Second Confidence Interval Always 1.00 For K-W? about ggstatsplot

[ludist]

Looking at the closed issues, I see you very recently addressed this at #906 here for a slightly different issue and I notice there's been a very recent commit. I'll update my package installation and investigate.

[ludist]

Wow, I don't know what changed in the last 72 hours or so but updating statsExpression sent me into an hour and a half dependency hell where I had to manually install a gfortran compiler and 3 other packages on macOS x64 before I could get it working again. If anyone reading this has the same problem, just save yourself the hassle and google the package names and "macOS binaries" to get to the r-project page and then install the .tgz manually.

Anyway, I tried running the code again, along with several attempts at parameter calls based on my review of the commits, and I still get 1.00 for the second CI. If it's fixed and requires a parameter call, could you expand on the documentation a bit? Thanks again!

[IndrajeetPatil]

Yes, this upper bound for the effect size for KW (rank epsilon squared) is 1.0.

For more, see: https://easystats.github.io/effectsize/reference/rank_epsilon_squared.html

[ludist]

Thank you!

[IndrajeetPatil]

The underlying functions used to run inferential statistics and effect size estimation are catalogued here, and more details can be found in linked pages:

https://indrajeetpatil.github.io/ggstatsplot/reference/ggbetweenstats.html#one-way-anova