ARCHIVED REPORTS_XR0013078 - Laserfiche WebLink

Home Browse Search

15 June 2009 AGE-NC Project No.05-1306 t». Page 20 of 33 6.2. MANN-WHITNEY U TEST The Mann-Whitney U Testis equivalent to the Wilcoxon Rank sum test and is applicable to data that may or may not exhibit seasonal behavior and is a non-parametric test for assessing whether two samples of observations come from the same distribution.The null hypothesis in the Mann-Whitney test is that the two samples are drawn from a single population, and therefore that their probability distributions are equal. It requires the two samples to be independent, and the observations to be ordinal or continuous measurements.The test involves the calculation of a statistic,called U,whose distribution under the null hypothesis is known. The test requires eight consecutive rounds of quarterly or semi-annual sampling results and for the purpose of the Mann-Whitney U Test,all non-detect values are assigned a data value of zero(0).The data are onto a spreadsheet in the order the data was collected. A rank is assigned to each sample value,with the smallest value ranked#1 and the largest ranked#8. Sum the ranks for the data in the first year. Denote this sum as Wrs (Wilcoxon rank sum) and calculate the U statistic where U=26 Wrs. To interpret the U statistic for 2 groups of 4 samples,at U<3,the probability that year 2 data show a decrease relative to year 1 is at least 90 percent,and so,U<3 will be acceptable to show that } contaminant concentration is declining.If there are ties in the sample data,an average rank value for +— %aw the tied data is calculated and this average rank is assigned to the tied sample data. The level of confidence(a)is calculated(probability and the U statistic)for values of U calculated for two groups of four samples each. ti.. The spreadsheet utilized allows for interpretation of increasing or decreasing trends.At a 90 percent confidence level, a U statistic of three (3) or less indicates a decreasing trend and a U statistic of `. thirteen(13) or more indicates an increasing trend. If the data does not pass either the increasing or decreasing trend test, the "no trend" result will be displayed. Additional information regarding the Mann-WhitneyU Test assumptions andprocedures is included in Appendix J. In order to utilize a program that was designed to analyze data sets from LUST sites that has formulas locked and not open to manipulation,the Mann-Whitney U Statistical Test Form 4400-216 (212001) from the Wisconsin Department of Natural Resources Remediation and Redevelopment Program was utilized. 6.2.1. Mann-Whitney U Test Data and Assumptions Data were assembled to include one or more contaminated monitoring wells near the down gradient plume margin, monitoring wells near the source zone, and monitoring wells along the flow line L. between the source and the down gradient plume margin well(s). Data were analyzed in groups of eight rounds per data set covering two years of quarterly sampling events. Therefore, wells with Advanced GeoEnvironmental,Inc. v