st: nlogitrum and nlogit: unbalanced data

[John Fulton <fulton@ssc.wisc.edu>]

Hi Statalist. Thanks for all the wonderful help.
Here's a puzzler:
I ran a model using nlogit, but nlogitrum kicks up an "unbalanced data" 
error.
I've read Heiss's 2002 article but haven't gleaned much in the way of 
what the problem might be, let alone the solution.

I don't think I've made a syntax error, but I've included the relevant 
command lines below.
This model has a degenerate nest and three levels. My indepvars for the 
top two nests are dummy interactions between chooser-specific 
characteristics and the nest variable - as is done in the "restaurant" 
example - so "var2" is "var2*(middle==1)" for all values. Similarly, 
"var3" is "var3*(top==1)."

Here's the syntax for nlogitrum. It outputs the correct tree, then gives 
an "unbalanced data" error.
nlogitrum chosen var1 var2 var3 [fw=count], group(group_id) nests(bottom 
middle top)

Here’s the nlogit command that works:
nlogit chosen (bottom=var1) (middle=var2) (top=var3) [fw=count] , 
group(group_id)


NB that my "levels" are reverse coded from stata's, so that "altsetb" 
variable is my "1" level, altset2 is "2" and altset1 is "3".

Thank you,
John.



*
*   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/