by Lyndon » Wed Mar 23, 2016 9:39 pm
I'm trying to do something about GOF in a robust design context. This means collapsing over secondary samples within primaries. The "Build groups for trap analysis (Fletcher)" treatment (p.14) uses data not collapsed to 0,1 but to the number of times each individual was captured within each primary. It makes sense to me not to waste these data (see Fletcher 1994 from refs). To do this I first ran "Gather occasions" (p.12), then "Fletcher" (p.14) then the GOF tests. It all seemed to work fine. Is this the way to do it? Your comments are very welcome.