Understanding information criteria for selection among capture-recapture or ring recovery models

We provide background information to allow a heuristic understanding of two types of criteria used in selecting a model for making inferences from rin ging data. The first type of criteria (e.g. AIC, AIC(c), QAIC(c) and TIC) a re estimates of: (relative) Kullback-Leibler information or distance and at tempt to select a good approximating model for inference, based on the prin ciple of parsimony. The second type of criteria (e.g. BIC, MDL, HQ) are 'di mension consistent' in that they attempt to consistently estimate the dimen sion of the true model. These latter criteria assume that a hue model exist s, that if is in the set of candidate models and that the goal of model sel ection is to find the true model, which in turn requires that the sample si ze is very large. The Kullback-Leibler based criteria do not assume a true model exists, let alone that if is in the set of models being considered. B ased on a review of these criteria, we recommend use of criteria that are b ased on Kullback-Leibler information in the biological sciences.