How to use directed acyclic graphs (DAGs), a guide for clinical researchers

Directed Acyclic Graphs are commonly used to illustrate and assess the hypothesised causal relationships in health and social research. They can illuminate an investigator’s assumptions and help clearly describe each possible explanation for associations observed in data given the researcher’s assumptions, ranging from causal relationships to confounding and selection bias, and thereby aid in identifying variables that can be used to reduce them. We explain how to construct, interpret, and present DAGs as part of clinical research studies and how they can help communicate a study's strengths or limitations.