Interpreting Crime Scene Fingerprint Evidence

      In 2002, at Chicago, Illinois, a man was arrested and charged with selling cocaine to an undercover agent. The seized cocaine had been packaged in "brick" form, wrapped in a common, plastic, trash bag and purportedly handled by the defendant. One aspect of the government's case was based primarily on a single, latent, fingerprint. Just days before the trial, Defense Attorney, Edward M. Genson, brought me in to resolve the fingerprint issue.
     The FBI, ergo, the government alleged that the fingerprint they had developed was that of  the defendant's left, middle, finger (F8). When I was shown into Attorney Genson's office, I observed a duplicate of the government's proposed (2' x 3')fingerprint chart evidence on top of some files on the floor. I could see from a point about six feet from the chart, a glaring disparity between the Inked fingerprint chart and the Latent fingerprint chart. On the Inked chart there were two, in-line, ridge endings, situated at the two o'clock position, three ridges out from Chart Point #8. Those ridge endings were not on the  Latent fingerprint chart.
      On June 29, 2005, AUSA Pope withdrew the fingerprint evidence.                  

        If you are a Latent Print Examiner, you know the above statement to be true.  If you are an attorney, police administrator or other non-expert, this may sound unrealistically strict.  Scientists generally concur that a reliability of 95% constitutes scientific certainty. 
      In many areas of science and many professions, such reliability serves society well.  For example, the acceptable error rate for Chicken Sexers (persons who examine the backsides of one-day-old chicks to sort males from females) is between 2.3 and 5 percent when examining 700 to 800 chicks per hour.
       A latent fingerprint or palm print "identification" is the determination that two corresponding areas of friction skin impressions originated from the same person to the exclusion of all others.  This absolute identification decision can carry great weight in police decisions about investigations, and court decisions about guilt or innocence.  Latent Print Examiner errors involving evaluation (whether marks are of value or not) and elimination (marks were not made by a person) are serious, but do not normally carry the same potential for harm as erroneous identifications.  An example of the degree of importance afforded identification decisions can be seen in scoring procedures for friction ridge impression comparison tests.  Universally, even if an exam has hundreds of latent prints and thousands of record (inked) prints, some mistakes involving evaluation or elimination are expected, but a single erroneous identification is automatic failure.  The acceptable finger or palm print identification error rate is zero. (Emphasis Added.)
        Because in any field of human endeavor there will always be errors, quality control and quality assurance standards are necessary to protect society.  Standards for initial "training to competency," case by case quality control, and long term quality assurance including life cycle training and periodic proficiency testing are essential.  Periodic proficiency testing can reveal bad habits developed since initial training to competency.  It may also reveal deteriorating eyesight or other human deficiencies impacting casework. 
        One set of standards is set forth in the Quality Assurance Guidelines of the Scientific Working Group on Friction Ridge Analysis, Study and Technology (SWGFAST).
        Upon detecting erroneous identification decisions (not just clerical errors and not just failing to find a "match") most agencies immediately suspend the examiner from further casework activity.  Remedies to ensure a zero error rate include the following: Return of the examiners to friction ridge identification duties only after retraining and evaluation indicate the experts can operate at zero identification error rate;  Permanent transfer to other duties; Employment termination. (Emphasis Added.)

          SOURCE: http://onin.com-fp-problemidents.html#Zero_tolerance_explanation