James-Stein Estimation from an Alternative Perspective

In the comparison of estimators, the typical viewpoint of mean squared error has been challenged by C. R. Rao. In this article we propose a method of selecting estimators in normal populations based on their regions of preference. These regions of preference are a natural consequence of Rao's emphasis on Pitman nearness. We apply the method in the case of estimation of the mean of a bivariate normal through the James-Stein class.     

A statistical method for the serial comparison of vector cardiograms

In the absence of quantitative clinical standards to detect serial changes in cardiograms, statistical procedures are proposed as an alternative. These procedures are preceded by a dimension reducing orthonormal transformation of the original digitized cardiogram into a lower dimensional feature space. In feature space, mul-tivariate test criteria are given for the detection of changes in covariance matrices or mean vectors of the cardiograms. The flexibility is provided to compare the cardiograms of the same individual pairwise or simultaneously. Some pertinent remarks are also made about controlling the overall level of significance and its impact on the application of these techniques to cardiograms of USAF pilots.     

The geometry of legal principles

We discuss several possible legal principles from the standpoint of Bayesian decision theory. In particular, we show that a compelling legal principle implies compatibility with decisions based on maximizing the expected utility.