The EDRM organization recently published a significantly updated version of their guide, Statistical Sampling Applied to Electronic Discovery. Many of the definitions and concepts have been revised so that they can be understood by a much broader audience.
The guide is divided into two main sections – an overview applicable to those who need to have a high level understanding of the sampling process, and an in-depth section for those who have tape on their glasses and know how to use a slide rule (or otherwise aspire to be a geek).
Although the title implies that the guide deals with statistical sampling (which is does), it also discusses judgmental, or intuitive sampling that legal professionals have used for years. It provides examples and examines the pros and cons of using each approach. It reviews both proportional sampling (for when you want to find the percentage of relevant records within a population) and acceptance sampling (for when you want to confirm that no records within a population are relevant). While proportional sampling can be used to confirm the lack of relevant records, acceptance sampling requires significantly less records be sampled.
Sampling forms the cornerstone of every form of technology assisted review. It is an efficient method to validate document review results. This guide will hopefully dispel the myth that sampling is too difficult to consider, and encourage more practitioners to adopt advanced e-discovery processes.