Sequential stratified inference for the mean

TL;DR

This paper introduces a new sequential stratified inference method for estimating the population mean, providing conservative, anytime-valid tests suitable for risk-limiting post-election audits with smaller expected sample sizes.

Contribution

It develops a novel approach combining test supermartingales and union-intersection hypotheses for stratified sampling, optimizing for minimal expected stopping time.

Findings

Expected sample size is significantly reduced compared to previous methods.

The method is applicable to risk-limiting post-election audits.

It allows flexible, sequential sampling with optional stopping.

Abstract

We develop conservative tests for the mean of a bounded population under stratified sampling and apply them to risk-limiting post-election audits. The tests are ``anytime valid'' under sequential sampling, allowing optional stopping in each stratum. Our core method expresses a global hypothesis about the population mean as a union of intersection hypotheses describing within-stratum means. It tests each intersection hypothesis using independent test supermartingales (TSMs) combined across strata by multiplication. A $P$-value for each intersection hypothesis is the reciprocal of that test statistic, and the largest $P$-value in the union is a $P$-value for the global hypothesis. This approach has two primary moving parts: the rule selecting which stratum to draw from next given the sample so far, and the form of the TSM within each stratum. These rules may vary over intersection hypotheses. We construct the test with the smallest expected stopping time and present a few strategies for approximating that optimum. In instances that arise in auditing and other applications, its expected sample size is substantially smaller than that of previous methods.