ESR John Buckleton ood Forensi S ien e. ESR John Buckleton Good Forensic Science A view from the bottom by John Buckleton With thanks to my great mentors.

John Buckleton ESR

ood Forensi S ien e

John Buckleton ESR

Good Forensic Science

A view from the bottom by John Buckleton

With thanks to my great mentors

Ian Evett and Bruce Weir

John Buckleton ESR

What can the 80’s and 90’s teach us?

In most cases the interpretation “arguments” were never settled.

They were superseded by technology or they remain

Product rule or not? The place of independence testing? Relatives? Two stain at a scene

John Buckleton ESR

Why we have been unable, by science alone, to settle the

interpretation matters of our day?

John Buckleton ESR

WHAT IS THE QUESTION?

John Buckleton ESR

Budowle B. National Forensic Science Training Centre, Florida, 1996.

“The germane question is: What is the frequency

in the database?”

John Buckleton ESR

First ‘germane’. Oxford gives this as: “Relevant, pertinent to the matter or subject”.

It must certainly be agreed that it is important to ask relevant questions so it certainly seems important to define the germane question.

John Buckleton ESR

Since this question is germane let us try answering it.

Imagine that we have a 13 locus STR match. • search the database of say 200 persons (a

plausible size in the USA) for this profile• almost certainly do not find it?

We can now answer the germane question:

John Buckleton ESR

“What is the frequency in the database?”

The frequency in the database is zero

“Hmmmm. What shall I do with that information?”

John Buckleton ESR

The question, if taken literally, is not much use.

If it cannot be taken literally then can the question be improved?

Perhaps the author didn’t mean database

perhaps he meant population.

John Buckleton ESR

“What I meant was, the germane question is: What is the frequency

in the population?”

John Buckleton ESR

Here the last word has been changed from database (where the frequency is zero) to population?

This seems quite a lot better. Armed with this, we approach our

database of 200 persons and attempt to answer this “germane” question.

What can we do next?

John Buckleton ESR

What has happened is that we have changed from database

completely characterized we can count any frequencies to talking about a “population” There is uncertainty about the population. do not KNOW the frequency of the genotype in

the population. may even have difficulty defining the population.

John Buckleton ESR

Accept that we are now estimating something There is uncertainty in the estimation process. We will need, at least: A population genetic model. A statistical theory to assess sampling error. Estimation of something very small is difficult.

John Buckleton ESR

“What I really meant was, What is the product rule estimate

from this database?”

It would be sad if the audience’s positive reaction was because they liked the simplicity of this statement or that they knew how to do it or that they didn’t have to learn anything new.

John Buckleton ESR

Evett IW. (1983) What is the probability that this blood came from that person: A meaningful question?

J Forensic Sci Soc 23:35-59.

“What is the probability that this blood came from that person?”

John Buckleton ESR

The second was set up deliberately to answer a question of relevance to the court.

The first did not start from this formalized line of thinking.

The second is the more “germane” question.

John Buckleton ESR

This is the lesson that can be taken from the analysis of this question:

If you don’t have some firm principles on which to base your forensic thinking you can be left relying on your intuition.

Sometimes this will be adequate but it will seldom stand up in comparison to a professional, scholarly and logical approach.

John Buckleton ESR

Uniqueness?

John Buckleton ESR

Your honor, this profile is unique, to a reasonable

degree of scientific certainty.

“That sounds very objective. I like that.What do you mean when you say

the profile is unique?”

John Buckleton ESR

I mean it has met objective criteria developed by my laboratory.

“What are these criteria?”

John Buckleton ESR

We defined a reasonable degree of scientific certainty by setting an

acceptable error rate for this decision.

“Ouch. That’s really my job”

John Buckleton ESR

We ignored all issues regarding relatedness.

“Why?”

John Buckleton ESR

We assumed a population genetic modelcalled the product rule.

“Ouch. How can you be sure that is correct?”

John Buckleton ESR

We assumed a population genetic modelcalled the product rule.

“Would it be better to accept that there is some uncertainty in the model?”

John Buckleton ESR

I’ve done lots of data analysis and I think it is OK or near enough.

“This sounds more like an opinion than an objective criterion.”

John Buckleton ESR

This is a fallacy.Independence testing cannot prove

the product rule

Independence testing shows that I’m right.

John Buckleton ESR

They are also consistent withminor subdivision

Yes, but the data are consistent with expectation for independence..

John Buckleton ESR

Caucasian Eastern Polynesian Western Polynesian

Pr Pr Pr VWA 0.698 0.297 0.195 THO1 0.395 0.755 0.266 D8 0.706 0.912 0.105 FGA 0.359 0.644 0.913 D21 0.714 0.752 0.709 D18 0.632 0.215 0.129 VWA/THO1 0.505 0.131 0.656 VWA/D8 0.482 0.232 0.424 VWA/FGA 0.112 0.506 0.893 VWA/D21 0.704 0.058 0.949 VWA/D18 0.283 0.715 0.751 THO1/D8 0.111 0.982 0.299 THO1/FGA0.535 0.964 0.324 THO1/D21 0.338 0.711 0.416 THO1/D18 0.057 0.241 0.025 D8/FGA 0.117 0.938 0.858 D8/D21 0.240 0.563 0.456 D8/D18 0.320 0.106 0.210 FGA/D21 0.209 0.668 0.901 FGA/D18 0.028 0.015 0.400 D21/D18 0.115 0.229 0.636

John Buckleton ESR

0

2

4

6

8

10

John Buckleton ESR

Post hoc rationalization Most unwise Budolwe et al. testing the FBI’s HBGG locus, appear

to have combined the A and C alleles. There is a significant departure from

independence if the alleles are not combined, caused largely by a single CC homozygote.

John Buckleton ESR

Post hoc rationalization Result does not appear in the paper Budowle et al. conclude that the data

are “consistent” with Hardy-Weinberg expectations

This is an unreasonable statement considering the (presumably) post hoc combination of two alleles.

John Buckleton ESR

We BELIEVE That the effects of subdivision are

relatively minor They do, however, add across loci

John Buckleton ESR

OK so it’s minor so we can safely ignore it.

It has no “forensic significance.”

“I won’t dare ask what Forensic significance means.

Would it not be more scientific to quantify the effect?”

John Buckleton ESR

Well that makes things quite complex.

“It seems to me more scientific . It is also a principle of law that doubt

should be given to the accused.”

John Buckleton ESR

Why are many countries dropping their point count rules in fingerprints

and accepting that a statement of common source in fingerprints is an expert opinion?

John Buckleton ESR

USA v Byron C Mitchell

Individualisation ..can result from comparisons.…containing a sufficient quality .. and

quantity of unique friction ridge detail

This sounds like expert opinion evidence, especially words like “sufficient quality”

John Buckleton ESR

Has the SCIENTIFIC basis for statements of common source in DNA been really thought

through? Published and accepted?

John Buckleton ESR

Yes you are probably right. But does this meet standards

for scientific testimony?

Not exactly but I’m probably right.

John Buckleton ESR

20 years from now will they call this good science?

John Buckleton ESR

Maybe, but I suspect the difference would be much less if

we had an accepted formal method for scientific decision making in forensic science.

Is it not simply that reasonable people can hold different points of view?

John Buckleton ESR

“What if Bayesian methods can help?”

John Buckleton ESR

When men understand what each other mean, they see for the most part, that controversy is either superfluous or hopeless.