M-equation logit / probit models in SPSS

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 2
Joined: Thu Jun 12, 2014 5:30 pm

M-equation logit / probit models in SPSS

Postby gysc » Mon Jun 16, 2014 11:16 am

I am not sure what to even call the system of equations that I have but here is a description of its form, maybe someone can tell me if there is a procedure in SPSS to estimate such models. My data set is longitudinal in nature, formatted in a long format, eg person-period.

The equations I am estimating are:

y1 = logit(x, y2)
y2 = logit(x, y1)

the y1 and y2 are 1 if a morbidity occurs for that person in that time period and zero otherwise. Once a person has a morbidity they contribute no more information. The x's are explanatory (mainly categorical) variables, eg age, gender, time.

Individually these are classic hazard models, easily estimated as logit, complementary log-log or even probit.

However having one morbidity (say, y1) will increase the changes of the other morbidity (ie y2) so it have utility as an explanatory variable.

So, is there a procedure in SPSS that will recognise this link between the 2 equations and estimate me a model? In other packages I have seen references to M-equation probit models.

Who is online

Users browsing this forum: No registered users and 1 guest