NCJ Number
192046
Journal
Journal of Forensic Sciences Volume: 46 Issue: 6 Dated: November 2001 Pages: 1456-1461
Date Published
November 2001
Length
6 pages
Annotation
Drug analysts are regularly faced with the problem of how many samples to analyze when a large number of apparently similar items are submitted; this paper describes a Bayesian approach to this problem, following Aitken.
Abstract
Recently, Aitken (1999) introduced an advance in the approach to decision making regarding drug sampling. The Bayesian approach favored by Aitken is based on a solid mathematical foundation. It differs from more classical approaches and is preferred in the context of drug sampling, both because it is more practical, in that it tends to suggest more realistic sample sizes, and because it makes better use of the information available. Implementation of this approach is described in this paper. It involves an explanation of the Bayes' theorem and the hypergeometric distribution. One section of the paper presents a step-by-step implementation of the mathematical equation that is amenable to programming in EXCEL. Three case examples are presented to illustrate the approach. Minor differences exist between the approach described in this paper and that developed by Aitken, both in the modeling of the prior probability and in dealing with the discrete nature of the samples. These differences do not detract from the sound mathematical foundation of the approach. 11 tables and 8 references