When we calculate some multiple comparison test (the Dunn's test, the Nemenyi test, or the Conover's test) using, for example, the PMCMR package in R, we can only get p-levels of this comparison. Is there a way to calculate the effect size and confidence intervals for these tests?
These are obvious depending on the method. See attached. They are often only t-tests so use that effect size or google your question. For Confidence intervals you could look at a statistics book.
David Booth
Hello Igor,
This resource is an excellent overview for using R to generate confidence intervals for a range of effect size statistics. It appears to cover the kinds of comparisons you propose to run: http://osc.centerforopenscience.org/static/CIs_in_r.html
Good luck with your work.
Effect_size_calculation_for_pairwise_comparisons_Sidak_adjustment
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package