GUIDANCE FOR DATA QUALITY ASSESSMENT Practical Methods for Data Analysis

FOREWORD This is the 1996 (QA96) version of Guidance for Data Quality Assessment, EPA QA/G-9. The Environmental Protection Agency (EPA) has developed the Data Quality Assessment (DQA) Process as an important tool for project managers and planners to determine whether the type, quantity, and quality of data needed to support Agency decisions has been achieved. This guidance is the culmination of experiences in the design and statistical analyses of environmental data in different Program Offices at the EPA. Many elements of prior guidance, statistics, and scientific planning have been incorporated into this document. This document provides general guidance to organizations on assessing data quality criteria and performance specifications for decision making. This guidance assumes that an appropriate Quality System has been established and that planning for data collection has been achieved using a scientifically-based information collection strategy. An overview of the Agency's recommended data collection procedure, the DQO Process, is included in this guidance in Chapter 1 and EPA QA/G-4. Guidance for Data Quality Assessment is distinctly different from other guidance documents; it is not intended to be read in a linear or continuous fashion. The intent of the document is for it to be used as a "tool-box" of useful techniques in assessing the quality of data. The overall structure of the document will enable the analyst to investigate many different problems using a systematic methodology. The methodology consists of five steps that should be iterated between them as necessary: (i) Review the Data Quality Objectives (ii) Conduct a Preliminary Data Review (iii) Select the Statistical Test (iv) Verify the Assumptions of the Test (v) Draw Conclusions From the Data This approach closely parallels the activities of a statistician analyzing a data set for the first time. The five step procedure is not intended to be a definitive analysis of a project or problem, but provide an initial assessment on the "reasonableness" of the data that have been generated. Sophisticated statistical analysis is often not necessary unless special or unusual circumstances have been encountered in the generation or collection of the data or the analysis is planned in detail before the data are collected. This guidance is directed towards the analysis of relatively small data sets containing data that have been collected in a relatively simple fashion. The analysis of survey data containing large data sets or a complex sampling scheme is best left for statistical experts. This document …