Friday, January 27, 2006

Fingered!

What makes fingerprint evidence interesting is the increasing unlikelihood that we all have different fingerprints. This, at least, is what ordinary old you and I might think. Fingerprints are not genetically determined: identical twins have different fingerprints. While their design falls broadly into a few types, at a finer level of detail they seem to be formed randomly.

There are more human fingers than there ever have been. Often, so-called identifications are made from incomplete prints left at the scene of a crime. Part of the print of one finger, it is said, may be sufficient for an expert to say that it is from the accused. However many fingers there are, there are many more "parts of fingers". Yet they are all, so it is said, sufficient to mark us out individually. This compression of information, in a small area of skin, is truly remarkable. Mathematicians can assert that there are something like 10 to the power of 20 different possible patterns in a small area of finger, so, if we accept that, our suspicions about the uniqueness of fingerprints are needlessly alarmist.

Let us assume, however, that the experts are not complacent and that they search with sustained vigour for people who have identical "parts of fingerprints". This amounts to assuming that the experts, who have made a career in reliance on the theory that we all have fingerprints that are unique, are prepared to risk making a discovery that will undermine their livelihoods. I exaggerate, of course: even if many people on earth had the same fingerprints, the finding of one at a crime scene that was, after allowance was made for the distorting effects that occur in real life, not significantly different from the accused's, would, on Bayesean analysis, be highly probative.

Far from being suspicious, the courts are apparently becoming more relaxed about fingerprint evidence. There is judicial acceptance of a trend away from the basing of fingerprint identifications on the number of specified points of identity as between the accused’s print and the sample found at the crime scene. Instead of showing the court enlarged photographs with arrows specifying points on the prints that the court should accept as demonstrating identity, the fingerprint experts are now permitted to base their opinions on the "quality and quantity of information in the images": R v Carter 19/12/05, CA155/05, a phrase used in a technical paper quoted by the Court at para 75.

Important, in this approach, is peer review of the expert witness’s opinion. This review must, in England and Wales, be carried out by three "independent" examiners.

In Carter the Court did not go so far as to say that three independent examiners are needed as peer reviewers before evidence of fingerprint identity will be admissible in New Zealand. Plainly, there would be difficulties finding people who are "independent", let alone people who are sufficiently experienced to qualify as experts, in a pool of talent that is, presumably, rather small. Nevertheless, the Court did reiterate a requirement that Tipping J had made of expert witnesses in the well-known case of R v Calder 12/4/95, HC Christchurch T154/94 (where the accused was ultimately acquitted of poisoning), that the evidence needed to have a "minimum threshold of reliability" to be admissible.

So, the need for the judge to warn the jury to exercise caution before accepting evidence of fingerprint identification where there are relatively few points of comparison (R v Buisson [1990] 2 NZLR 542 (CA)) has apparently been replaced by a more relaxed approach. In Carter, the expert was unwilling or unable to indicate whether there were any points of comparison identified (para 60), but the Court of Appeal held that no particular direction to the jury was needed on this, because the defence, in its closing, had not suggested to the jury that the expert was wrong. Instead, the defence, when the time came to close its case, had merely suggested that there were innocent explanations for the accused’s fingerprint being at the scene.

Incidentally, the real question on which the expert's evidence is relevant, is whether there are any significant differences between the print found at the scene and the accused's fingerprint. As a prosecution witness, the expert will be expected to say there are none, and it is his explanations for apparent differences that are material.

One might wonder whether, in a case where defence counsel had apparently been discouraged (the judge having ruled the expert’s evidence admissible) from a full assault on the credibility of the expert, and where peer review witnesses were not called, the jury were given a proper opportunity to reject the fingerprint evidence.

This raises the question of when a decision made as a matter of trial tactics can be regarded as an error that caused a loss of a real chance of acquittal (assuming that if there was a doubt about the fingerprint identification then there would have been a doubt about guilt); as the Supreme Court held in Sungsuwan (blogged here, 26 August 2005), it is the effect of the error, not its magnitude, that matters.

Update: for a critique of fingerprint evidence, see Gary Edmond, "Fingerprint Evidence in New Zealand's Courts: the Oversight of Overstatement" (2020) 29 NZULR 1. And further: Gary Edmond, Fingerprint Evidence in New Zealand's Courts: the Limits of Legal "Challenge" (2020) 29 NZULR 347.

No comments: