Phylogenetic ANOVA: Group‐clade aggregation, biological challenges, and a refined permutation procedure

• DOI: 10.1111/evo.13492
• Publication Date: 2018-04-23T05:26:25Z
• Authors: Dean C. Adams; Michael L. Collyer
• Keywords: 0106 biological sciences; Analysis of Variance; Data Interpretation, Statistical; Multivariate Analysis; Animals; Urodela; 01 natural sciences; Phylogeny

Phylogenetic regression is frequently used in macroevolutionary studies, and its statistical properties have been thoroughly investigated. By contrast, phylogenetic ANOVA has received relatively less attention, the conditions leading to incorrect biological inferences when comparing multivariate phenotypes among groups remain underexplored. Here, we propose a refined method of randomizing residuals permutation procedure (RRPP) for evaluating phenotypic differences while conditioning data on phylogeny. We show that RRPP displays appropriate both models, univariate datasets. For ANOVA, find exhibits higher power than methods utilizing simulation. Additionally, investigate how group dispersion across phylogeny affects inferences, reveal highly aggregated generate strong significant correlations with phylogeny, which reduce subsequently affect interpretations. discuss broader implications this aggregation, relation challenges encountered other comparative where one or few transitions discrete traits are observed Finally, recommend studies continuous trait use assessing significance indicator variables as sources variation.