Quantile Non-Parametric Test
Quantile Non-Parametric Test
See Also:
Wilcoxon Non-Parametric Group Comparison
Description:
The Quantile test is provided in the U.S. Navy 1999 Guidance Document as a test to determine if compliance point measurements are statistically elevated when compared to background. It is a non-parametric test, and therefor a normal distribution is not required and the test is suitable for any number of non-detects. In ChemStat, the test can be performed at the 99% or 95% confidence levels.
Use:
In ChemStat, the Quantile test can be run in two ways:
- It can be used to determine if compliance point observations as a group are statistically elevated when compared to background point measurements as a group, however, it can not determine which individual monitoring points are elevated. This is how the test is presented in the U.S. Navy 1999 Guidance Document.
- It can be used as an intra-well comparison to determine if a single compliance well is elevated when compared to all background measurements, however, a large number of data points from the compliance well may be required. This implementation is more appropriate for RCRA regulated sites.
Use the right-click menu in ChemStat to select the confidence level and the well comparison type.
Remarks:
For use in RCRA monitoring sites when detecting a statistical significance in individual compliance wells, a large number of samples may be required to perform a suitable analysis. Pay attention to the required sample sizes on page 67 of the U.S. Navy 1999 Guidance.
ChemStat obtains r and k from Tables A11 and A13 in the U.S. Navy 1999 Guidance. Therefore, ChemStat can not calculate values of r and k for n and m that exceed the range of these tables. For intermediate values of n and m not listed in the table, ChemStat uses the next highest value of n and m for which there is a value in the table.
Also note that the Type I error rate of 0.01 or 0.05 is approximate. Tables A11 and A13 in the U.S. Navy 1999 Guidance Document provide precise Type I error rates.