[Limdep Nlogit List] selection model with the selection equation as a mulitnomial logit model

[Barry Quinn]

Hi Diane,

If i am reading your problem correctly the first stage is the selection equation " a dichotomous variable (0,1)" and the second stage is your outcome equation "multinomial model"??

If so then you are dealing with estimation of a multinomial model with selectivity which is already pre-programmed in Limdep.

So the suggested command would be something like 

Probit; lhs=dichotomous selection variable ;rhs=characteristics affecting selection;hold$
LOGIT 	; Lhs = dependent variable(s)(coded y=0,1,,,,,J OR a set of J+1 proportion variables)
		; Rhs = characteristics affecting multinomial choices; select$


Barry Quinn


Queen's University 
Management School 
Queen's University Belfast
25 University Square
Belfast
Northern Ireland | UK | 
BT7 1NN
Email: b.quinn at qub.ac.uk
www.qub.ac.uk/mgt



-----Original Message-----
From: limdep-bounces at limdep.itls.usyd.edu.au [mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of Diane Dancer
Sent: 27 August 2010 05:05
To: limdep at limdep.itls.usyd.edu.au
Subject: [Limdep Nlogit List] selection model with the selection equation as a mulitnomial logit model

Dear All,

I have a problem with a sample selection model and am wondering if anyone has programmed this model.

In the selection equation we have a dichotomous variable (0,1) . If the respondent answers 1, we then have a multinomial model in the second stage with three possible responses.

I would be very grateful for any help in programming this model.
Diane Dancer
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au