This paper presents a general loss function under quadratic loss structure and discusses the comparison of risk functions associated with the unbiased least squares and biased Stein-rule estimators of the coefficients in a linear regression model.