Search arXivSearch

arXiv · 2502.10317

A Framework for Covariate-Adjusted Bivariate Causal Discovery

Abstract

Ascertaining causal direction from observational data is a fundamental challenge in scientific inquiry. Of particular interest is the problem of covariate-adjusted bivariate causal discovery, i.e., determining the causal direction between X and Y in the presence of Z. While unadjusted bivariate causal discovery has seen significant advances (Hoyer et al., 2008; Ni, 2022), there is a lack of methodology dealing with real-world bivariate relationships, which are often modulated by a set of covariate(s), Z. Building on previous work in Purkayastha and Song (2025), we introduce a novel, nonparametric framework for the covariate-adjusted bivariate causal discovery problem and propose a conditional asymmetry coefficient to track said direction of causation. We develop a robust estimation procedure using kernel-based conditional density estimation with cross-fitting and also provide rigorous uncertainty quantification using a nonparametric kernel smoothing technique, addressing a key limitation of many existing algorithms. As a key application, we demonstrate how this framework can be applied to the problem of collider detection, a persistent challenge in causal structure learning. Simulation studies show our method's superior performance in identifying causal structures. We apply our approach to an epigenetic study (Perng et al., 2019), investigating the role of blood pressure in regulating the effects of DNA methylation. In summary, our work offers methodological advancement by providing a robust, inferential toolkit for dissecting complex, moderated bivariate causal relationships in observational data.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Soumik Purkayastha, Peter X. -K. Song. 2025-07-28. A Framework for Covariate-Adjusted Bivariate Causal Discovery. https://arxiv.org/abs/2502.10317

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Switchback Experiments under Geometric Mixing

The switchback is an experimental design that measures treatment effects by repeatedly turning an intervention on and off for a whole system. Switchback experiments are a robust way to overcome cross-unit spillover effects; however, they are vulnerable to bias from temporal carryovers. In this paper, we consider properties of switchback experiments in Markovian systems that mix at a geometric rate. We find that, in this setting, standard switchback designs suffer considerably from carryover bias: Their estimation error decays as $T^{-1/3}$ in terms of the experiment horizon $T$, whereas in the absence of carryovers a faster rate of $T^{-1/2}$ would have been possible. We also show, however, that judicious use of burn-in periods can considerably improve the situation, and enables errors that decay as $\log(T)^{1/2}T^{-1/2}$. Our formal results are mirrored in an empirical evaluation.

stat.ME

Saddlepoint approximations for plug-in resampling

Resampling-based procedures can improve on normal approximations in sparse, large-scale testing problems, but their computational cost can be prohibitive. We recognize that several existing procedures belong to a faster plug-in resampling subclass, fixing fitted nuisance parameters during resampling. When the resampled statistic is a sum of conditionally independent terms, the saddlepoint approximation (SPA) for the resampling $p$-value offers further acceleration, replacing resampling with an analytical tail approximation. However, standard Edgeworth-based approximation-error bounds impose regularity conditions that are hard to verify for plug-in resampling laws. We use an alternative approach to establish a finite-sample relative-error bound for the Lugannani-Rice approximation under more tractable conditions, which we apply in two contexts. In statistical genetics, we identify response resampling procedures as the targets of existing SPAs and establish guarantees in a representative setting. In conditional independence testing, we introduce spaCRT, an SPA for the distilled conditional randomization test (dCRT), which has been applied successfully in biology. Our rates quantify the effects of sparsity and signal strength, with matching lower bounds in special cases. We additionally establish asymptotic Type-I error control of the corresponding plug-in resampling procedures under growing sparsity. In simulations and single-cell CRISPR data analysis, spaCRT closely approximates dCRT $p$-values and preserves its statistical performance while accelerating computation by up to 250-fold.

stat.ME

NIRVAR: Network Informed Restricted Vector Autoregression

High-dimensional panels of time series often arise in finance and macroeconomics, where co-movements within groups of panel components occur. Extracting these groupings from the data provides a coarse-grained description of the complex system in question and can inform subsequent prediction tasks. We develop a novel methodology to model such a panel as a restricted vector autoregressive process, where the coefficient matrix is the weighted adjacency matrix of a stochastic block model. This network time series model, which we call the Network Informed Restricted Vector Autoregression (NIRVAR) model, yields a coefficient matrix that has a sparse block-diagonal structure. We propose an estimation procedure that embeds each panel component in a low-dimensional latent space and clusters the embedded points to recover the blocks of the coefficient matrix. Crucially, the method allows for network-based time series modelling when the underlying network is unobserved. We derive the bias, consistency and asymptotic normality of the NIRVAR estimator. Simulation studies suggest that the NIRVAR estimated embedded points are Gaussian distributed around the ground truth latent positions. On three applications to finance, macroeconomics, and transportation systems, NIRVAR outperforms competing models in terms of prediction and provides interpretable results regarding group recovery.

stat.ME