Reduced rank regression

General
1

So with reduced rank regression we identify response variables associated with our outcome of interest (Y) and model relationships between the independent variables say dietary data vs. the selected responses. If there is an association between the selected factor from this initial modelling and the outcome variable (Y), could I conclude that the association is due to mediating effects of the response variables or to the independent effect of the selected factor? Thanks

2

Hi! Sorry it took us a while to respond.

So this is a really broad question. And not really Stan specific, right? Do you have a model or a specific scenario where you run into this?

This kind of sounds like there are causal assumptions going into the model. I think it is really hard without knowing more about the problem to guide you further on this one.

Also, with “response variable” you actually mean independent variables? I usually hear “response” being used synonymously for dependent/outcome variable. Maybe I’m missing something here?

3

Thanks Max_Mantei for the response.
No this is general question regarding the RRR methodology. I did some analysis a while back and was just thinking about it in general terms. I meant dependent variables/response variables.

Thanks