A recent U.K. decision held that DNA evidence, without corroborating evidence, was not sufficient evidence to convict under the particular circumstances of the case, heard in the Court Of Appeal (Criminal Division) : R. v. Watter, October 19, 2000.
In Watter, the judge conceded that if the appellant had two brothers the probabilities involved would reduce to 1 in 267 and 1 in 32 respectively. She continued, based on these probabilities, that DNA evidence in itself was not proof.
Police forces admitted that they had wrongly secured hundreds of convictions on the basis of DNA testing. The police concluded that, as far back as April of 1999, they had matched a sample taken from the scene of a burglary to six loci on the DNA molecule of one of 700,000 persons whose DNA was collected without consent and stored in the national database. The odds of DNA being wrongly matched against that of the crime scene were said to be one in 37 million.
Was the result an "erroneous" or "false" identification? No.
There was indeed a "match" at six loci. All but the initiated believe that when a DNA "match" is reported with odds of one in 37 million, we will encounter a like match in the DNA pattern only once in 37 million people.
To test whether this is a common misunderstanding, the author of the study, Professor Moenssens, when lecturing asked attendees what was understood by the meaning of the testimony. All who responded said that this defendant's particular DNA pattern would occur only once in 37 million individuals. But this is blatantly untrue.
DNA scientist Keith Inman, co-author with Norah Rudin of the recently published treatises An Introduction to Forensic DNA Analysis (CRC Press, 1997) and Principles and Practice of Criminalistics (CRC Press, 2000), said that the calculated frequency is an estimate, and can be off by an order of magnitude in either direction and that
"studies show that when databases grow, more loci (more discriminating loci) are required to support a strong inference of a common source."
In other words, despite the statistical calculation of 1 in 37 million on six loci, that does NOT mean that the six loci cannot match more than one person in 37 million. According to population geneticists, it is indeed possible to have the six loci match in perhaps many dozens of individuals whose DNA is contained in a databank of 700,000.
In short, DNA profiling is creating a society of criminals where none actually exist.
* Karen Clark Stapleton is an ex-Police and Prison Officer who is currently an MSc Student in Criminal Justice Research.