The
Y7/Shirley McKie case – logical and statistical aspects
I
am a computer programmer working in the field of statistics for industry. I
suspect that the confident beliefs that Shirley McKie lied about entering the
murder scene were based on an error of logic.
The
case is extremely unusual because there is no reason, other than the
fingerprint, to believe that any wrongdoing occurred - there is no independent
evidence of a crime. This is a major and significant abnormality, the case
might even be unique in the history of forensic evidence presented in a court
of law. I suspect the significance of the abnormality was not taken into
account when the evidence against Shirley McKie was evaluated.
Known
fingerprint errors are very rare but it does not follow that we can assume that
every identification will carry only a very small risk of being wrong (to do so
would be a serious error of logic known as the Prosecutor’s Fallacy). After a
suspect has disputed forensic evidence the likelihood of error in that case is
tied (inversely) to the likelihood of the suspect’s guilt. This can only be
assessed by balancing the likelihoods of the different explanations for the
facts of the case.
While
fingerprint errors are rare it is not rare for perpetrators of crimes to leave
fingerprints while committing them. So if a fingerprint from a location
connected with a crime plausibly incriminates somebody in that crime, and they
deny it, this is much more likely to be the perpetrator lying than an innocent
person telling the truth due to a rare error.
When
a single fingerprint is the only reason to believe that wrongdoing has
occurred, as well as suggesting who did it, the balance of likelihoods is not
so clear. Misidentifications are rare but rare events happen. Do we have a valid
reason to believe that we are not simply reacting to a rare error?
A
police officer making an unauthorised entry to a secured and guarded murder
scene, who would deny it if told of fingerprint evidence, would surely be a
rare thing and be very unlikely if there was no reason to suspect it had
happened. Committing perjury over such a thing will be even more rare and
unlikely. So when evaluating the disputed fingerprint evidence in the Shirley
McKie case, taking the context into account, the likelihood of one rare and
unlikely explanation for the facts should be compared and balanced with the
likelihood of another rare and unlikely explanation. The result of balancing
two similar likelihoods is uncertainty.
If
this had been recognised in 1997, suspicions that Shirley McKie was lying might
have been dropped.
Steve Horn.
West Lothian
It struck me while watching news clips of Scottish Parliament inquiry into the Shirley McKie case that if, generally speaking, fingerprint identifications carry a small risk of error, then a confident belief that McKie deposited Y7 could be a case of The Texas Sharpshooter Fallacy.
Fingerprinting
Innocentville is a way of looking at the logic using a few simple “back of
envelope” type calculations.
When looking at expert
evidence we have to be aware of Transposing the
Conditional
Since
1997 the McKie case has been repeatedly studied and investigated by police
officers, a procurator fiscal (prosecutor), judges, forensic experts and, in
2006, by a committee of the Scottish Parliament. In all this time nobody has
thought to question the evaluation of the evidence that gave rise to the case.
LINKS:
I think that the principle described here is important and should be clearly understood by justice professionals. I have joined discussions in the CLPEX forum for fingerprint examiners and the linked pages below are copies of some of my contributions.
From a logical point of view using
a denied elimination identification as a means of detecting a hidden act of
wrongdoing has similarities to using mass medical
screening to detect hidden diseases. In mass screening a medical test can
be very accurate yet someone who tests positive will probably not have the
disease.
There are logical
similarities with the Sally Clark case (the Royal
Statistical Society used the term “serious error of logic” in connection with
that case).
More about relative
likelihood and the McKie case.
An exercise to see how varying the error rate alters the likelihood of guilt
produced some interesting results.
OTHER LINKS:
Facts
about the Shirley McKie case can be found here, here and here.
Here is a page on reasoning
about evidence in Court by Norman Fenton, Professor of Computer Science at
Queen Mary University, London.
Link to Scottish Parliament report
into the McKie case: http://www.scottish.parliament.uk/business/committees/justice1/reports-07/j1r07-03-vol1-00.htm
Read the Crown Prosecution Service
warnings about the Prosecutor’s
Fallacy regarding matching of DNA profiles.
Please contact me if you think there are any errors in my reasoning (or post on the CLPEX forum chat board).