Detecting Confounding in Multivariate Linear Models via Spectral Analysis
Open Access
- 28 October 2017
- journal article
- research article
- Published by Walter de Gruyter GmbH in Journal of Causal Inference
- Vol. 6 (1)
Abstract
We study a model where one target variable Y $Y$ is correlated with a vector X := ( X 1 , … , X d ) $\textbf{X}:=(X_1,\dots,X_d)$ of predictor variables being potential causes of Y $Y$ . We describe a method that infers to what extent the statistical dependences between X $\textbf{X}$ and Y $Y$ are due to the influence of X $\textbf{X}$ on Y $Y$ and to what extent due to a hidden common cause (confounder) of X $\textbf{X}$ and Y $Y$ . The method relies on concentration of measure results for large dimensions d $d$ and an independence assumption stating that, in the absence of confounding, the vector of regression coefficients describing the influence of each X $\textbf{X}$ on Y $Y$ typically has ‘generic orientation’ relative to the eigenspaces of the covariance matrix of X $\textbf{X}$ . For the special case of a scalar confounder we show that confounding typically spoils this generic orientation in a characteristic way that can be used to quantitatively estimate the amount of confounding (subject to our idealized model assumptions).
Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Latent variable graphical model selection via convex optimizationThe Annals of Statistics, 2012
- Estimation of causal effects using linear non-Gaussian causal models with hidden variablesInternational Journal of Approximate Reasoning, 2008
- Causality: Models, Reasoning and InferencePhilosophical Review, 2001
- Random Vectors in the Isotropic PositionJournal of Functional Analysis, 1999
- Modeling of strength of high-performance concrete using artificial neural networksCement and Concrete Research, 1998
- Rank One Perturbations, Approximations, and Selfadjoint ExtensionsJournal of Functional Analysis, 1997
- Rank One Perturbations with Infinitesimal CouplingJournal of Functional Analysis, 1995
- Perturbation theory for semigroups of operators1995
- Classes of orderings of measures and related correlation inequalities. I. Multivariate totally positive distributionsJournal of Multivariate Analysis, 1980
- An Inverse Matrix Adjustment Arising in Discriminant AnalysisThe Annals of Mathematical Statistics, 1951