Andi Fugard @andifugard.info · 18/09/2026G'morning :) That was painless (thankfully)! I'll upload the code somewhere - probably worth looking at more details of the fits... 100
Andi Fugard @andifugard.info · 18/09/2026Have you found an example where it makes much of a difference to slopes? I had go by deleting data for a level one by one and then trying with and without the empty level, using MASS::polr... Level 2 just had 5 observations so didn't expect it to make much difference. 120