[Limdep Nlogit List] hausmantest

Caroline Johanna Biehl Caroline-Johanna.Biehl at gmx.de
Wed Jun 18 23:44:42 EST 2008


Dear all, 

Unfortunately i am still a beginner in dealing with choice experiments and now I have the following problem: 
i want to conduct a hausman test. 
Alternative A has 3 alternative specific attributes and 2 generic ones
Alternative B has 5 alternative specific attributes and 4 generic ones
Alternative C has 2 alternative specific attributes and 2 generic ones
the fourth one is "opt out" option

For exclusion of alternative A it works fine.
But since B has lots of attributes and when i want to exclude B, my restricted model has less rows than my full model and thus the hausmantest cannot be conducted. is there a way to do it or is it just not possible?

This is my test for exclusion of alternative b

?Exclusion of b
?Hausman test
sample; all$
reject; ort=12$
reject; cset=-999$
nlogit
;lhs= choice, cset, alti
;choices= tap, public, borehole, other
;show
; model: 
U(tap) = asct+ fr1*fre1 + fre2*fre2  + tapp*price+ pincTap*pincTap/ 
U(public) = ascp + p1*p1 + p2*p2 + cons*consumpt + pubfp*pubfp + fre1*fr1 + fre2*fre2 + w2*w2+ p*price /
U(borehole) = ascb+ cons*consumpt + pump*pump + p*price + incbore*incbore/
U(other)=0 $
matrix; j1=[
		0,1,0,0,0,0,0,0,0,0,0,0,0,0,0/
		0,0,1,0,0,0,0,0,0,0,0,0,0,0,0/
		0,0,0,1,0,0,0,0,0,0,0,0,0,0,0/
		0,0,0,0,1,0,0,0,0,0,0,0,0,0,0/
		0,0,0,0,0,0,1,0,0,0,0,0,0,0,0/
		0,0,0,0,0,0,0,1,0,0,0,0,0,0,0/
		0,0,0,0,0,0,0,0,1,0,0,0,0,0,0/
		0,0,0,0,0,0,0,0,0,1,0,0,0,0,0/
		0,0,0,0,0,0,0,0,0,0,1,0,0,0,0/
		0,0,0,0,0,0,0,0,0,0,0,1,0,0,0/
		0,0,0,0,0,0,0,0,0,0,0,0,0,1,0/
		0,0,0,0,0,0,0,0,0,0,0,0,0,0,1]$
matrix; bu=j1*B; Vu=J1*VARB*j1'$

create; csetnew=3$
create; if(alti=1)altinew2=1$
create; if(alti=3)altinew2=2$
create; if(alti=4)altinew2=3$
reject; alti=2$
reject; cs2=1$
reject; altinew2=0$


?Exclude b
reject; csetnew=-999$
nlogit
;lhs= choice, csetnew, altinew2
;choices= tap, borehole, other
;show
; model: 
U(tap) = asct+ fr1*fre1 + fre2*fre2  + tapp*price+ pincTap*pincTap/ 
?U(public) = ascp + p1*p1 + p2*p2 + cons*consumpt + pubfp*pubfp + fre1*fr1 + fre2*fre2 + w2*w2+ p*price /
U(borehole) = ascb+ cons*consumpt + pump*pump + p*price + incbore*incbore/
U(other)=0 $

matrix; j2=[0,0,0,0,0,0,0,0,0,0/
		0,1,0,0,0,0,0,0,0,0/
		0,0,1,0,0,0,0,0,0,0/
		0,0,0,1,0,0,0,0,0,0/
		0,0,0,0,1,0,0,0,0,0/
		0,0,0,0,0,0,0,0,0,0/
		0,0,0,0,0,0,1,0,0,0/
		0,0,0,0,0,0,0,1,0,0/
		0,0,0,0,0,0,0,0,1,0/
		0,0,0,0,0,0,0,0,0,1]$

here i get the Error   177: MATRIX - MSUM or MDIF, matrices must have same dimensions

Thank you very much, 
caroline


-- 
GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen!
Jetzt dabei sein: http://www.shortview.de/?mc=sv_ext_mf@gmx



More information about the Limdep mailing list