I am estimating a 2-level nested logit. The 1st level has 3 branches, each
of which has then 3 bottom level choices. I am using alternative-specific
explanatory vbles only (my expl vbles are the exact same throughout, so I'm
just interacting them with dummies for the particular choice).
For the bottom level (level2), can I include 8 sets of explanatory vbles -
and 8 alternative-specific constants? If I do this some of the parameters
are not identified. If instead I just include 6 expl vbles and 6 ASC's (and
omit the default in each of the 3 bottom level nests - like a mlogit), I
'identify' (ie get estimates of!) all of the coefficients except for one of
the inclusive values.
I'd be grateful for any suggestions,
thanks,
Emla.
_________________________________________________________________
Get Hotmail on your mobile phone http://www.msn.co.uk/msnmobile
