[gov2001-l] multiple equation models - Gov2001 - lists.gking.harvard.edu

List overview All Threads
Download

[gov2001-l] multiple equation models

[gov2001-l] Grabbing standard...

[gov2001-l] Recoding variables

olau＠fas.harvard.edu

27 Apr 2003 27 Apr '03

1 a.m.

Dear all, Does anyone know of a paper that deals with a multiple equation model with two different distributions? I think I have a case where I have stochastically dependent Ys (probability of crisis, length of crisis). The problem is that the first Y (pr crisis) is logisitic and the second is weibull. Any suggestions? Thanks, Olivia

Reply

Show replies by date

kimai＠fas.harvard.edu

27 Apr 27 Apr

9:58 a.m.

This could be done if you model it hierarchically (you also need a variable that belongs to one equation but not to the other). But, it may be beyond what we learned in the class... You can still model the two decisions independently, though. Kosuke On Sun, 27 Apr 2003, Olivia Lau wrote:

Dear all, Does anyone know of a paper that deals with a multiple equation model with two different distributions? I think I have a case where I have stochastically dependent Ys (probability of crisis, length of crisis). The problem is that the first Y (pr crisis) is logisitic and the second is weibull. Any suggestions? Thanks, Olivia _______________________________________________ gov2001-l mailing list gov2001-l(a)fas.harvard.edu http://www.fas.harvard.edu/mailman/listinfo/gov2001-l

Reply

Gary King

10:07 a.m.

you could do something where the distibutions are stochastically independent but parametrically dependent. then you merely multiply the densities to get the joint density, but you have to estimate them together becasue fo the shared parameters. oh, and nothing is beyond the scope of this class!, but there is a question as to whether the setup would make a difference in your situation. Gary On Sun, 27 Apr 2003, Kosuke Imai wrote:

This could be done if you model it hierarchically (you also need a variable that belongs to one equation but not to the other). But, it may be beyond what we learned in the class... You can still model the two decisions independently, though. Kosuke On Sun, 27 Apr 2003, Olivia Lau wrote:

Dear all, Does anyone know of a paper that deals with a multiple equation model with two different distributions? I think I have a case where I have stochastically dependent Ys (probability of crisis, length of crisis). The problem is that the first Y (pr crisis) is logisitic and the second is weibull. Any suggestions? Thanks, Olivia _______________________________________________ gov2001-l mailing list gov2001-l(a)fas.harvard.edu http://www.fas.harvard.edu/mailman/listinfo/gov2001-l

_______________________________________________ gov2001-l mailing list gov2001-l(a)fas.harvard.edu http://www.fas.harvard.edu/mailman/listinfo/gov2001-l

Reply

7676

days inactive

7676

days old

gov2001@lists.gking.harvard.edu

Manage subscription

2 comments

3 participants

tags (0)

participants (3)

Gary King
kimai＠fas.harvard.edu
olau＠fas.harvard.edu