Closed joshyam-k closed 1 year ago
We currently cannot test that adequately from the GitHub CI. Tracey will need to take a look.
yep, I'll definitely spend some time walking through this with tracey to make sure it works how we want it to work!
closing this for now while I troubleshoot modelselect behavior more thoroughly
The idea is to try to preserve additivity in MA estimates by doing model selection once at the very beginning instead of during each model fitting to a given rowvar and/or colvar level in the data.
I think that I added this code into the right section of modMAtree.R, but would appreciate a second opinion on that.