Hi Svend,
I used stset, but in your case why did you stset testdate, as opposed to enddate? I thought the stset 'timevar' is supposed to represent the end of the study period (i.e. the point at which failure occured or the subject was no longer being observed)?
In any case, I used stset with my data and I am almost certain I used the correct command. However, when I look at the output I see "probable error" and the number of failures being reported doesn't match with what I know of my data.
Also, you are right. In my case the failure event is determined based on measurements that are taken on my subjects periodically (not pre-determined). How does that change the situation?
I used stset, but in your case why did you stset testdate, as opposed to enddate? I thought the stset 'timevar' is supposed to represent the end of the study period (i.e. the point at which failure occured or the subject was no longer being observed)?
In any case, I used stset with my data and I am almost certain I used the correct command. However, when I look at the output I see "probable error" and the number of failures being reported doesn't match with what I know of my data.
Also, you are right. In my case the failure event is determined based on measurements that are taken on my subjects periodically (not pre-determined). How does that change the situation?
Comment