I have highly positively skewed outcome variable. I need to publish my results in an academic journal.

tidybayes allows calculating credible intervals (CI) using median or mean with high density intervals (HDI) or quantile intervals (QI).

Should I report mean and HDI or median and QI? Probably reporting mean and QI is not a good practice?

  • $\begingroup$ Do you have to choose either or? Are the two very different (probably since it is very skewed)? What do you want to communicate? Perhaps you could show the distribution itself in a figure? $\endgroup$ – einar Aug 19 '20 at 7:54
  • $\begingroup$ Considering that most of the research bases on frequentist statistics, which of them is more close to confidence intervals. HDI? $\endgroup$ – st4co4 Aug 19 '20 at 9:52
  • $\begingroup$ In frequency arguments as in Bayesian ones there are many many ways of defining an interval that captures, say, 79% of the distribution. Many of them make sense. It is really your choice based on what makes sense for the application. This question is similar to yours and might help: stats.stackexchange.com/questions/24588/… $\endgroup$ – einar Aug 19 '20 at 11:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.