Time Intervals Vary Between Groups

[KristenHeinStrohm]

I know how to set up unequal time intervals in Mark if all groups in the study share those intervals. Currently, though, I'm working with 3 sites that were sampled in different years. Site 1 was sampled every year between 2001-9, site 2 was sampled in 2002-3 and 2008-9, and site 3 was sampled in 2004-9. CJS comparisons between sites are of biological and management interest for the years in which all sites were sampled, and trends over time within and between sites are also of interest. Would it be correct for me to enter annual intervals in the setup page and model the site-specific interval differences in the design matrix with zeros for the unsampled sites & years? Or is there a way that I'm missing on the setup page to distinguish interval length between groups? Thank you in advance for your guidance.

[jlaake]

Define largest set of time intervals 2001-2009 and then set p=0 (fixed parameter values not in Design Matrix) for occasions in which some groups are not sampled. Your survival will be annual values but you won't be able to estimate each annual survival for each site. You can restrict that by assigning the same index number to Phi entries where there is no intervening occasions with p>0. Alternatively, you can set some Phi's to 1 and then let the other Phi in the interval absorb the multi-year survival. I don't recommend that as it gets tricky to model covariate effects on Phi. Better off having them all be annual rates with some restricted to be the same. For example, for group 1 time-specific PIM for Phi could be 1 to 8 for survival in intervals between 2001-2002,2002-2003,...,2008-2009. Then for group site 2, you could have 10,11,12,12,12,12,12,13 (where Phi(10) fixed to 1 or 0 because there are no data). For site 3, it would be 10,10,10,14,15,16,17,18, again with Phi(10) fixed. Where the PIM values are equal you can estimate a common rate across those years but not a year-specific rate for those years at that site.

--jeff

[KristenHeinStrohm]

Thank you! I forgot about the option to fix parameters in the run window. This makes perfect sense. I appreciate your time!