Ven Popov

Found this thread while looking if I can specify a gamma distribution for the random effects. I know above you said you have no plans to do that, but it...

judging by the numbers, if I had to guess, in this case type=response treats the cumulative model like logistic regression with a single threshold at the grand mean, so you...

Yes, this is what it's doing. a is your reference level, so it gets a fixed estimate of 0 on the latent scale, and the b and c are relative...

btw @benwhalley you are confusing the print output format of emmeans with the object that it produces. The return value of emmeans is a complex S4 objects, but when you...

When using the new functions I noticed that cmdstan gives an error if running a model with adapt_engaged=FALSE and save_metric=TRUE. So I also added a test showing the failure and...

Strange, the cmd check passed locally. Should I try to fix the issues or do you want to review the general approach first?

Ok, I'll see to fixing the checks. As for the reading the metadata from the new files, I have noted a few questions to discuss first, but I'll probably have...

> When using the new functions I noticed that cmdstan gives an error if running a model with adapt_engaged=FALSE and save_metric=TRUE. So I also added a test showing the failure...

All the tests pass, except for WSL, which gives an error not being able to find the metric files. I'm not sure what to do about that - I don't...

Just bumping this up since it seems just that small issue with the WSL backend test is left. I could check the WSL tests in principle, but I would have...