This question concerns the set up of data for estimation using nlogit
and nlogitrum commands.
Do I need to populate the dataset with observations for all the choices
the individual did not make? Let "chosen" be the dependent variable, I
only have data for the observations where "chosen" equals 1. For
example, of possible choices: A, B, C, D, I have an observation that
they selected D and can set the variable "chosen" equal to 1. Do I need
to generate three more observations where "chosen" takes a value of 0
for choices A, B, and C, thus making my dataset 4 times as large? My
understanding is that I do, but it doesn't seem to be very efficient
considering I have a very large dataset. Thanks in advance for any
help.
=====================
Jeremy Prahm
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/