A Bayesian Approach to EPA's Data Quality Objectives Process

• Activity Number: 231
• Type: Invited
• Date/Time: Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Paul K. Black*+ and Mark Fitzgerald and Tom Stockton
• Companies: Neptune and Company, Inc. and Neptune and Company, Inc. and Neptune and Company, Inc.
• Keywords: DQOs ; Bayesian methods ; Scientific Method ; statistical planning

Abstract

In order to improve the quality of its decisions, the Environmental Protection Agency recommends the use of the Data Quality Objectives (DQO) Process. The intent of the DQO process is to offer a systematic approach to planning for data collection consistent with the goals of the Scientific Method. EPA has focused on the use of classical statistical methods to implement the implied decision analysis. Arguably, the Scientific Method is best implemented using Bayesian methods. When the DQO process was first introduced, computer programs for implementing a Bayesian approach were not available. With the advent of new algorithms for Bayesian analysis, this is no longer an issue. We will present a contrast between the two approaches, demonstrate the benefits of using a Bayesian approach to DQOs and, more generally, to decisionmaking, and provide some real world examples of its implementation.