the latest JASP gauntlet!

Discuss the jamovi platform, possible improvements, etc.

by scroon » Thu Oct 25, 2018 6:07 am

Hmmm, the above and regression features in the latest JASP are pretty handy ones. I don't suppose these are on your list for future jamovi versions?
Posts: 15
Joined: Tue Feb 27, 2018 3:52 am

by jonathon » Thu Oct 25, 2018 6:40 am

what did you have in mind?

to me it's a funny suggestion, because i think linear regression is an area where we really outshine jasp.

User avatar
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by scroon » Thu Oct 25, 2018 10:43 am

Well, these would be handy things,

in classical linear regression:
Bootstrap confidence intervals of coefficients
Cook’s distance (case-wise diagnostics)

in ANOVA Games-Howell post-hoc tests

But, yes, I agree with you
Posts: 15
Joined: Tue Feb 27, 2018 3:52 am

by Ravi » Thu Oct 25, 2018 1:02 pm

Hi scroon,
So the goal is to make the analyses as complete as possible without them becoming unwieldy and hard to use. Sometimes this means not adding certain functionality. Adding functionality is easy, doing it in a way that ensures usability is not. Of course we aim for the basic analyses to be as complete as possible as well, so if there is a big demand for certain functionality we would like to add it.

For instance, we want to add the case-wise diagnostics, but we first have to build a way that allows analyses to add columns to the spreadsheet (I don't think providing a table with the diagnostics for each case is very elegant, because it becomes increasingly hard to read as the number of cases increases). Once we have succeeded in doing this we will add the possibility of creating a new column in the spreadsheet with case-wise diagnostics such as Cook's distance. For the moment we provide summary statistics of the Cook's distance which gives you some insight into the data set.

About the Games-Howell post-hoc tests, we actually do provide them for One-Way ANOVA, so have a look to see whether that meets your needs :)

User avatar
Posts: 140
Joined: Sat Jan 28, 2017 11:18 am

by kris_ariya » Sun Apr 07, 2019 10:13 am

For the Cook's distance, is it possible to flag cases with a certain cutoff and output the row numbers into a summary table/graph? That way, the users can still identify the problematic data points without a need to add columns to the spreadsheet.
Posts: 5
Joined: Sat Mar 09, 2019 8:27 pm

Return to General