as an expert on those occasions?

# 50 A:

I have.

# 51 Q:

And have any of those occasions been in courts in California?

# 52 A:

Yes. I think three or four times.

# 53 Q:

Let me now show you now, Dr. Weir --

# 54 MR. LAMBERT:

We'll put on the Elmo, as well, and this will be the next exhibit in order, Your Honor which is?

# 55 THE CLERK:

2231.

# 56 MR. LAMBERT:

2231.

# 57 (The instrument herein referred to as Document entitled Summary of Mixture Frequencies was marked for identification as Plaintiffs' Exhibit No. 2231.)

# 58 A:

Thank you.

# 59 (Witness is handed document by counsel.)

# 60 (Exhibit 2231 is displayed on Elmo Screen.)

# 61 Q:

(BY MR. LEONARD) Would you please, Dr. Weir, describe what this document is?

# 62 A:

Well, overall, it's a series of numbers that helps us to assess the evidentiary strengths of some of the stains in this case, particularly those stains where there are DNA from more than one person. So we call these -- it's so called mixed stains.

# 63 Q:

So the stains that are listed on the document which are listed on the left-hand side, item 31, items 303, 304, and 305, and the various nine items, those are all mixed stains, Dr. Weir?

# 64 A:

That's right. All those stains contain DNA for more than one person.

# 65 Q:

And you've done a -- have you done a frequency calculation in regard to those stains?

# 66 A:

Yes, I have. These -- these are numbers I displayed that tell us the probability of finding those particular types of mixed stains if they came from two unknown people.

# 67 Q:

So let's take, for example, item number 31, the top item, where the Department of Justice obtained PCR results, indicating as a possible source of that item, Mr. Simpson and Mr. Goldman.

What does your calculation in the right-hand column tell us about that particular item?

# 68 A:

That 6800 number tells us if we went into the population and took from the population two people and typed them, we would find that they had the same PCR profile as those in item 31.

You've got the 6800 plus the 19,000, and that's a fairly wide range that reflects the fact that I used different data bases corresponding, broadly speaking, to a different racial group of these two unknown people.

# 69 Q:

Taking the next item, 303, 304 and 305, where the Department of Justice obtained RFLP results indicating possible sources of that stain, that being Mr. Simpson and Mr. Goldman, could you explain your frequency numbers there, please.

# 70 A:

Yes.

Both the numbers are a great deal bigger in this case because the RFLP profile, I believe, was based on four different probes. So it's quite an extensive profile.

And in order to get that particular set of alleles, that set of bands, and no other bands from two people taken at random is a fairly rare event. And I've estimated that to be somewhere between 1 and 800 million and 1 and 4 billion.

KEY QUOTE

# 71 Q:

And the other calculations on this exhibit are all of the same sort that you've described for 31 and 303?

# 72 A:

That's correct. They all talk about the probability of seeing the stain from two unknown people.

# 73 Q:

Now, in doing this calculation, the frequency calculation that you've done, you've calculated for two possible unknown persons being the contributor to the stain.

Is it possible that three or four persons could have been the contributors to those stains?

# 74 A:

Well, it's certainly possible. We can't tell simply by looking at the profile how many people are represented in the DNA.

We can tell by looking at the profile, because there were never more than four alleles or four bands for each probe. That suggests that there were any two contributors. But certainly there could be three or four or even a bigger number, I suppose.

# 75 Q:

And why did you choose to just calculate on this list for two persons, as opposed to more?

# 76 A:

For simplicity, mainly. That is, I think, the easiest and the most -- the most obvious explanation of these profiles if they did not come from the people named as the possible sources.

# 77 Q:

And if you were to calculate the number of unknown contributors and frequency of unknown contributors, three or four were considered, in what way would the numbers change?

# 78 A:

Generally speaking, these numbers become more extreme as the number of contributors go up. It becomes increasingly difficult to have a large number of people with only this particular set of alleles.

Now, there are exceptions, of course, to all general rules. And the in the cases where there are fewer than four alleles, some of the RFLP only had three bands. And then we have to do some very conservative calculations to allow for there being an unseen band. So in those cases, the numbers could go the other direction.

