Hello,
I am analyzing the delay, measured by number of days, it took an entity to implement one of three policies, after they detected a problem. Therefore, in my dataset I observe entities after they have detected a problem. Many would not have implemented a policy in the time period observed. I have multiple observations per entity, one per each policy. I am wondering if I should declare my data using the ID function on stset, where I would use the company ID to distinguish between entities. I also wanted to keep the type of policy as a predictor variable, so I am wondering if this would be redundant. Also when I run model using the stcox command, should i be using the strata option, with company ID as the variable?
I am still familiarizing myself with this so any information regarding the above or any relevant information regarding model implementation would be appreciated.
Thanks!
Amit
I am analyzing the delay, measured by number of days, it took an entity to implement one of three policies, after they detected a problem. Therefore, in my dataset I observe entities after they have detected a problem. Many would not have implemented a policy in the time period observed. I have multiple observations per entity, one per each policy. I am wondering if I should declare my data using the ID function on stset, where I would use the company ID to distinguish between entities. I also wanted to keep the type of policy as a predictor variable, so I am wondering if this would be redundant. Also when I run model using the stcox command, should i be using the strata option, with company ID as the variable?
I am still familiarizing myself with this so any information regarding the above or any relevant information regarding model implementation would be appreciated.
Thanks!
Amit
