What Should We Measure Next? Finding Identification Strategies by Refining Mechanisms
Abstract
Canonical approaches in causal inference treat model specification as fixed, assuming that researchers directly translate all relevant domain knowledge into a causal model, which can then be used to deduce its logical implications. Yet, in practical applications, model specification is often an iterative process, and involves exploration and introspection: of the many aspects of the phenomenon one could investigate, which ones actually matter for the identification of the causal effect of interest? In this paper we study the problem of iterative identification in partially specified causal models. We focus on determining where observing variables that intercept a direct effect or a confounding path between two variables could enable identification in semi-Markovian models. We give necessary and sufficient graphical conditions for when observing such variables can render an unidentifiable query identifiable, together with an efficient algorithm for locating all such opportunities in a given causal diagram. Our results can help analysts better navigate the model space by drawing attention to the parts of their substantive knowledge that could result in a successful identification strategy.
Related
- Finding New Limit Points of Mahler Measure by Methods of Missing Data Restoration
- When the Same Prompt Gives Different Answers (and the Same Answer Isn't Always Right): A Taxonomy of Failure Mechanisms, Testing, and Mitigation Strategies for LLM-Assisted Evidence Synthesis Read-and-Evaluate Tasks
- Refining invariants of finite groups with class functions
- Refining GLSDC for Orbit Determination
- Characterizing k-Naples parking functions and refining bounds via obstructed parking functions
BibTeX
@misc{vaananen2026what,
title = {{What Should We Measure Next? Finding Identification Strategies by Refining Mechanisms}},
author = {Mikko Väänänen and Fanyu Cui and Carlos Cinelli and Santtu Tikka},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.08092},
eprint = {2610.08092},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.08092},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - What Should We Measure Next? Finding Identification Strategies by Refining Mechanisms
AU - Mikko Väänänen
AU - Fanyu Cui
AU - Carlos Cinelli
AU - Santtu Tikka
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.08092
UR - https://arxiv.org/abs/2610.08092
AB - Canonical approaches in causal inference treat model specification as fixed, assuming that researchers directly translate all relevant domain knowledge into a causal model, which can then be used to deduce its logical implications. Yet, in practical applications, model specification is often an iterative process, and involves exploration and introspection: of the many aspects of the phenomenon one could investigate, which ones actually matter for the identification of the causal effect of interest? In this paper we study the problem of iterative identification in partially specified causal models. We focus on determining where observing variables that intercept a direct effect or a confounding path between two variables could enable identification in semi-Markovian models. We give necessary and sufficient graphical conditions for when observing such variables can render an unidentifiable query identifiable, together with an efficient algorithm for locating all such opportunities in a given causal diagram. Our results can help analysts better navigate the model space by drawing attention to the parts of their substantive knowledge that could result in a successful identification strategy.
ER -