search

jamovi / walrus

Walrus - Robust Statistical Methods
2 stars 4 forks source link

No A*B post hoc in Robust ANOVA #3

Open pratinhos opened 4 years ago

pratinhos commented 4 years ago

I have been trying to perform a robust ANOVA to check for the lack of differences between two groups across time in jamovi. However, while everything worked smoothly, I have been unable to get the post hocs for the interaction effect groups*time.

I have read the documentation of Walrus and this appears to be an expected behaviour. However, after reading the manual from the original WRS2 package, I came to notice that they appear to have implemented post hocs for the interaction of variables.

If my understanding is indeed correct, would it be possible to add this to Walrus? Thanks a bunch for your work!

jonathon-love commented 4 years ago

hi,

could you point me to the appropriate function in WRS2 ... i just had a quick glance and couldn't see it.

cheers

pratinhos commented 4 years ago

Sure, either in the WRS2 documentation (https://cran.r-project.org/web/packages/WRS2/vignettes/WRS2.pdf) or the R Companion website (https://rcompanion.org/rcompanion/d_08a.html) https://rcompanion.org/rcompanion/d_08a.html, I can see that at least the mcp2a and mcp2atm produce post hoc outputs containing, among others, lines similar to the following (taken from the WRS2 documentation):

gender1:alcohol1 -3.33333 -20.49551 13.82885 0.61070 gender1:alcohol2 -29.16667 -45.67245 -12.66089 0.00025 gender1:alcohol3 -25.83333 -42.53204 -9.13463 0.00080

As a comparison, on jamovi I am 'only' able to get two post hoc tables, each for within-variable comparisons (gender or alcohol).

So, either I misunderstood the documentation of WRS2 (apologies in advance, but I just recently begun to learn R motivated in part by jamovi) or there is actually something missing here...

Thanks for the hard work!

18/07/2020, 23:12 por notifications@github.com:

hi,

could you point me to the appropriate function in WRS2 ... i just had a quick glance and couldn't see it.

cheers

— You are receiving this because you authored the thread. Reply to this email directly, > view it on GitHub https://github.com/jamovi/walrus/issues/3#issuecomment-660555687> , or > unsubscribe https://github.com/notifications/unsubscribe-auth/AGWYNTNHFMHADMBJ2BV4XQ3R4IT5XANCNFSM4PAJ3RYA> .