KEY QUOTE

# 79 Q:

In this presentation you've made, you believe this is the most presentation of the frequencies for these particular stains?

# 80 A:

I think it's the one consistent with the fact that there were never more than four alleles at each probe.

# 81 Q:

Now, these numbers that you've given us, these frequency numbers, are they intended to be precise numbers, Dr. Weir?

# 82 A:

Certainly not. These are estimates.

I'm a statistician. I cannot simply give you a single number. You've always got to accompany this by some indication of how much reliance we can put on these numbers.

# 83 Q:

And what -- how would you accompany these particular numbers to show the degree of reliance we can put upon them?

# 84 A:

Well, the bottom line is that I would, particularly for the more extreme numbers, both provide, and divide, and multiply by a factor of 10 to the 800 million. That's the best estimate from the data we have available.

But I must say we should accompany that 800 million by a range anywhere from 80 million to 8 billion, just in the same way when we read public-opinion surveys in the newspaper, we see 47 percent support the president, plus or minus 3 percentage points. The plus or minus in these calculations is a lot bigger than 3 percent; it's about a factor of 10.

# 85 Q:

For example, on the RFLP results for 303, 304 and 305, if you applied the factor of 10 to both the highest and lowest frequency, another lowest would -- could become 89 -- lowest would be 89 million and the highest would be 40 billion?

# 86 A:

Yes. If the low end of the smallest number is 89 million and biggest end of the largest number is 40 billion.

# 87 Q:

The frequencies you put up here, your best estimate of those frequencies based upon the available data?

# 88 A:

Yes. I believe they are the appropriate numbers to report.

# 89 MR. LAMBERT:

Your Honor, I'd move into evidence Exhibit 435 and this new exhibit, number 2231.

# 90 THE COURT:

Okay.

# 91 (The instrument previously marked as Plaintiffs' Exhibit 435 was received in evidence.)

# 92 (The instrument previously marked as Plaintiffs' Exhibit 2231 was received in evidence.)

# 93 MR. LAMBERT:

I have no further questions, Your Honor.

Thank you.

Temperature

procedural

Key Quotes (3)

Bruce Weir

I'm a statistician. I cannot simply give you a single number. You've always got to accompany this by some indication of how much reliance we can put on these numbers.

Establishes Weir's credibility by acknowledging uncertainty — a statistically honest framing that actually strengthens rather than weakens his testimony.

Bruce Weir

Both the numbers are a great deal bigger in this case because the RFLP profile, I believe, was based on four different probes. So it's quite an extensive profile. And in order to get that particular set of alleles, that set of bands, and no other bands from two people taken at random is a fairly rare event. And I've estimated that to be somewhere between 1 and 800 million and 1 and 4 billion.

Core statistical finding — places the odds of a random match at up to 1 in 4 billion for the key mixed stains tied to Simpson and Goldman.

Bruce Weir

Generally speaking, these numbers become more extreme as the number of contributors go up. It becomes increasingly difficult to have a large number of people with only this particular set of alleles.

Preemptively addresses a potential defense argument that more contributors could explain the stain — Weir turns it against that theory.

Evidence (2)

Plaintiffs' 435

Curriculum vitae of Dr. Bruce Weir

introduced and admitted into evidence

Plaintiffs' 2231

Document entitled 'Summary of Mixture Frequencies' — statistical frequency calculations for mixed DNA stains from items 31, 303, 304, and 305

introduced, displayed on Elmo screen, admitted into evidence

Notable Exchanges (2)

Tom Lambert•Bruce Weir

Lambert walks Weir through the frequency table, having him explain the difference between PCR results (1-in-6,800) and RFLP results (1-in-800-million to 1-in-4-billion) for stains listing Simpson and Goldman as possible contributors.

strategic

Tom Lambert•Bruce Weir

Lambert elicits testimony that using two unknown contributors is the conservative, simple choice — and that assuming more contributors would only make the numbers more extreme against the defense.

strategic

Objections

None recorded

Proceeding 8504 • 93 utterances • Plaintiff witness

Civil Trial

Department 103