WORK PLANS - Laserfiche WebLink

Home Browse Search

3. F Test to compare two variances (dispersions)—Parametric Test 4. Wilcoxon-Mann-Whitney (WMW) Test to compare two locations, comparability of two continuous distributions—Nonparametric Test 5. Quantile Test to compare the upper tails of two continuous distributions -Nonparametric Test 6. Gehan Test to compare two locations -Nonparametric Test T-tests and F-test assume normality of the data sets under comparison. Some details of these approaches are described in ProUCL 4.0 Technical Guide. It should be noted that Gehan test, WMW test and Quantile test are also available for data sets with NDs. Gehan's test is specifically meant to be used on data sets with multiple detection limits. The Quantile test is a nonparametric test and is useful to detect a shift in the right tail of the site data distribution. The Quantile test when used in parallel with the Wilcoxon Mann Whitney (WMW) test provides the user with stronger evidence to make decisions about the comparability of site and background distributions, leading to more reliable conclusions whether the site has attained remediation levels or not. It is suggested that for best results,both WMW test and Quantile tests should be used on the same data set. Note on Comparability of Data Sets The samples collected from the two (or more)populations under comparisons should all be of the same type obtained using similar analytical methods and apparatus. In other words,the collected site and background samples should be all discrete or all composite (obtained using the same number of discrete samples, same design and pattern), and be collected from the same medium(soil) at similar depths (e.g., all surface samples or all subsurface samples) and time (e.g., during the same quarter in groundwater applications) using comparable (preferably same) analytical methods. Some good soil sample collection methods and sampling strategies are described in EPA, 2003 guidance document. Note on Influence of Outliers and Use of Lognormal Distribution Typically, in environmental data sets collected from impacted sites or monitoring wells(MWs), an outlier represents an observation coming from a potentially contaminated site location. This is especially true,when the data are collected from a site specific background area. The outlying observations need to be identified before computing the background statistics (and other estimates and test statistics) as outliers when present distort all statistics of interest, which in turn may lead to incorrect remediation and cleanup decisions for the site under investigation. For an example, inclusion of an outlier may distort the t-test statistic resulting in distorted and incorrect decision errors (Type 1 or Type 2 errors),which can lead to incorrect conclusion about the hypotheses testing. The incorrect decisions may adversely affect the human health and the environment. The main objective of using a statistical procedure is to model the majority of the data representing the main dominant population, and not to accommodate a few low probability outliers that may yield inflated and impractical statistics, results, and incorrect conclusions. For