Summary comments on analyses of the island closure experiment

• Main Author: Doug Butterworth (10005953)
• Format: Report
• Language: unknown
• Published: 2021
• Subjects: Biological Mathematics; Applied Statistics; Biostatistics; penguin island closure experiment; penguin analyses; summary characteristics; penguin; penguins; Penguin Island
• Online Access: https://doi.org/10.25375/uct.13669637.v1

Based on a summary of issues raised in recently submitted documents, the conclusion is drawn that existing analyses based on annually-aggregated data are fully acceptable for consideration in developing management recommendations regarding possible future island closures., but that those based on individual data are not. The reason for this last statement is that results based on a methodology which an unchallenged proof has shown to be flawed are necessarily considered to be unreliable. It is also suggested that sufficient analyses of existing data using the estimation model of the 2016 Panel Algorithm are available to allow for proceeding towards formulation of recommendations. However, discussion is first needed on the response variables to be considered, reconciliation of apparent conflicts amongst some of these, and the criteria/considerations to be taken into account in developing those recommendations. In future, an important prerequisite will be a re-emphasis of the need to follow agreed protocols when engaged in a comparative analysis exercise such as this.