Sample size adaptation with two sample t-test
Abstract
The maximum type 1 error rate of the naïve t-test and different ways of adjusting the statistic are computed in the scenario where the sample size is adjusted after observing part of the data. The maximum type 1 error rate of the naïve t-test can be more than double the targeted rate. There is a valid test based on the fixed weight combination of independent test statistics. However, inverting that test does not provide a simple formula for the confidence limits or point estimate of the mean difference.
Fields of Study & Subjects
Statistical Methods in Clinical Trials
Probability and Statistical Research
Statistical Methods and Bayesian Inference
Publication details
- Journal / source
- Communications in Statistics - Simulation and Computation
- Pages
- 1–11
- Publication date
- 4 March 2026
- Language
- EN