[Limdep Nlogit List] Interaction between ASC and socio-economic variables

[F WD]

Dear Fredik Carlsson,

Actually, I don't use ASC in the second utility function in my model. I
made a mistake to write the syntax in the previous mail. I use the syntax
as below:

RPLOGIT   ; Lhs = CHOICE ; Choices = 1,2,3
;model:
U(1,2) = PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 + X63*X63 +
LOCATION*LOCATION + X65*X65 + QUANTITY*QUANTITY +
AGE*ASCXAGE + EDUC*ASCXEDUC + ORGANIC*ASCXORGA + RISK*X104 + KTPA*ASCXKTPA
+ COMMUNITY*ASCXKTCO + INCOME*ASCXINCO /
U (3) =  ASC + PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 +
X63*X63 + LOCATION*LOCATION + X65*X65 + QUANTITY*QUANTITY
; fcn = ASC(n), Price(c), X76(n), HANDLING(n), X62(n), X63(n), LOCATION(n),
X65(n), QUANTITY(n)
; Maxit = 50 ; Pts = 25 ; Halton ; Output = 3 ; Pds = 3 $

when I run the result is Error   1027: Models - estimated variance matrix
of estimates is singular..


Pada tanggal Kam, 18 Jun 2020 pukul 00.40 Fredrik Carlsson <
fredrik.carlsson at economics.gu.se> menulis:

> You have an ASC in all utility fcns in the second specification
>
> Best
> Fredrik
>
> Skickat från min iPhone
>
> > 17 juni 2020 kl. 16:51 skrev F WD via Limdep <
> limdep at mailman.sydney.edu.au>:
> >
> > Dear all,
> >
> > I am using a mixed logit model to estimate the preferences. In my model,
> > there are two types of alternatives and a status quo (alternative
> specific
> > constants). So basically, I had three utility functions. When I tried to
> > estimate the basic model without interaction with some socio-economic
> > variables, it works. However, When I tried to add the model with
> > interaction between ASC and socio-economic variables, it did not work. As
> > for your information, I use ASC to measure utility of status quo
> > alternative. I use dummy variable coding 1 if the respondent chooses the
> > status quo, and 0 otherwise. I want to know the effect of socio-economic
> > variable on the probability to opt “status quo”.
> >
> > The syntax of mixed logit is below for basic model and it run well
> >
> > RPLOGIT   ; Lhs = CHOICE ; Choices = 1,2,3
> >
> > ;model:
> >
> > U(1,2) = PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 + X63*X63
>     +
> > LOCATION*LOCATION + X65*X65  +  QUANTITY*QUANTITY /
> >
> > U (3) =  ASC + PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 +
> > X63*X63 + LOCATION*LOCATION + X65*X65 + QUANTITY*QUANTITY
> >
> > ; fcn = ASC(n), Price(c), X76(n), HANDLING(n), X62(n), X63(n),
> LOCATION(n),
> > X65(n), QUANTITY(n)
> >
> > ; Maxit = 50 ; Pts = 25 ; Halton ; Output = 3 ; Pds = 3 $
> >
> >
> >
> > However, when I run to interact between ASC and some socio-economic
> > variables, it does not work.  I got the message: *Error   1027: Models -
> > estimated variance matrix of estimates is singular*
> >
> > I use the syntax below:
> >
> > RPLOGIT   ; Lhs = CHOICE ; Choices = 1,2,3
> >
> > ;model:
> >
> > U(1,2) = ASC + PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 +
> > X63*X63 + LOCATION*LOCATION + X65*X65 + QUANTITY*QUANTITY +  AGE*ASCXAGE
> +
> > EDUC*ASCXEDUC + ORGANIC*ASCXORGA + RISKst*ASCXRISK + KTPA*ASCXKTPA +
> > COMMUNITY*ASCXKTCO /
> >
> > U (3) = ASC + PRICE*PRICE + X76*X76 + HANDLING*HANDLING + X62*X62 +
> X63*X63
> > + LOCATION*LOCATION + X65*X65 + QUANTITY*QUANTITY
> >
> > ; fcn = ASC(n), Price(c), X76(n), HANDLING(n), X62(n), X63(n),
> LOCATION(n),
> > X65(n), QUANTITY(n),
> >
> > ; Maxit = 50 ; Pts = 25 ; Halton ; Output = 3 ; Pds = 3 $
> >
> >
> >
> > Does anyone know why the model in the interaction is not run? Is there
> any
> > wrong in the syntax? Thank you for your help.
> >
> > best regards,
> > F
> > _______________________________________________
> > Limdep site list
> > Limdep at mailman.sydney.edu.au
> > http://limdep.itls.usyd.edu.au
>


-- 

*Fanny Widadie, S.P., M.Agr*
Department of Agricultural Economics and Agribusiness
Faculty of Agriculture
University of Sebelas Maret
Jalan Ir. Sutami No. 36 A Kentingan Surakarta - Central Java
Indonesia 57126