8127/frame/fm

 
e.
Combining Source Determination and Inference of Contact
We have outlined two processes in forensic science that require inference —
source determination and contact. We are left with the question of how we
are to incorporate the inference of common source into our inference about
contact. Because of the constant confusion in the literature about association
and interpretation, this issue has not been addressed with clarity. Clearly, a
Bayesian framework provides the most explicit and quantitative approach.
However, perhaps because the process is so case specific, no general sugges-
tions or guidelines have been set forth in the literature about how to do this.
We challenge the field to consider this problem in applying the results of a
physical evidence analysis to a case.
 
f.
O. J. and Inference — Is the Tube Half Full or Half Empty?
The 
 
Simpson
case provides an interesting example of inferential thinking. One
noteworthy example (math not required) of LRs in (intuitive) action is found
in the criminal trial of O. J. Simpson. In that case the formulation of compet-
ing hypotheses was crucial to the verdict. The prosecution hypothesis was that
the defendant had committed a double-homicide, and when considering the
physical evidence, this translated into the hypothesis that Simpson was present
at the crime scene. Multiple biological evidence samples were collected from
at least three different locations (Rockingham, Bundy, the Bronco). DNA
analysis provided overwhelming correspondence frequencies in favor of the
prosecution hypothesis. The alternate hypothesis posed by the defense was
not merely that Simpson was not present, but that the police had planted
evidence from which we could would infer his presence at the various loca-
tions. The competing hypotheses are translated into an LR as follows:
LR
P E
Simpson present
(
)
P E
Blood planted
(
)
------------------------------------------------------
1
=
=
8127/frame/ch07 Page 175 Friday, July 21, 2000 11:46 AM

---

176
Principles and Practice of Criminalistics
Under this set of hypotheses, the LR = 1, because we would expect to find
the defendant’s blood in either case. The jury could then ignore the DNA
evidence, because it did not provide compelling evidence in favor of either
proposition. This left the determination of the probability of his presence at
the scene dependent wholly on the prior odds. In this case, the prior odds
are directly concerned with the juror’s belief that the police were both capable
and willing to plant evidence, compared with their belief that the defendant
would murder two people.
The defense then had to demonstrate that the opportunity existed for
the planting to take place, and also had to rely on the jurors’ subjective prior
belief in the proposition that the police were willing to plant evidence. The
ratio of these beliefs would have to rise above the standard of “beyond a
reasonable doubt” in favor of the prosecution hypothesis in order to convict.
The verdict of “not guilty” says that their evaluation of the prior odds of the
police planting blood was very high. In mathematical notation:
∴
In other words, the results of the physical evidence were essentially not
considered because the jurors believed they had no relevance to the case.
Thus, the prior odds completely determined the posterior odds. This reiter-
ates the point that physical evidence merely updates our belief in a proposi-
tion; it does not substitute for the belief. Different assumptions or facts will
change our acceptance of an inference.