TOSTER
TOSTER copied to clipboard
Implement robust test
Story: I want to apply the TOST on a sample that is not ideal for a t-test: Outliers, non-normal distribution or even ordinal data.
Possible solution: Instead of a t-Test, use
- the Mann–Whitney U test (or also: "Wilcoxon Rank Sum Test") on independent samples
- the Wilcoxon signed-rank test on paired samples.
Both, Man-W-U and Wilcoxon-s-r tests are implemented in the R base package: http://stat.ethz.ch/R-manual/R-patched/library/stats/html/wilcox.test.html
I agree this is a desirable test to add. I'm working on some new functions for proportions, and updating the power functions so they calculate power, given a sample size. But after that.
I would also like to see an implementation of non-parametric equivalence test and/or an option for Bootstrapping parameters.
See documentation