Computation of the uniform minimum variance unbiased estimator of a normal mean following a group sequential trial.

The sampling distribution of data collected in a group sequential trial is such that the usual fixed-sample estimates of treatment effect are biased. Improved estimates can be obtained by taking the group sequential stopping rule into account. In particular, in the case of inference about the mean of a normal distribution, the sample mean is no longer the uniform minimum variance unbiased estimator (UMVUE). In this paper, I present a way in which the UMVUE for a normal mean can be calculated using software capable of determining the operating characteristics of a group-sequential test.