Confounding Equivalence in Causal Inference
Abstract
The paper provides a simple test for deciding, from a given causal diagram, whether two sets of variables have the same bias-reducing potential under adjustment. The test requires that one of the following two conditions holds: either (1) both sets are admissible (i.e., satisfy the back-door criterion) or (2) the Markov boundaries surrounding the manipulated variable(s) are identical in both sets. Applications to covariate selection and model testing are discussed.
- Publication:
-
arXiv e-prints
- Pub Date:
- March 2012
- DOI:
- 10.48550/arXiv.1203.3505
- arXiv:
- arXiv:1203.3505
- Bibcode:
- 2012arXiv1203.3505P
- Keywords:
-
- Statistics - Methodology;
- Computer Science - Artificial Intelligence
- E-Print:
- Appears in Proceedings of the Twenty-Sixth Conference on Uncertainty in Artificial Intelligence (UAI2010)