A General Identification Condition for Causal Effects

Abstract

This paper concerns the assessment of the effects of actions or policy interventions from a combination of: (i) nonexperimental data, and (ii) substantive assumptions. The assumptions are encoded in the form of a directed acyclic also called graph, in which some variables are presumed to be unobserved. The paper establishes a necessary and sufficient criterion for the identifiability of the causal effects of a singleton variable on all other variables in the model, and a powerful sufficient criterion for the effects of a singleton variable on any set of variables.

Cite

Text

Tian and Pearl. "A General Identification Condition for Causal Effects." AAAI Conference on Artificial Intelligence, 2002. doi:10.5555/777092.777180

Markdown

[Tian and Pearl. "A General Identification Condition for Causal Effects." AAAI Conference on Artificial Intelligence, 2002.](https://mlanthology.org/aaai/2002/tian2002aaai-general/) doi:10.5555/777092.777180

BibTeX

@inproceedings{tian2002aaai-general,
  title     = {{A General Identification Condition for Causal Effects}},
  author    = {Tian, Jin and Pearl, Judea},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2002},
  pages     = {567-573},
  doi       = {10.5555/777092.777180},
  url       = {https://mlanthology.org/aaai/2002/tian2002aaai-general/}
}