[Limdep Nlogit List] RPL panel data random effects model

[Iain Pardoe]

Thanks Patrick and Maksym (and also Bill Greene and David Hensher for additional responses). 

To answer Maksym's question below, I updated to the most recent software build and was able to get results for the code on page N7-20 of the nlogit manual that were similar to those reported (differences are probably due to a different random number seed and/or the use of Halton draws).

I think Patrick's ideas may well come in handy in the future though as I start working with my own data :)

Cheers, Iain

-----Original Message-----
From: Maksym Polyakov [mailto:polyama at auburn.edu] 
Sent: Thursday, January 04, 2007 5:14 PM
To: Limdep and Nlogit Mailing List
Subject: Re: [Limdep Nlogit List] RPL panel data random effectsmodel [SEC=UNCLASSIFIED]

Laplagne, Patrick wrote:
> Iain,
>
> I am not an NLogit guru, but I have had the same problem as you fitting a RPL model with panel data [...] 
The problem is that these are data and examples which comes with the 
software or with the book and we should be able to reproduce results. 
Anyone was able to reproduce results of RPL models from NLOGIT manual or 
from the CD which comes with Econometric Analysis book?

-- 
Maksym Polyakov

> -----Original Message-----
> From: limdep-bounces at limdep.itls.usyd.edu.au 
> [mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of 
> Iain Pardoe
> Sent: Thursday, 4 January 2007 11:45 AM
> To: limdep at limdep.itls.usyd.edu.au
> Subject: [Limdep Nlogit List] RPL panel data random effects model
> 
> I'm a beginning nlogit user with what is almost surely a 
> naïve question (sorry).  I'm working through the nlogit v3.0 
> reference guide and have come to a grinding halt at the part 
> I'm most interested in, that is fitting a random parameters 
> logit model to panel data (section N7.6.1).  With the code on 
> page N7-20 (attached below) for the CLOGIT data I am 
> currently getting error message 806: "Line search does not 
> improve fn. Exit iterations."  I thought I better get a known 
> example working correctly first before I started with my own 
> data, so any guidance would be greatly appreciated.
> 
> Best wishes, Iain
> 
> Iain Pardoe <ipardoe at lcbmail.uoregon.edu>
> Assistant Professor of Decision Sciences
> Charles H. Lundquist College of Business
> 1208 University of Oregon
> Eugene, OR 97403-1208, USA
> Ph: 541-346-3250, Fax: 541-346-3341
> http://lcb1.uoregon.edu/ipardoe
> 
> NLOGIT ; Lhs=mode
>     ; Choices=air,train,bus,car
>     ; Rhs=gc,ttme
>     ; Rh2=one
>     ; RPL=hinc
>     ; Fcn=a_air(N),a_train(N),a_bus(N),gc(N),ttme(N)
>     ; Correlation
>     ; Parameters
>     ; Pds=3
>     ; Pts=10 $
> 
> Line search does not improve fn. Exit iterations. Status=3
>   Error   806: Line search does not improve fn. Exit 
> iterations. Status=3
> Function=  .19997662311D+03, at entry,  .12364020610D+03 at exit
>   Error  1025: Failed to fit model. See earlier diagnostic.