Latent Print Proficiency Testing
One of the issues raised with latent print examination was the need for proficiency testing. I would want to test does an examiners’ work quality differ depending on the context of the examination. This would require testing examiners with them knowing they were being “tested” and then blind testing the examiners. This could be done by adding test prints into their normal case flow. With regards to the studies in the report, I assume the examiners knew they were part of a study. This knowledge could have led them to examine the prints more carefully than they would during regular cases. There are two factors that would need to be controlled: confirmatory bias and contextual bias. Confirmatory bias could be reduced if the verifier (if using the ACE-V method) was not allowed to know the outcome of the initial examination. Also, the examiner should analyze the unknown before looking at the known. To control contextual bias, the examiner should be given limited (if any) information about the facts of the case.
The treatments in this study would be casework vs. a study (i.e. whether the examination was performed during regular case work or during a study). The units would be the examiners.
The responses of the study would be the error rate differences, measured as percentages. The error rate would be calculated based on false positives.
Each examiner could be assigned a number, a number could be chosen at random and assigned to either the control or experimental group. To achieve replication, the same examiners could be re-assessed over a period of time.