Inferential Statisticshard
0:00.0

In the Rao-Blackwell theorem, if θ^\hat{\theta} is an unbiased estimator for θ\theta and TT is a sufficient statistic, how does the variance of the improved estimator θ~=E[θ^T]\tilde{\theta} = E[\hat{\theta} | T] compare to the variance of θ^\hat{\theta}?