Object Conditioning for Causal Inference

David Jensen, Javier Burroni, Matthew Rattigan
Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, PMLR 115:1072-1082, 2020.

Abstract

We describe and analyze a form of conditioning that is widely applied within social science and applied statistics but that is virtually unknown within causal graphical models. This approach, which we term object conditioning, can adjust for the effects of latent confounders and yet avoid the pitfall of conditioning on colliders. We describe object conditioning using plate models and show how its probabilistic implications can be explained using the property of exchangeability. We show that several seemingly obvious interpretations of object conditioning are insufficient to describe its probabilistic implications. Finally, we use object conditioning to describe and unify key aspects of a diverse set of techniques for causal inference, including within-subjects designs, difference-in-differences designs, and interrupted time-series designs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v115-jensen20a, title = {Object Conditioning for Causal Inference}, author = {Jensen, David and Burroni, Javier and Rattigan, Matthew}, booktitle = {Proceedings of The 35th Uncertainty in Artificial Intelligence Conference}, pages = {1072--1082}, year = {2020}, editor = {Adams, Ryan P. and Gogate, Vibhav}, volume = {115}, series = {Proceedings of Machine Learning Research}, month = {22--25 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v115/jensen20a/jensen20a.pdf}, url = {https://proceedings.mlr.press/v115/jensen20a.html}, abstract = {We describe and analyze a form of conditioning that is widely applied within social science and applied statistics but that is virtually unknown within causal graphical models. This approach, which we term object conditioning, can adjust for the effects of latent confounders and yet avoid the pitfall of conditioning on colliders. We describe object conditioning using plate models and show how its probabilistic implications can be explained using the property of exchangeability. We show that several seemingly obvious interpretations of object conditioning are insufficient to describe its probabilistic implications. Finally, we use object conditioning to describe and unify key aspects of a diverse set of techniques for causal inference, including within-subjects designs, difference-in-differences designs, and interrupted time-series designs.} }
Endnote
%0 Conference Paper %T Object Conditioning for Causal Inference %A David Jensen %A Javier Burroni %A Matthew Rattigan %B Proceedings of The 35th Uncertainty in Artificial Intelligence Conference %C Proceedings of Machine Learning Research %D 2020 %E Ryan P. Adams %E Vibhav Gogate %F pmlr-v115-jensen20a %I PMLR %P 1072--1082 %U https://proceedings.mlr.press/v115/jensen20a.html %V 115 %X We describe and analyze a form of conditioning that is widely applied within social science and applied statistics but that is virtually unknown within causal graphical models. This approach, which we term object conditioning, can adjust for the effects of latent confounders and yet avoid the pitfall of conditioning on colliders. We describe object conditioning using plate models and show how its probabilistic implications can be explained using the property of exchangeability. We show that several seemingly obvious interpretations of object conditioning are insufficient to describe its probabilistic implications. Finally, we use object conditioning to describe and unify key aspects of a diverse set of techniques for causal inference, including within-subjects designs, difference-in-differences designs, and interrupted time-series designs.
APA
Jensen, D., Burroni, J. & Rattigan, M.. (2020). Object Conditioning for Causal Inference. Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, in Proceedings of Machine Learning Research 115:1072-1082 Available from https://proceedings.mlr.press/v115/jensen20a.html.

Related Material