[Limdep Nlogit List] Request to post on discussion group -- seeking for guidance on running DCE analysis in Nlogit

[Shayma Mohammed Selim]

Good afternoon,

I was recently introduced to the List Server for Limdep/Nlogit. Relating to this, I'm hoping to ask if it would be possible to post a message on the Limdep/Nlogit discussion board to seek help/advice on the following:

---------

I'm working on a DCE analysis (using Nlogit) and am hoping to ask for help/advice on “fixing” the following.

Context: We conducted a DCE study (sample size =361). For the analysis, we are hoping to do a latent class model. We combined two datasets together (unconditional and conditional). The conditional choice refers to instances where depending on the option chosen by the participant, an additional question appears – hence, each respondent has a different number of choice observations.



We are hoping to have class assignments based on age (continuous), sex (categorical), and total number of appointments – apptsum (continuous).



The Nlogit syntax that we used is below:



sample ;all $



Timer

Nlogit

;lhs = decision,cset,choice_s

;choices = 1,2,3,4,5

;lcm=age, sex_2, apptsum

;maxit=200

;pts=2

;pds=panel3

;checkdata

;model:

U(1) = asc1 + mod_2*mod_2 + mod_3*mod_3 + mod_4*mod_4 + mod_5*mod_5 + tim_2*tim_2 + tim_3*tim_3 + con_2*con_2 + con_3*con_3 + int_2*int_2 /

U(2) = asc2 + mod_2*mod_2 + mod_3*mod_3 + mod_4*mod_4 + mod_5*mod_5 + tim_2*tim_2 + tim_3*tim_3 + con_2*con_2 + con_3*con_3 + int_2*int_2 /

U(4) = asc3 + mod_2*mod_2 + mod_3*mod_3 + mod_4*mod_4 + mod_5*mod_5 + tim_2*tim_2 + tim_3*tim_3 + con_2*con_2 + con_3*con_3 + int_2*int_2 /

U(5) = mod_2*mod_2 + mod_3*mod_3 + mod_4*mod_4 + mod_5*mod_5 + tim_2*tim_2 + tim_3*tim_3 + con_2*con_2 + con_3*con_3 + int_2*int_2

$


Queries:

• When we run the following syntax with ONLY one specified variable in the lcm (e.g., ;lcm=age), we generate an output. However, when we have attempted to run the same syntax with more than one specified variable in the lcm (e.g., ;lcm-age, sex_2, apptsum), we received the following error message and no output was generated. We are hoping to ask for guidance on how we can fix this.


Error: 1027: Models - estimated variance matrix of estimates is singular

Thank you very much for the guidance.

---------

Thank you and looking forward to your response.

Kind Regards,
Shayma