jasp-stats
[Feature Request]: more graphical assumption testing in linear regression
Description
It would be super nice in JASP to have the possibility to easy check assumptions through depictions when running a linear regression.
Purpose
More elegant assumption testing in linear regression analysis.
Use-case
No response
Is your feature request related to a problem?
No
Is your feature request related to a JASP module?
Regression
Describe the solution you would like
This is possible though R: s library “Performance” and using the "Check model" function. I think this could be a helpful, modern and attractive added feature to JASP´s regression module.
Describe alternatives that you have considered
No response
Additional context
More information on this can been seen on this video about 16.55 minutes into the video: https://www.youtube.com/watch?v=BNTn_f43U04&t=618s and also around 7.20 minutes here: https://www.youtube.com/watch?v=EPIxQ5i5oxs
@PerPalmgren thx for the request.
Would be cool to have loess lines in our residual plots, yeah. Currently I resort to flexplot for that.
Other than that I do not see much in those videos. Most of the graphs shown are available in JASPs regression module in the "Plots" section.
The 2 plots from the video missing in JASP are:
- check for multicolinearity
- cooks distance
I agree, that the cooks distance plot looks nice, but if you just check for residuals >3z in one of the "normal" residual scatter plots, then you get that information already.
The colinearity plot does present the already available JASP-table of VIF and Tolerance, but with the legend you get some added information. So yeah, would be nice to have those, but since all the information is already available in some forme it may be of low priority to implement.
If I did misunderstand something, please do not hesitate to tell me!
