I am using the program Mark via package ‘RMark’ in R to analyze the nest survival data of bird nests with the model 'Nest'.
In my analysis, I considered a covariate "researcher's nest visit frequency", which calculated by the number of visits to the nest divided by the total number of days that the nest was monitored during a given breeding stage (like the egg stage or nestling stage). But one of the reviewers thinks it should be coded as time-dependent.
Reviewer comments:
(The nest survival model as implemented in MARK estimates survival probability for each time unit (here nest day) under observation as a function of covariates. If these covariates are naturally not constant, they should be coded as time-dependent, which is quite easy. The problem with the present analysis is that the covariate of primary interest – frequency of nest visits – is being treated as a nest-level covariate. This may, in my opinion, obscure its real effect on DSR. In the present analysis, the DSR for a particular day is being modelled as a function of “mean” visit frequency, i.e. including also future (with respect to given day) visits. I suggest entering visit frequency as a time-dependent covariate, in which case the DSR for a particular day would be modelled as a function of frequency of previous visits (visit frequency calculated from the number of visits and number of observation days up to this day), which is biologically more sensible. )
I considered it need to add a covariate of observer effects in the analysis, but I don't know how to do it. I know this can be achieved in the program Mark by adding a covariate like VisitDay1, VisitDay2,... I want to know can I achieve this step in R by package ‘RMark’?