Marshall University Math Colloquium
January 27, 2016
Most of the standard statistical hypothesis testing procedures are based on the normality assumption, i.e., the data is normally distributed and also on few other assumptions. What happens if these assumptions do not hold? The answer lies in the Permutation testing approach also known as randomization or re-randomization tests. In this talk, I will discuss the randomization principle and also how these tests are impervious to complications that defeat other classical statistical significance testing techniques.
Keywords: Permutation testing, Statistical significance, Randomization, Distribution-free approach, P-values.