Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: vce(cluster group) or GLLAMM?
From
Stas Kolenikov <[email protected]>
To
"[email protected]" <[email protected]>
Subject
Re: st: vce(cluster group) or GLLAMM?
Date
Thu, 11 Jul 2013 09:03:23 -0500
If this is a simultaneous move cooperative game, random effects as
specified may be OK, and the gllamm model as specified looks fine
(although -xtmelogit- will run faster... I don't know if you are going
to care about the minute technical differences between probit and
logit). If this a sequential move or a prisoner's dilemma kind of a
game, you may need to distinguish the "origin" and "destination"
subjects, like it is sometimes done in dyadic analysis. Kit Baum has
probably done some work of that kind, as far as I recall one of his
presentations some five years ago on international trade.
-- Stas Kolenikov, PhD, PStat (SSC)
-- Senior Survey Statistician, Abt SRBI
-- Opinions stated in this email are mine only, and do not reflect the
position of my employer
-- http://stas.kolenikov.name
On Thu, Jul 11, 2013 at 8:56 AM, FredWan <[email protected]> wrote:
> Dear fellow STATAlist members,
>
> I am trying to build a dynamic probit model with RE clustered on group
> level.
>
> My data:
> I have experimental data from 177 subjects, which play a repeated game in a
> fixed matching with two other subjects (59 groups of three).
>
> What I want:
> I want a dynamic probit on a binary decision - but need to control for
> correlations between observations of subjects in the same group.
> Random-effects should be introduced on the subject (level2) and group level
> (level3).
>
> As far as I am aware, xtlogit does not allow the vce(cluster group) option,
> so I have been looking into GLLAMM.
>
> xi: gllamm y y_1 i.treatment xxx , i(id group_index) link(logit) family(bin)
> adapt seems to work, though I am a bit reluctant to use a method in a paper,
> if I am not 99.9% sure it is correctly specified :)
>
>
> Is there an alternative to using GLLAMM, but does something similiar to
> xtprobit vce(cluster group)?
>
> If I use the last periods group average of decision y in xtprobit, which
> should be sufficient to control for the effect the group had?
>
>
>
> Many thanks!
>
> Fred
>
>
>
> --
> View this message in context: http://statalist.1588530.n2.nabble.com/vce-cluster-group-or-GLLAMM-tp7580395.html
> Sent from the Statalist mailing list archive at Nabble.com.
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/