Hi,
I am hoping to get some advice on the proper way to format data input files for Mark/RMark that censor individuals based on certain criteria. Specifically, I am using multistate models to estimate survival and transcience probabilities 5 different geographical states. Using acoustic telemetry, I've tagged/detected 135 individuals across a 2-year period (25 sampling occasions).
I would like to censor individuals on a specific encounter occasion if the tag battery life has expired (according the the manufacturer's estimated tag life).
I know this is normally done by setting the frequency to -1 which tells Mark that after the last encounter (or detection) the fish should be removed from the study. However, the problem I am running into is that there are true absences (0's) prior to the encounter occasion that should be censored so I'm not sure how to distinguish these true 0's from the 0's that should be censored.
For example, assume I have 5 sampling occasions across 3 strata and a transmitter expires on the fourth occasion, then the encounter history would look something like,
A0000
where the first two 0's are true absences (the fish was not detected) and the last two zeroes are absences as a result of the tag expiring. If I assign a frequency of -1 to this encounter history, then the model will ignore everything after the initial encounter (the first two zeroes will not enter the model likelihood although they are true values and should be treated as such).
Can anyone recommend the best way to approach this?