A closed-form sample size correction k^(α, β, t0) for always-valid inference that achieves target power within ~3pp in Gaussian simulations across three boundary families and saves 8-20% versus the last-point rule.
David Siegmund.Sequential Analysis: Tests and Confidence Intervals
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A closed-form sample size correction for always-valid inference with optional stopping
A closed-form sample size correction k^(α, β, t0) for always-valid inference that achieves target power within ~3pp in Gaussian simulations across three boundary families and saves 8-20% versus the last-point rule.