Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.30019

CAQ: Clustered Asymptotically Anytime-Valid Quantile Treatment Effects for Online Experiments

Diane Lu·Eric J. Oh
View original source ↗Read source document ↗Export metadata ↓

Abstract

Confidence sequences let an experiment be monitored continuously without inflating error, but the existing constructions for quantiles assume that observations arrive independently. In online experiments this rarely holds: randomization assigns users, while the metric is computed over the many correlated events each user generates, so intervals built for independent data are too narrow and miss the truth more often than their nominal rate allows. We propose CAQ, a confidence sequence for a quantile that is both robust to this within-user dependence and valid at arbitrary stopping times, together with a sequential test for the difference in a quantile between two treatment arms. The width adapts to the dependence actually present, widening beyond the independent-data benchmark only when correlation within users is strong. In simulations, CAQ is the only method we tested that holds both coverage and false-positive rate at their nominal levels across clustering strengths and heavy-tailed data, while the fixed-horizon baseline fails under continuous monitoring and the independence-assuming sequential baseline fails as within-user correlation grows.

Explore related subjects

stat.ME

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Diane Lu, Eric J. Oh. 2026-10-21. CAQ: Clustered Asymptotically Anytime-Valid Quantile Treatment Effects for Online Experiments. https://arxiv.org/abs/2610.30019

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

KEEP EXPLORING

Related papers

arXiv2026

A robust multivariate, non-parametric outlier identification method for scrubbing in fMRI

Functional magnetic resonance imaging (fMRI) data contain high levels of noise and artifacts. To avoid contamination of downstream analyses, fMRI-based studies must identify and remove these noise sources prior to statistical analysis. One common approach is the "scrubbing" of fMRI volumes that are thought to contain high levels of noise. However, existing scrubbing techniques are based on ad hoc measures of signal change. We consider scrubbing via outlier detection, where volumes containing artifacts are considered multidimensional outliers. Robust multivariate outlier detection methods are proposed using robust distances (RDs), which are related to the Mahalanobis distance. These RDs have a known distribution when the data are i.i.d. normal, and that distribution can be used to determine a threshold for outliers where fMRI data violate these assumptions. Here, we develop a robust multivariate outlier detection method that is applicable to non-normal data. The objective is to obtain threshold values to flag outlying volumes based on their RDs. We propose two threshold candidates that embark on the same two steps, but the choice of which depends on a researcher's purpose. Our main steps are dimension reduction and selection, robust univariate outlier imputation to get rid of the effect of outliers on the distribution, and estimating an outlier threshold based on the upper quantile of the RD distribution without outliers. The first threshold candidate is an upper quantile of the empirical distribution of RDs obtained from the imputed data. The second threshold candidate calculates the upper quantile of the RD distribution that a nonparametric bootstrap uses to account for uncertainty in the empirical quantile. We compare our proposed fMRI scrubbing method to motion scrubbing, data-driven scrubbing, and restrictive parametric multivariate outlier detection methods.

stat.ME↗
arXiv2026

Optimized Conformal Selection: Powerful Selective Inference After Data-Dependent Score Construction

We study data-driven score construction in conformal selection, which uses conformal p-values and multiple testing to select "interesting" instances with large unobserved labels from a pool of unlabeled data, while controlling the FDR in finite sample. While existing solutions require the score function to be constructed independent of the calibration and test data, it is desirable to select or construct powerful conformity score in a data-driven manner and to leverage all labeled data for score training when presented with many score choices and limited labeled data. We propose Optimized Conformal Selection (OptCS), a general framework that allows valid multiple testing after flexible data-driven score construction. We introduce general conditions under which OptCS constructs valid conformal p-values despite substantial data reuse and handles complex p-value dependencies to maintain finite-sample FDR control. We instantiate this general recipe to propose three FDR-controlling procedures, each optimizing the models differently: (i) selecting a powerful score function among multiple candidates, (ii) training the score with all labeled data without sample splitting, and (iii) combining full-sample training and model selection. We demonstrate the efficacy of our methods via simulation studies and real applications in drug discovery and large language models.

stat.ME↗
arXiv2026

Network Cross-Validation and Model Selection via Subsampling

Complex and larger networks are becoming increasingly prevalent in scientific applications in various domains. Although a number of models and methods exist for such networks, cross-validation on networks remains challenging due to the unique structure of network data. In this paper, we propose a general cross-validation procedure called NETCROP (NETwork CRoss-Validation using Overlapping Partitions). The key idea is to divide the original network into multiple subnetworks with a shared overlap part, producing training sets consisting of the subnetworks and a test set with the node pairs between the subnetworks. This train-test split provides the basis for a network cross-validation procedure that can be applied on a wide range of model selection and parameter tuning problems for networks. The method is computationally efficient for large networks as it uses smaller subnetworks for the training step. We provide methodological details and theoretical guarantees for several model selection and parameter tuning tasks using NETCROP. Numerical results demonstrate that NETCROP performs accurate cross-validation on a diverse set of network model selection and parameter tuning problems. The results also indicate that NETCROP is computationally much faster while being often more accurate than the existing methods for network cross-validation.

stat.ME↗