A predictor x
with measure error can be modeled with brms
as
brm(y ~ me(x, sdx), data = ...)
If I have a grouping variable subject
, can I do the following?
brm(y ~ me(x, sdx) + ( me(x, sdx) | subject ), data = ...)
A predictor x
with measure error can be modeled with brms
as
brm(y ~ me(x, sdx), data = ...)
If I have a grouping variable subject
, can I do the following?
brm(y ~ me(x, sdx) + ( me(x, sdx) | subject ), data = ...)
If I remember correctly you want the grouping variable like this me(x, sdx, gr = subject).
I should have mentioned that, without measurement error, the original model is
brm(y ~ x + ( x | subject ), data = ...)
So, I’m uncertain whether the following would still work when measurement error is added:
brm(y ~ me(x, sdx) + ( me(x, sdx) | subject ), data = ...)