What is it about?
This paper introduces the R package dagitty – used to create and work with directed acyclic graphs (DAGs) in the application of causal inference modelling – and explains key features of this package (though there are many more): identifying optimal minimal adjustment sets of confounders; evaluating data-DAG consistency; and deriving DAG ‘equivalence classes’ where arcs in a DAG may be in error but do not affect the implied testable assumptions and provides adjustment sets unaffected by these potential DAG errors.
Featured Image
Why is it important?
Data-DAG evaluation is made simple in this package. It is also poorly understood that a DAG need not be absolutely right for robust causal inference to be achievable - DAG equivalence classes provide an improved selection of minimal adjustment sets that are robust to uncertainties in the DAG.
Perspectives
Read the Original
This page is a summary of: Robust causal inference using directed acyclic graphs: the R package ‘dagitty’, International Journal of Epidemiology, January 2017, Oxford University Press (OUP),
DOI: 10.1093/ije/dyw341.
You can read the full text:
Contributors
The following have contributed to this page