The problem of mooted models for analyses of microbiome causality

Abstract

Lynch, Parke, and O’Malley highlight the need for better evaluative criteria for causal explanations in microbiome research. They propose new interventionist criteria, show that paradigmatic examples of microbiome explanations are flawed using those criteria, and suggest numerous ways microbiome explanations can be improved. While we endorse their primary criticisms and suggestions for improvements in microbiome research, we make several observations regarding the use of mooted causal models in microbiome research that have significant implications for their overall argument. In sum, we contend that their critique is too modest and that even flawed causal inferences like those they criticize can be valuable for generating better causal models and evaluating explanatory outcomes in individual cases.