Input file for Pradel model

[nperlut]

Hello all,

I am using Pradel models to look at survival and recruitment. I have ten sampling locations - of which, half have four years of data and half have three years of data (the '3-year' sites began on the second year of the study). I am wondering about the input file - as the capture history for these sites with only three years of data all begin with 0, which is not neccesarily accurate. This way all sites have four years of capture history - unforutantely, a missing value "." for the three year sites is not an option.

In fact, many of these birds were probably present, but we simply were not working on the field. Will these 0's artificially inflate the recruitment or detection probability estimates? For example, the history 0111 for a site with four years of data suggests that the bird was recruited into the population during the second year (or present and missed the first year, or alive but not present the first year). However, when the same 0111 history represents a field with only three years of data, then the bird in fact may not have been a recruit (because we were not present the previous year to sample), but is otherwise being treated as a recruit (or the other described situations)- when included in a four year data set. I am not entire sure how to address this situation without throwing away the first year of data for the four year set..
Any thoughts would be greatly appreciated.

noah

[gwhite]

You should fix p1 for the unsampled group to zero, so that the correct likelihood is generated.

Gary