Parametric Empirical Bayes Methods for Ecological Applications

This paper reviews parametric empirical Bayes (PEB) methods and their application to ecological data. One of the main objections to Bayesian statistics is the subjectivity caused by specifying parameters. Attempting to remove that subjectivity, the PEB philosophy uses the Bayesian mathematical framework, but then estimates, rather than specifies, all unknown parameters. Three examples are given to illustrate PEB methods. The first is a PEB analysis for monitoring the yearly population of harbor seals, the second uses simulated data and shows how a PEB analysis combines single capture-recapture estimates, and the third is a PEB method for spatial patch analysis of data from a transect through grassland vegetation. From the examples, the general methodology and philosophy of a PEB analysis are explained. See full-text article at JSTOR