On identifying total effects in the presence of latent variables and selection bias
Published 9 July 2008
Zhihong Cai, Manabu Kuroki
Citations23
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
Abstract
Assume that cause-effect relationships be-tween variables can be described as a directed acyclic graph and the corresponding linear structural equation model．We consider the identification problem of total effects in the presence of latent variables and selection bias between a treatment variable and a response variable. Pearl and his colleagues provided the back door criterion, the front door cri-terion (Pearl, 2000) and the conditional in-strumental variable method (Brito and Pearl, 2002) as identifiability criteria for total ef-fects in the presence of latent variables, but not in the presence of selection bias. In or-der to solve this problem, we propose new graphical identifiability criteria for total ef-fects based on the identifiable factor models. The results of this paper are useful to iden-tify total effects in observational studies and provide a new viewpoint to the identification conditions of factor models. 1
Keywords
Computer ScienceMathematicsDecision Sciences
Wiley series in probability and statisticsLinear Statistical Inference and its Applications
10,509 Citations1973C. Radhakrishna Rao
EpidemiologyA Structural Approach to Selection Bias
2,630 Citations2004Miguel A. Hernán, Sonia Hernández–Dı́az +1 more
This work argues that the causal structure underlying the bias in each example is essentially the same: conditioning on a common effect of 2 variables, one of which is either exposure or a cause of exposure and the other is either the outcome or acause of the outcome.
TechnometricsGraphical Models in Applied Multivariate Statistics
1,827 Citations1991Colin Goodall, Joe Whittaker
EpidemiologyQuantifying Biases in Causal Models: Classical Confounding vs Collider-Stratification Bias
812 Citations2003Sander Greenland
The results suggest that bias from stratifying on variables affected by exposure and disease may often be comparable in size with bias from classical confounding, whereas other biases from collider stratification may tend to be much smaller.
The Annals of StatisticsAncestral graph Markov models
544 Citations2002Thomas S. Richardson, Peter Spirtes
A class of graphical independence models that is closed under marginalization and conditioning but that contains all DAG independence models, called maximal ancestral graphs, which lead to a simple parametrization of the corresponding set of distributions in the Gaussian case.
MIT Press BooksCausation, Prediction, and Search, 2nd Edition
211 Citations2001Peter Spirtes, Clark Glymour +1 more
The authors axiomatize the connection between causal structure and probabilistic independence, explore several varieties of causal indistinguishability, formulate a theory of manipulation, and develop asymptotically reliable procedures for searching over equivalence classes of causal models.
eScholarship (California Digital Library)Identification of Joint Interventional Distributions in Recursive Semi-Markovian Causal Models
155 Citations2006Ilya Shpitser, Judea Pearl
arXiv (Cornell University)Pearl's Calculus of Intervention Is Complete
106 Citations2012Yimin Huang, Marco Valtorta
arXiv (Cornell University)Generalized Instrumental Variables
92 Citations2012Carlos Brito, Judea Pearl
This paper provides a generalization of the well-known method of Instrumental Variables, which allows its application to models with few conditional independeces, and provides a description of domain knowledge encoded in the form of a directed acyclic graph.
BiometrikaMiscellanea. Identification of a single-factor model using graphical Gaussian rules
27 Citations1997Elena Stanghellini
BiometrikaOn the identification of path analysis models with one hidden variable
27 Citations2005Elena Stanghellini, Nanny Wermuth
BiometrikaOn identification of multi-factor models with correlated residuals
27 Citations2004Michel Grzebyk
arXiv (Cornell University)A Bayesian Method for Causal Modeling and Discovery Under Selection
23 Citations2013Gregory F. Cooper
A Bayesian method for combining data under selection with prior beliefs in order to derive a posterior probability for a model of the causal processes that are generating the data in the population of interest.
Identifying linear causal effects
18 Citations2004Jin Tian
The paper shows how techniques developed for identifying causal effects in causal Bayesian networks can be used to identify linear causal effects, and thus provides a new approach for assessing linear causal results in structural equation models.
BiometrikaOn recovering a population covariance matrix in the presence of selection bias
17 Citations2006Manabu Kuroki, Zhihong Cai
The problem of using observational data in the presence of selection bias to identify causal effects in the framework of linear structural equation models is considered and a criterion for testing whether or not observed statistical dependencies among variables are generated by conditioning on a common response variable is proposed.
EnvironmetricsInstrumental variables in Gaussian directed acyclic graph models with an unobserved confounder
11 Citations2004Elena Stanghellini
The Joint Causal Effect in Linear Structural Equation Model and Its Application to Process Analysis
6 Citations2003Manabu Kuroki, Zhihong Cai
Ouyou toukeigakuIdentificability Criteria for Total Effects in the Presence of Unmeasured Confounders
1 Citations2007Manabu Kuroki