On the use of directed acyclic graphs in behavioural ecology and evolution

Mirjam Borger , Aparajitha Ramesh

Directed acyclic graphs (DAGs) are graphical models to visualise hypotheses. DAGs are generally used in the field of causal inference and their use is spreading across different fields. However, in biology and especially in behavioural ecology and evolution, DAGs are still underutilised. Here, we point out why DAGs are such useful tools for these fields. Using concrete examples, we demonstrate that including DAGs in empirical studies is helpful for summarising all the important underlying assumptions about the ecology of the study species. With that, DAGs increase the readability and transparency of papers, which could help solve the replication crisis. Moreover, it makes the work of reviewers and meta-analysis researchers easier. Lastly, DAGs can be used to make researchers aware of bad controls and help them to explicitly think through the relationship between variables and their inclusion in statistical models. With this paper, we hope to encourage all biologists to include DAGs in their empirical papers.