A simplified two sample sequentialt-Test

From MaRDI portal
Publication:3685886





Current formulations of the two-sample sequential t-test have the practical disadvantage that they require sophisticated computational procedures for their implementation. A simplified two-sample sequential t-test, based on the fixed-sample t-test is proposed and its properties investigated. Using Monte Carlo simulation, the decision boundaries are approximated as are the distributions of the sample sizes required to make a decision under different alternative hypotheses. Truncated gamma densities are used to fit the distributions of the sample size and a simple relation is given for the calculation of the expected sample size and its variance for a class of alternative hypotheses. The proposed test is shown to be a closed test and the relative efficiency of this test to the fixed sample t-test is examined. It is found that there is a saving in sample size of between 10 \% and 43 \% using this sequential test. The usefulness of this test is illustrated using data arising from a problem in medical research.




Cited in
(22)








This page was built for publication: A simplified two sample sequentialt-Test

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3685886)