Use of non-parametric tests
The TBT was stopped early due to disproportionate adverse outcomes between groups. The study was originally designed as a single tailed test, but the intermediate analysis was designed as a two tailed test. Upon termination of the study, the data did not conform to the assumptions of parametric tests (normalcy, variance, and linearity) and therefore had to be analyzed with non-parametric tests. While non-parametric tests are valid and do offer statistically significant findings, they are less generalizable since the sample does not conform to the population profile.
| « Previous: Data analysis | » Next: Initial findings |