I analyse data on bahaviour of crayfish with PHREG, and I encounter problems if a time-dependent covariate is added. I observed a number of crayfish over 30 minutes, and recorded time after which they were leaving their shelter (0 (censored) if they did not leave after 30 minutes; 1 - success if they left their shelter) in 2 groups (hereafter "treatment"). Water in tanks changed its quality after 15 minutes (time-dependent state - time dependent variable "status" with two levels 0 and 1).
I used the following model:
proc phreg data=tak;
model time*censored(0)= treatment status;
if time < 15 then status= 0.0;
else status= 1.0;
Something goes wrong with this procedure, as I obtain 0 values (results table) for the effect of "status" variable. I should be grateful for your suggestions what can be wrong. Maybe something with my SAS program or with my usage of this procedure?