I am PhD student and working on repeatability of dynamic penetration testing. When a test is performed in a number of times. It is required to determined whether one value in a set of determinations is an outlier, which can be excluded from the set and to be disregarded in the analysis. Contaminated data could be distinguished by engineering judgment or suggested statistical tests. It is difficult for me to choose a proper statistical test to exclude contaminated data because of large number of suggested method. What statistical test may suggest to be popular in geotechnical engineering? Best regards --------------------------------- Mahdi Khodaparast PhD Student, Geotechnical Group, Civil Engineering Department, University of Tehran.