Martin,
Thank you for your reply.My dataset is set up such that individuals in my
sample have a choice between 3 different alternatives and there are quite a
few(constituting about 20% of the sample observations)who choose more than one
alternative. Now asclogit has the flexibility to handle this as it allows
multiple alternatives to be chosen for each case. However, as far as I
understand it (and as you explain) -estat mfx- does not seem to be able to
incorporate this feature of the data,as -k- needs to be set at either 1 or 2(for
my example) for all cases. The results are clearly different depending on which
option I specify. So, my question is- which is the correct way to compute
marginal effects?
Murali Kuchibhotla
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
