Fingerprints on Trial

Science, pseudoscience, and the curious case of Shirley Mckie

Next month, a “Public Judicial Inquiry” - announced by Kenny MacAskill MSP, the Cabinet Secretary for Justice, in the Scottish Parliament on 14 March 2008 – into the case of H.M Advocate v McKie will finally report its findings. (ref)

This report (about a case which seems to have faded into media obscurity) may have profound implications for legal systems (north and south of the border if not worldwide) and potentially for (a topic very much in the news) border control.

It all began back in 1997 when, a police woman, Shirley Mckie, was arrested and charged with perjury over a (left thumb) fingerprint – identified as hers by four experts from the Scottish Criminal Records Office (SCRO) – found on a right door jamb inside a house where a police team were investigating the vicious murder of one Marion Ross in Kilmarnock, Scotland.

The authorities insisted that Shirley must have been present at the crime scene at some stage during the investigation and was now lying about it. She insisted she had never entered the house.

In the end, Shirley Mckie was acquitted of perjury (though the case did go to trial) and (in separate developments) David Asbury – the man jailed for Marion Ross’s murder, also on the basis of a disputed fingerprint[1] – was exonerated and freed from jail.

You can read Shirley Mckie’s version of events here.

Now if the four experts at the SCRO had simply admitted they had made a mistake, that would have been the end of the matter. In fact, they stuck to their guns and have been supported by a number of outside experts. Meanwhile, many more experts – including several international experts - have supported Shirley Mckie and insisted that the fingerprints in question were not hers or Marion Ross’s.

The balance of findings of various court cases and inquiries thus far has tended to support the Mckie camp, but the fact that a number of profound questions remain open and are still disputed by different experts has led to the current Public Judicial Inquiry. The evidence presented to that inquiry is available online and, probably most interestingly, this evidence includes that of fingerprint expert Peter Swann – who was originally consulted by the Mckie family but who is now very much in the opposing camp and who still insists to this day that the identifications of Shirley Mckie's and Marion Ross's prints were unequivocally correct.

I have no idea what the findings of the Judicial Inquiry will be but, whatever they are, profound questions will be raised about the use of fingerprinting.

Let me say at the outset that I know nothing (other than what I have picked up in the course of reading about this case) about the science of finger printing. I do, however, know quite a lot about scientific methodology in general.

The Professor of Logic and Scientific Method Sir Karl Popper is often quoted (and often misunderstood) both by defenders of the “scientific method” (something harder to define than many scientist admit or realize) and by defenders of all sorts of pseudo-scientific nonsense. What is less often pointed out is that many of Popper’s ideas have been found wanting, but one of his insights stands up rather well. Popper realized that one of the key things that distinguishes true science from pseudo-science is that pseudo-scientists look (by and large) only for verification of their theories. True scientists test their theories by looking for falsification.

To give a simple example: it is very easy for an astrologer to go out and find 16 (or more) people born under the sign of Capricorn who are extremely capricious (or whatever Capricorns are supposed to be). The true scientist, however, looks to see if there are any Capricorns who are un-capricious. The finding of just one such person falsifies the astrologer’s claim – regardless of how many people the astrologer has lined up in support of his claim.

Of course claims like this are more often expressed in probabilistic terms: “Capricorns are more likely to be capricious”. The same basic principles apply however. The true scientist starts counting un-capricious Capricorns and capricious non-Capricorns and performs a statistical analysis. The astrologer simply goes on saying: “look at all these people who are Capricorns and who are extremely capricious”.

So what has all this got to do with finger printing? At the time of the original events, the website of the SCRO used to boast that “in order to comply with the existing standard in Scotland” they insisted on identifying sixteen characteristics, in sequence and agreement, before concluding that a match had been found (unfortunately I have no cache of this site but I assume that the exact wording will be available somewhere in the inquiry evidence).

I should be interested to know how widely this “standard” is (or was) used in the world at large – my understanding is that the former SCRO (now, I understand, part of the Scottish Police Services Authority Forensic Service) now use a “non-numeric” standard.

The potential fallacy implicit in the “sixteen points in agreement” approach (to anyone who has read Popper) is plain to see. Let us assume (as fingerprint scientist claim) that every fingerprint in the world really is unique (an un-testable hypothesis, but one that there are good reasons for accepting). It does not follow at all from this that every fingerprint in the world is unique with respect to sixteen (or however many) carefully selected characteristics. After all, I am pretty confident that (just considering the topology) I could find at least sixteen “points of agreement” between the London and Paris underground maps if I set my mind to it.

No matter how many points of agreement there are between two fingerprints – 16, 45, or 145 – logically speaking, one single point of disagreement clinches the matter: the two prints are not the same.

Of course, in the real world, things are never quite so clear cut. Points of agreement and disagreement will be a matter of judgment by experts - often working with less than perfect images – and several points of disagreement may have to be found to really clinch the matter. But the basic principle holds: Any true scientist comparing finger prints will look not just for points that confirm his hypothesis that two prints are the same, but for any points that falsify that hypothesis.

So does this mean that fingerprinting is or was or often is a pseudoscience? I am really not qualified to say – though I can think of research programmes that could clinch the matter and I shall certainly read the coming report with a great deal of interest.

Two alternative conclusions seem hard to avoid however:

either

1) UK fingerprinting was (and possibly still is) riddled with incompetent and/or dishonest and/or deluded “experts”.

or

2) Fingerprinting really is not a true science and different “experts” will honestly reach different judgments when presented with the same evidence.

Again, I have no idea which of these conclusions is correct (perhaps a bit of both?) but I would find either of them profoundly disconcerting.


[1] In this case, the print (on a tin in David Asbury's possession) was identified - by the SCRO and later by Peter Swann - as belonging to the murder victim Marion Ross.

POSTSCRIPT

The report has been published:
summary
full report

The Chairman, Sir Anthony Campbell, would seem to have come down very firmly in support of suggestion "2" above: "Fingerprinting really is not a true science and different 'experts' will honestly reach different judgments when presented with the same evidence."