University of North Carolina at Chapel HillAbstract/ManuscriptAlexander BreskinApproval & CommercializationYesR&D
DAG With Omitted Objects Displayed (DAGWOOD): a framework for revealing causal assumptions in DAGs.
DAG With Omitted Objects Displayed (DAGWOOD): a framework for revealing causal assumptions in DAGs
Challenge
Directed acyclic graphs are widely used to represent causal assumptions in epidemiologic research, but the implicit background assumptions and omitted variables in a DAG often go undisclosed, creating opportunities for hidden causal assumptions to bias study design and analysis.
Solution
Target RWE researchers introduced the DAGWOOD framework—a structured approach for revealing and documenting the causal assumptions typically hidden in conventional DAGs, including omitted common causes, mediators, and selection nodes.