Gavin Simpson

Also, what are you hoping to get out of this? Differences for 1. `s(x, by=fac, bs="tp")`, or 2. `s(x, participant, bs="fs", m=2)`, or 3. both sets of smooths ? I'm...

I think there's a missing word in > `difference_smooths()` works on a model which does not include any `bs="re"` smooths, only random smooths of the form `s(x, participant, bs="fs", m=2)`...

Turns out I was massively over-thinking this, at least for the case of including group means (the complexity where there are multiple factors just cancels for any reasonable comparison because...

Thanks @michaelmcanally `appraise()` should now work for models fitted with the `gaulss` family and version 0.4 will shortly be sent to CRAN.

See: https://docs.github.com/en/repositories/configuring-branches-and-merges-in-your-repository/managing-branches-in-your-repository/changing-the-default-branch

Oops; referenced the wrong bug in that commit message

This is probably some conflict with selecting terms (should be smooths); does it work if you don't specify `parm`? I'll try to take a look later.

Having had another think about this, `gratia:::confint.gam` is doing the right thing. `confint` methods are traditionally for *parameters* of a model, not predictions from the model. As factor `by` smooths...

I've been playing with a few designs/ideas; assuming I get a few minutes before semester starts I hope version 0.8 of gratia comes with a hexsticker

Hi @LisaHuelsmann No developments as yet, but I think I have a very easy way to implement this using existing functionality in {gratia} so I may be able to implement...