Multicategorical x in bmlm mediation?

Modeling
1

Hiya all, new to Stan here and I have a simple data-wrangling issue I’d like some advice on.

I’m aiming to run a mediation model from a simple repeated measures study, where x is the condition a given participant was in, m is a continuously measured mediator (measured only once), and y is a continuously measured DV measured at two time-points. Though it’s not necessarily tailored to 2-2-1 mediations, I was thinking I could still use bmlm–here’s a quick mock dataset of what the data would broadly look like:

image
image992×715 32.4 KB

And here’s the kind of basic model creation in bmlm I’m picturing at the moment

fit <- mlm(d = df,
           id = "id",
           x = "x", #??? 
           m = "m",
           y = "y", 
          iter = ...
          etc...)

The problem I’m seeing for use in bmlm is that x is multicategorical, not continuous, but I’m not sure how to specify that for bmlm purposes–any thoughts or ideas how to make a categorical x work in bmlm? Or is there perhaps an alternative way of going about this kind of model build?

2

replying for post visibility, any ideas on using categorical vars in bmlm? Or is there perhaps a better place for me to ask this kind of question?
Cheers!

3

Sorry for the late reply. I am not sure who are the bmlm folks are here on the forum. @andrjohns might know? Or you could also try posting over the bmlm github page here: https://github.com/mvuorre/bmlm