Influence of groups of points on estimates of the parameters of moult

Abstract

The Underhill–Zucchini model of avian moult is a statistical simplification for the analysis of primary moult, a complex biological system in birds. As with all such models, conclusions can be biased if a subset of the data used to fit the model is strongly influential on the results. In the analysis of primary moult of a species, it is standard practice to assemble as many records as possible, frequently from different years or localities. We developed an approach, based on the concepts underpinning Cook’s distance in regression, to evaluate the influence on the estimated moult parameters of groups of records that differ from the bulk of the data. The method is applied to a simulated dataset.