Technical Documentation and Verification for the Buildings Module in the Visual Sample Plan (VSP) Software

Abstract

Visual Sample Plan (VSP) is an easy-to-use visual and graphic statistically-based software tool being developed by the Pacific Northwest National Laboratory (PNNL) to help determine the appropriate number and location of environmental samples so that environmental decisions can be made with the required confidence. The VSP software, which is available free at http://dqo.pnl.gov/vsp, is a significant aid in developing probability-based sampling designs (number and location of samples and measurements) using the Data Quality Objectives (DQO) planning process developed by the U.S. Environmental Protection Agency (EPA). VSP also has the capability of conducted statistical analyses to provide descriptive statistical summaries of data sets, to test whether data are normally distributed, and to compute upper confidence limits on means. This report is the latest in a series of reports that document the statistical methods used in VSP [Davidson (2001), Gilbert et al. (2001), Gilbert et a l. (2002), and Gilbert et al. (2003)] and the quality assurance (QA) activities conducted by PNNL to verify that VSP computations are correct and accurate. This report focuses on the VSP buildings module that was developed with support from the Department of Homeland Security (DHS), Combating Terrorism Technology Support Office (CTTSO), Technical Support Working Group (TSWG). Section 1.0 provides an introduction and overview of the buildings module, while Section 2.0 describes the statistical computations and methods used. Section 3.0 provides the results of the QA activities. Section 4.0 is the reference list. The QA verification results in Section 3.0 demonstrate that VSP is providing correct and accurate computations for: (1) the number of samples required for the various sampling objectives and design options; (2) all statistical calculations, including descriptive statistics, statistical tests for evaluating if data are normally distributed, and upper confidence limits on the mean, and (3) the graphs of data used to visually evaluate if data are normally distributed