search

jasp-stats / jasp-issues

This repository is solely meant for reporting of bugs, feature requests and other issues in JASP.
58 stars 29 forks source link

CI correction based on holm's method #970

Open saurabhr opened 4 years ago

saurabhr commented 4 years ago

how to get corrected CI based on holm's correction instead of corrected CI being based on bonferroni in analysis of post-hoc of interactions in ANOVA?

boutinb commented 4 years ago

@JohnnyDoorn Could you answer this request?

JohnnyDoorn commented 4 years ago

Hi @saurabhr

If I am not mistaken, we use the Tukey method by default and not the bonferroni method. Unfortunately this method cannot be specified in JASP, but it would be very nice to have this. This is mostly because multiple p-value corrections can be specified, so we would not know which specific one the user wants to use for the CIs.

Cheers Johnny

tomtomme commented 8 months ago

@saurabhr @JohnnyDoorn To solve this request we could: a) display a completetly new table for each correction selected. This way also the CIs can be recomputed for the selected method without bursting the width of the table into infinity b) we could allow only to select one correction at a time, instead of multiple ones. CIs and ps would then respect that correction.

I would favour b)

tomtomme commented 7 months ago

@JohnnyDoorn What do you think?