I am trying to build a discrete time event model. The event occurs once in 20 years period of time. Most of the example i saw use only covariates which are same over period of time. e.g. sex,race etc.
But i have observed values of covariates for different time periods.And there are around 10-15 covariates which are observed for 20 years.
I want to model all the time varying covariates for each year. e.g. in SAS the option is like: model event=var1 var2 var3 number_of_years
Here they have var1 var2 as race, sex etc, which are constant over period of time(time independent). But i have like different var1 values observed for number of years. So for second entity, its again different for number of years.
The event occurs once for given number of years in 1/0 form
I am using SAS, but can use R as well if that helps.
Please let me know, if you have any questions.
To clarify further:
Please refer to: http://ccsr.ac.uk/qmss/summer/Paris09/documents/SingerWillet.pdf
in the last you can see logistic regression code.
I want to know how to include predictors like cash value along with dummy variables D01-D12 which are varying for 12 time periods.