Hello,
I'm running some GOF tests in a mark-recapture dataset of bottlenose dolphins. When I runned the tests with the full dataset it turn out as expected with significance for Test3SR (as I have many transients) and, not so expected for Test3SM. To turn this around I left truncated my dataset, eliminating the fist capture of all individuals whitch reduced my dataset considerably (from 300 individuals to 69 and 11 years to 5). When I runned the GOF tests again it only test for transience (Test 3SM), all the other tests were inconclusive. With this, the c^ of the overall test was <1, but I simply don't know if I should be ok with this. The results of the estimates when I runned the truncated dataset in CJS look OK, but with wide CI and a large SE (not very informative). Of course, when I adjust the c^, they get mutch better but I don´t know if in this case I should set c^=1 (as explained in Mark book) and don't adjust at all.
Thanks for the help!