You are here:
PERFORMANCE OF STATISTICAL TESTS FOR SITE VERSUS BACKGROUND SOIL COMPARISONS WHEN DISTRIBUTIONAL ASSUMPTIONS ARE NOT MET
Citation:
ENGLUND, E. J. PERFORMANCE OF STATISTICAL TESTS FOR SITE VERSUS BACKGROUND SOIL COMPARISONS WHEN DISTRIBUTIONAL ASSUMPTIONS ARE NOT MET. U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-07/020 (NTIS PB2007-107708).
Impact/Purpose:
The overall objectives of this task are to: 1) provide ORD state-of-the-science technical support and assistance to Regional staff; 2) facilitate the evaluation and application of site characterization technologies at Superfund and RCRA sites; and 3) improve communication among Regions and ORD laboratories.
Description:
Statistical distributions of site and background soil samples often do not meet the assumptions of statistical tests. This is true even of non-parametric tests. This paper evaluates several statistical tests over a variety of cases involving realistic population distribution scenarios and sampling schemes. Over the range of cases, performance was erratic for most tests. When planning a project, the sampling scheme must be designed together with the statistical test, and the choice of test may vary depending on which scenario best matches the conceptual model for the site.