rank-ordered logit models aka exploded logit model

[Jan Verbelen]

Dear list members,

Has anyone of you ever used a technique called 'exploded logit model' aka
'rank-ordered logit' models.  I understand it is sometimes used in
marketing research and health research.
I'm looking for experience or references where SPSS was used to accomplish
such an analysis.

I'm trying to analyze the ranking of several items.  The items are groups
of people.  Respondents were asked to rank five social groups in a surevy.
 They were asked to rank these groups (from most preferred as neighbour to
least preferred as neighbour).   If they were in a position to choose
their neighbours what would their prefrences look like?
The outcome is a ranking of five categories.  Now I'm trying to model thes
preference data, these rankings, to explore what factors have an impact.

Can anyone point me to references  or people with experience in doing this
with SPSS?

many thanks in advance,
-----------
Jan Verbelen

Studiecel / L806
Vlaams Agentschap voor  Personen met een Handicap

Study Department / L806
Flemish Agency for  Persons with a Disability

Sterrenkundelaan 30
1210 BRUSSEL

TEL ++32 (0)2 225 86 32
FAX ++32  (0)2 225 84 05

URL http://www.vlafo.be

---------------------------------------------------------------------------------
DISCLAIMER :
De personeelsleden van het agentschap doen hun best om in e-mails
betrouwbare informatie te geven. Toch kan niemand rechten doen gelden op
basis van deze inhoud. Als in de e-mail een stellingname voorkomt, is
dat niet noodzakelijk het standpunt van het agentschap. Rechtsgeldige
beslissingen of officiele standpunten worden alleen per brief toegestuurd.

[Anthony Babinec]

Jan,
I don't have an example of doing this in SPSS.

You can take a look at the website for Latent Gold Choice
and download a PDF file of the User's Guide at

http://www.statisticalinnovations.com/products/choice.html

If you browse the User's guide and do a Find on "rank," you
will find a discussion of a sequential logit model for
ranks along with a few citations. This particular model
is illustrated in Tutorial 5 in that document.

Issues for doing this model in or out of SPSS are:

- SPSS does not have a dedicated conditional logit model procedure.
You could "trick" Cox Regression into fitting a conditional logit
model with the right setup.
- With ranks as the dependent variable, you would need to fit
a sequence of conditional logit models.
- You might need to account for respondent heterogeneity by
using either latent classes or a mixed logit or hierarchical Bayes
model.

You might approach Jeroen Vermunt at the University of Tilburg to
pursue this topic.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jan
Verbelen
Sent: Thursday, December 21, 2006 8:35 AM
To: [hidden email]
Subject: rank-ordered logit models aka exploded logit model

Dear list members,

Has anyone of you ever used a technique called 'exploded logit model' aka
'rank-ordered logit' models.  I understand it is sometimes used in
marketing research and health research.
I'm looking for experience or references where SPSS was used to accomplish
such an analysis.

I'm trying to analyze the ranking of several items.  The items are groups
of people.  Respondents were asked to rank five social groups in a surevy.
 They were asked to rank these groups (from most preferred as neighbour to
least preferred as neighbour).   If they were in a position to choose
their neighbours what would their prefrences look like?
The outcome is a ranking of five categories.  Now I'm trying to model thes
preference data, these rankings, to explore what factors have an impact.

Can anyone point me to references  or people with experience in doing this
with SPSS?

many thanks in advance,
-----------
Jan Verbelen

Studiecel / L806
Vlaams Agentschap voor  Personen met een Handicap

Study Department / L806
Flemish Agency for  Persons with a Disability

Sterrenkundelaan 30
1210 BRUSSEL

TEL ++32 (0)2 225 86 32
FAX ++32  (0)2 225 84 05

URL http://www.vlafo.be

----------------------------------------------------------------------------
-----
DISCLAIMER :
De personeelsleden van het agentschap doen hun best om in e-mails
betrouwbare informatie te geven. Toch kan niemand rechten doen gelden op
basis van deze inhoud. Als in de e-mail een stellingname voorkomt, is
dat niet noodzakelijk het standpunt van het agentschap. Rechtsgeldige
beslissingen of officiele standpunten worden alleen per brief toegestuurd.