sure, that's a good approach.
Gary
On Sat, 10 May 2003, Maria Popova wrote:
Hi Gary and Kosuke,
We want to check the causal mechanism and we're using a mnl. Do we just
include all the variables that we think are causally posterior to our main
variable of interest and look for the coefficient of the main variable of
interest to drop as close to 0 as possible?
Thanks!
Maria and Stan