An official website of the United States government
We propose a new method for evaluating the mean square error (mse) of a possibly biased estimator, or, rather, the class of estimators to which it belongs. The method uses confidence intervals c of a corresponding unbiased estimator and makes its assessment based on the extent to which c includes the (possibly biased) estimator of interest. The method does not require an estimate, implicit or explicit, of the bias of the estimator of interest, is indifferent to the bias/variance breakdown of its mse, and does not require surety of the model on which it is based.