𝑡-test bootstrapping

[N1X3L]

Module development request: could bootstrapping be added to robust 𝑡-tests for 𝑀-estimators please? Thank you in advance for the consideration!

[jonathon-love]

ooh, let's ask @MAgojam

[MAgojam]

Hey Jth,
Walrus is under maintenance to add robust ANCOVA and MANCOVA, I will also add @N1X3L's request to the list.

[MAgojam]

Hey @N1X3L,
are you suggesting an option for bootstrap to play Robust Paired Samples T-Test or is it something else?
TA

[N1X3L]

Hi @MAgojam
Thank you for asking! I'm suggesting that bootstrapping also be made available for 𝑀-estimation when conducting robust 𝑡-tests on independent samples, much like how it is available for robust ANOVAs.

[MAgojam]

Hey @N1X3L,
thanks for replying with the attached screenshot, because it showed me that there is a possible bug to fix.
There is no message reminding that the group variable (Species) has three levels, but it accepts it by deleting one arbitrarily.
Since you don't have a Filter active to exclude a level, otherwise F in robust ANOVA should be equal to t^2 and it isn't.