JIAB-2026-AST2
Randomization Inference for Small Samples: Power and Size Control
Published
Editor
Sofia Reyes
Handling editor for this submission.
Assigned editor: Sofia Reyes
Update the handling editor to rebalance workload or hand off the submission.
Abstract
Randomization inference (RI) is often recommended for small-sample experiments, yet standard RI tests can be severely underpowered relative to parametric alternatives. We derive power bounds for common RI procedures and show that matching on pre-treatment covariates recovers most of the power loss at no cost to size control, even under sharp-null violation.
Reviews
2 itemsR1 Henry Weber Submitted Jul 25, 2026
The paper is technically sound and fills a clear gap. I have suggestions for presentation, especially around the power-bound intuition, but I do not require another round of review.
R2 Isabelle Dubois Submitted Jul 30, 2026
The matching-on-covariates result is convincing and the simulations are well documented. I recommend acceptance with minor editorial revisions.
Files
1 manuscriptReviewers
2 of 2 receivedDrag names to change the reviewer order for people who have not been invited yet.
Henry Weber
Isabelle Dubois
Messages
1 messageA Author Cover letter
This paper fills an important gap in the randomization inference literature on small-sample power.
Timeline
3 events
Published
Oct 29, 2026
In Production
Sep 6, 2026
Waiting for Reviewers
Jun 18, 2026
Status override
Override status only to correct a workflow state set in error. Notification actions should use Compose.