survival calculation: unequal time intervals with covariates

[Eurycea]

I would like to know how/if MARK adjusts survival estimates when covariates are used for a dataset with unequal time intervals. For example, take ragged time intervals 1 2 1 2: phi can be adjusted as phi^1, phi^2, phi^1, phi^2. Let's call the adjusted vector "phi.s". What happens when a covariate is involved? Is the expression

Code: Select all

logit(phi) = linear predictor

or

Code: Select all

logit(phi.s) = linear predictor

?

I'm partly asking because I'm not only try to understand what MARK does, but also to code the analogous model in BUGS. Thanks.

[gwhite]

The value of phi is computed from the expit of the linear equation, then raised to the appropriate power for the time interval.

Gary