Anytime-Valid Inference in Linear Models with Clustered Outcomes: Exact Multivariate Tests and Cluster-Robust Confidence Sequences

Michael Lindon

Home/Archive/arXiv

arXiv · 2610.25613

Anytime-Valid Inference in Linear Models with Clustered Outcomes: Exact Multivariate Tests and Cluster-Robust Confidence Sequences

Michael Lindon

View original source ↗Read source document ↗Export metadata ↓

Abstract

We extend anytime-valid inference for linear models from independent scalar outcomes to units that contribute a vector of dependent outcomes. We develop a parametric and a nonparametric treatment. For the Gaussian multivariate linear model with an unstructured error covariance, we derive a closed-form Bayes factor under the right-Haar prior and a matrix-normal mixture alternative and prove that it is an exact e-value at every sample size and a test martingale in the invariant filtration. To our knowledge, this Bayes factor is new in the generality considered here: a d x q coefficient matrix, nuisance regressors and an arbitrary fixed positive-definite prior precision under the right-Haar nuisance prior. It yields sequential Hotelling and MANOVA tests with exact confidence sequences. Under a common within-cluster design, the cluster-robust Wald statistic of a pooled regression is equal to Hotelling's T^2 statistic computed from the per-cluster least-squares summaries, giving an exact Gaussian benchmark with G-1 degrees of freedom. For general clustered population least-squares projections, the corresponding sandwich statistic is an asymptotic e-process, without Gaussian, common-design or linearity assumptions. Each treatment includes growth and first-order expected stopping-time results; under stronger regularity, direct comparisons of expected log evidence identify a fixed mixture precision that maximizes expected log growth through constant order within a specified family, for a specified frequentist alternative. Applied to randomized experiments with several outcomes, the robust procedure gives asymptotically anytime-valid joint tests, confidence ellipsoids and simultaneous confidence sequences for the vector of average treatment effects from ordinary least-squares estimates and cluster-robust standard errors.

Explore related subjects

stat.ME math.ST stat.TH

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Michael Lindon. 2026-10-19. Anytime-Valid Inference in Linear Models with Clustered Outcomes: Exact Multivariate Tests and Cluster-Robust Confidence Sequences. https://arxiv.org/abs/2610.25613

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

KEEP EXPLORING

Related papers

arXiv2026

Estimation of Treatment Effects Under Nonstationarity via the Truncated Policy Gradient Estimator

Randomized experiments (or A/B tests) are widely used to evaluate interventions in dynamic systems such as recommendation platforms, marketplaces, and digital health. In these settings, interventions affect both current and future system states, so estimating the global average treatment effect (GATE) requires accounting for temporal dynamics, which is especially challenging in the presence of nonstationarity; existing approaches suffer from high bias, high variance, or both. In this paper, we address this challenge via the novel Truncated Policy Gradient (TPG) estimator, which replaces instantaneous outcomes with short-horizon outcome trajectories. The estimator admits a policy gradient interpretation: it is a truncation of the first-order approximation to the GATE, yielding explicit bias and variance bounds in nonstationary Markovian settings. For a fixed truncation size, we establish a central limit theorem and, under an additional mean-stability condition, the consistency of a variance estimator for nonstationary single-trajectory data. We validate our theory with two real-world case studies. The results show that relative to existing approaches, a well-calibrated TPG estimator can achieve a favorable balance between bias and variance in nonstationary settings, highlighting the value of the policy-gradient perspective for designing effective estimators under complex dynamics.

stat.ME↗

arXiv2026

Weak Factor Analysis via Temporally Weighted PCA

Principal component analysis (PCA) is arguably the most widely used approach for large-dimensional factor analysis. While it is effective when the factors are sufficiently strong, it can be inconsistent when the factors are weak and/or the noise has complex dependence structure. We argue that the inconsistency often stems from bias and introduce a general approach to restore consistency. Specifically, we propose a general weighting scheme for PCA and show that suitable weights can restore consistency for much weaker factors under the dependent-noise structures allowed by our model. Under the stronger signal-to-noise conditions needed for inference, we also establish asymptotic normality of the loading and factor estimators. While an oracle weighting matrix may require knowledge about the factors and covariance of the idiosyncratic noise that are not known a priori, we develop an agnostic approach to adaptively choose from a large class of weighting matrices that can be viewed as PCA for weighted linear combinations of auto-covariances among the observations. Theoretical and numerical results demonstrate the merits of our methodology over the usual PCA and other recently developed techniques for large-dimensional approximate factor models.

stat.ME↗

arXiv2026

Combining e-values using demi-supermartingales

We present a new method for combining e-variables through demi-supermartingales, which settles an old conjecture in the literature on nonparametric mean testing. It also provides an explicit concentration bound for a certain Kullback--Leibler-type statistic arising in the stochastic multi-armed bandit literature. All of these combination results hold for independent e-variables as well as for the class of co-valid e-variables, whose dependence structure lies somewhere between independence and sequential validity. The results are further generalized to compound e-variables. The proofs proceed by analyzing elementary symmetric polynomials and their behavior as nonnegative demi-supermartingales.

stat.ME↗