Statistical Model Adequacy, Goodness-of-Fit, and Model Selection: A Theoretical Review
摘要
The concepts of model adequacy, model goodness-of-fit, and model selection are often mischaracterised, likely due to being ineluctably intertwined. As a result, some descriptions of these ideas are imprecise and/or incomplete in ecological literature. The aim of this chapter is to foster a nexus between the theoretical concepts of model adequacy, goodness-of-fit, and selection, and their application to complex problems in statistical ecology. We will facilitate a progression from novice to expert knowledge by (i) unambiguously defining model adequacy, model goodness-of-fit, and model selection from a statistical standpoint, (ii) emphasising the importance of understanding the integration of these concepts and their differences, with the main associated benefit being obtaining more reliable biological inferences in ecological research, and (iii) succinctly outlining what each analytical step entails to help eradicate future misuse.