evidence enters a courtroom with an authority that its scientific basis does not always support.
Several techniques accepted for decades were never before they were used to convict.
Bite mark comparison, hair microscopy and several pattern disciplines fall into this category.
Reviews found that no court had been told about and that no defence could have known to ask for.
Even techniques with a sound foundation are presented in ways the evidence cannot carry.
A between two samples is a statement about probability and is routinely delivered as a statement about identity.
The distinction is not pedantic, since a jury that hears identity has been given a conclusion rather than evidence.
compounds the problem in a way that surprised the profession itself.
Analysts who knew the case background reached different conclusions on the same prints than analysts who did not.
, which releases information to the examiner only as it becomes necessary, removes most of that effect.
It costs nothing and has been adopted slowly, because it implies that the previous method was unsound.
Laboratories that report to a police force face a structural conflict that no individual integrity can resolve.
Independent and blind proficiency testing address it, and both require money that is easier to spend on prosecutions.
The defence side of the problem is rarely discussed and is arguably larger.
Challenging expert evidence requires an expert, and legal aid rates in most systems do not fund one.
An system that supplies expertise to one side only is in name.
Reform has concentrated on standards for the laboratory, which is the visible half of the problem.
The other half sits in a legal aid budget and attracts no scientific attention at all.