Recross-examination of William Shields

📄 Recross-examination of William Shields — Thursday, June 22, 1995

Address:

C:\DEPT103\CRIMINAL\1995\JUN\22\RECROSS-EXAMINATION-OF-WILLIAM.DOC

TRIAL

▲ Day 101 of 167

◀ Redirect examination of William Shields Further redirect examination of William Shields ▶

Witness: Dr. William Shields
Examiner: George Clarke
Called by: Defense • Date: Thursday, June 22, 1995 • Utterances: 14

In a brief recross-examination, prosecutor George Clarke gets Dr. William Shields to confirm that the mixture-statistics method used by defense witness Dr. Bruce Weir in the United Kingdom relies on likelihood ratios that account for the known genetic types of people connected to the case. Clarke then highlights that the specific frequency number Weir offered at trial (exhibit 71) did not use this more inclusive likelihood-ratio approach.

# 1 MR. CLARKE:

Just a moment if I might, your Honor.

# 2 (Discussion held off the record between the Deputy District Attorneys.)

# 3 MR. CLARKE:

Just briefly, your Honor.

# 4 THE COURT:

Briefly.

RECROSS-EXAMINATION BY MR. CLARKE

# 5 MR. CLARKE:

Dr. Shields, with regard to mixtures, are you aware of whether or not mixture calculations are made by the methods described by Dr. Weir in the United Kingdom?

# 6 DR. SHIELDS:

I believe they are.

# 7 MR. CLARKE:

And that actually includes likelihood ratios; does it not?

# 8 DR. SHIELDS:

It does. They go so far as to go all the way to a bayesian analysis.

KEY QUOTE

# 9 MR. CLARKE:

And likelihood ratios even incorporate the known types of the individuals involved in the case, don't they?

KEY QUOTE

# 10 DR. SHIELDS:

Yes.

# 11 MR. CLARKE:

Dr. Weir's offered frequencies such as the one in 71 that we discussed don't even incorporate the known types of the individuals who may be associated with the case, do they?

KEY QUOTE

# 12 DR. SHIELDS:

That's correct.

# 13 MR. CLARKE:

Thank you.

# 14 MR. CLARKE:

Nothing further, your Honor.

Temperature

procedural

Key Quotes (4)

George Clarke

And likelihood ratios even incorporate the known types of the individuals involved in the case, don't they?

Sets up the contrast between the more rigorous UK method and the frequency figure Weir actually presented to the jury

Dr. William Shields

It does. They go so far as to go all the way to a bayesian analysis.

Establishes that likelihood ratio/Bayesian methods are a more sophisticated alternative to simple frequency statistics

George Clarke

Dr. Weir's offered frequencies such as the one in 71 that we discussed don't even incorporate the known types of the individuals who may be associated with the case, do they?

Prosecution's closing point — suggesting Weir's numbers understate the strength of the DNA match by omitting known genotypes

Dr. William Shields

That's correct.

Witness concedes the limitation of the defense expert's statistical approach

Evidence (1)

DNA mixture frequency statistic offered by defense witness Dr. Bruce Weir

discussed

Notable Exchanges (1)

George Clarke•Dr. William Shields

Clarke walks Shields through a comparison between UK likelihood-ratio/Bayesian mixture methods and the simpler frequency statistic Dr. Weir presented, getting Shields to agree Weir's number omitted known genotype data.

strategic

Credibility Attacks (1)

⚔ Dr. Bruce Weir

methodological critique via a different witness

Clarke uses Shields' testimony to suggest Weir's mixture frequency (exhibit 71) is statistically less rigorous than likelihood-ratio methods that account for known individuals' genotypes

Objections

None recorded

Proceeding 6519 • 14 utterances • Defense witness

Criminal Trial

Department 103