Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Problem using asclogit regression

From (Brendan Halpin)
Subject   Re: st: Problem using asclogit regression
Date   Tue, 29 May 2012 21:47:36 +0100

On Tue, May 29 2012, Muhammad Akram wrote:

> Dear Stata Users,
> I am trying to use asclogit model for my estimation purpose. I use command like this
> asclogit  val pris menuvariation quetime, case(id) alternative(alternativ123)
> This command comes out with the error 
> "variable alternativ123 has replicate levels for one or more cases; this is not allowed"

The alt() option identifies the values of the outcome variable, and
these values have to be mutually exclusive. 

> Actually in a survey question we asked receipent to choose the
> alternative from group (1) A B C and from group(2) E F G. Every
> individual was too select an alternative from each group. The dependent
> variable val is their choice from both groups. If some one can guide me
> what is actually wrong with this command or some variable error can be.

Clearly you have two choices being made. You could model that in two
separate models, or possibly you could create a new variable that is the
combination of the choices. This will have 9 categories (AE, AF, AG, BE,
BF, etc), which may be a lot for the model to estimate, unless you have
lots of data. Certainly you'll have a lot of parameter estimates.

Brendan Halpin,   Department of Sociology,   University of Limerick,   Ireland
Tel: w +353-61-213147  f +353-61-202569  h +353-61-338562;  Room F1-009 x 3147    ULSociology on Facebook:         twitter:@ULSociology
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index