Modeling nest survival data: a comparison of recently developed methods that can be implemented in MARK and SAS

摘要: Estimating nest success and evaluating factors potentially related to the survival rates of nests are key aspects many studies avian populations. A strong interest in has led a rich literature detailing variety estimation methods for this vital rate. In recent years, modeling approaches have undergone especially rapid development. Despite these advances, most researchers still employ Mayfield’s ad–hoc method (Mayfield, 1961) or, some cases, maximum–likelihood estimator Johnson (1979) Bart & Robson (1982). Such permit analyses stratified data but do not allow more complex realistic models rate that include covariates vary by individual, age, time, etc. may be continuous or categorical. Methods rigorously assess importance biological might affect can now readily implemented Program MARK SAS’s Proc GENMOD NLMIXED. Accordingly, use without first need cannot justified. With goal increasing flexible methods, we describe likelihood used then consider question what effective sample size is computation AICc. Next, advantages disadvantages different programs terms ease input model construction; utility/flexibility generated estimates predictions; selection; ability estimate variance components. An example set analyzed using both SAS demonstrate implementation with various contain nest–, group– (or block–), time–specific covariates. Finally, discuss improvements would, if they became available, promote better general understanding rates.