Comparison of Student – t, Welch s t, and Mann – Whitney U Tests in Terms of Type I Error Rate and Test Power

In this study, we compared the Student's t-test, Welch's and Mann-Whitney U test, in terms of their type I error rate statistical power when assumptions parametric tests are violated different situations. Materials used consisted random numbers generated using Numpy library Python programming language. All were from a normal distribution with N (0, 1) parameters. Balanced unbalanced experimental conditions simulated 50 000 times for each combination. The study revealed that, comparison to other tests, Welch’s t - test was particularly more conservative rate. It discovered that Student-t had higher values than mainly only small sample size observations analysis. This simulation indicated is robust preserving normal. Therefore, practice, use Welch t-test recommended based on findings study. One recommendations question should also be evaluated cases where have distributions.