AbstractTo investigate how biodiversity arose, the field of macroevolution largely relies on model-based approaches to estimate rates of diversification and what factors influence them. The number of available models is rising steadily, facilitating the modeling of an increasing number of possible diversification dynamics, and multiple hypotheses relating to what fueled or stifled lineage accumulation within groups of organisms. However, growing concerns about unchecked biases and limitations in the employed models suggest the need for rigorous validation of methods used to infer. Here, we address two points: the practical use of model adequacy testing, and what model adequacy can tell us about the overall state of diversification models. Using a large set of empirical phylogenies, and a new approach to test models using aspects of tree shape, we test how a set of staple models performs with regards to adequacy. Patterns of adequacy are described across trees and models and causes for inadequacy – particularly if all models are inadequate – are explored. The findings make clear that overall, only few empirical phylogenies cannot be described by at least one model. However, finding that the best fitting of a set of models might not necessarily be adequate makes clear that adequacy testing should become a step in the standard procedures for diversification studies.
TEST SCENARIO SPECIFICATION LANGUAGE FOR MODEL-BASED TESTING
