Mark-Resight Data Types

Contents – Index

Mark-Resight Data Types

Full details of these models are given in the PDF file in Chapter 18 of the Gentle Introduction, where proper formatting of mathematical symbols is provided.  What follows below is taken from this file, but much of the math has been removed.  If you cannot access this file directly by clicking on the above link, copy the URL to your browser.

Three additional mark-resight models have been added that are analogous to the original 3 data types, but which now model the probability of determining the identity of a marked animal when it is observed.  This additional parameter, r, is the probability that the identity is determined, and is in addition to the original parameters.  The original method to incorporate unidentified marks was ad-hoc (and not particularly satisfying) and does not hold (i.e., biased and too precise) if there are a decent amount of these.

When all marked animals are identified, the original model and the new model produce identical results with r = 1.  However, the likelihoods are not comparable when r is not estimated to be 1.