From reachishantsharma at gmail.com  Fri Jul 22 02:08:46 2022
From: reachishantsharma at gmail.com (Ishant Sharma)
Date: Thu, 21 Jul 2022 11:08:46 -0500
Subject: [Limdep Nlogit List] Marginal effects for Latent class choice model
 (ordered logit kernel) in NLOGIT6
Message-ID: <CABHmYCDj9URNT8_PsX+9srZXYZzqdKHVEw5V_jWFDUfJUz+HTg@mail.gmail.com>

Hello,

I am using NLOGIT6 to estimate a latent class choice model with an ordered
logit kernel. The model specifications include two classes and three
outcomes levels of the ordinal dependent variable. The model runs
successfully, except it gives an error for estimating marginal effects:

"RP cannot compute ME for ORDERED when J > 9.."

Here is the script:

CREATE ; LC1, LC2 $

ORDERED; Lhs = EV_BUY

; Rhs = one, AGE2, AGE3, AGE4, GENDER,TIER1 ,EV_FAMIL, PTPV_CHE,
RS2,CDIST,CTIME,PTDIST, CP2, CAROWNER, EMP1, PT_FAMIL,TW, CAB, PT, AUTO

; Model = logit; maxit=100; Partial Effects

; LCM = ONE,PROEV1,GC1,FEP1; Pts = 2; classp = LC1, LC2$


>From the error, it seems that NLOGIT6 does not support calculating marginal
effects for this model specification. Please assist me with this issue.

--
Thanks,
Ishant