|
The Two-Stage Approach (TSA) for the interpretation of evidence is a statistical methodology used to support the inference of source problem in forensic science. This method is based on first considering the traces with an unknown source and comparing them to control samples from the specific source of interest. If the specific source cannot be excluded (usually with a alpha-level hypothesis test) as the source of the trace objects, the examiner will make an assessment concerning the rate at which alternative sources can be excluded as the source of the traces. The formal statistical development of this approach was first laid out in the late 60's by Parker at the UK Atomic Energy Authority.
In this talk we present a modified TSA which aims to suspend all decisions concerning the source of the traces until the evidence is presented to the decision maker. This new approach instead makes use of ROC curves to summarize the atypicality of the traces for each of the competing models for how the evidence has arisen. We will discuss how this new approach is related to classical and formal Bayesian methods for the presentation and interpretation of scientific forensic evidence.
|
