Nest Survival: Unbalanced data sets between 2 sites

[rjfriendly2]

Hello,

I am trying to estimate the daily survival rate (DSR) using the nest survival module in program MARK implemented through the RMARK package in program R. I am fairly new to creating nest survival models and came across a problem with one of my models. I had a simple model DSR~site+year+site*year (year site interaction model) and the number of parameters I got didn't look right.

To give background information of my data. In one site, (A), I have nest data from 1994-2015 (22yrs), on the other site, (B), I have data from 2010-2018 (9yrs). In my INP file, I have 'SITE' as a categorical variable that denotes A=1 and B=0. 'YEAR' is set as a factor or group variable and spans from 1994-2018, BUT nests from site A only go up to 2015 and nests from site B go from 2010-2018.

After running the year*site interaction model I saw in the model output that there were estimates for site A that went to year 2018 (even there was no data for it), and estimates from 1994-2009 for site B (again, event there is no data for it). I've been trying to figure out how to navigate around this or find the right way to get these estimates. Can anyone help me this? Thanks!

[jlaake]

Use ~-1+site:year. This is similar to age-year example in documentation and formula for Psi transitions where stratum and tostratum are incompletely crossed.