Admissibility of confidence estimators in the regression model

In the regression model, we assume that the independent variables are rando m instead of fixed. Consider the problem of estimating the coverage functio n of a usual confidence interval for the unknown intercept parameter. In th is paper, we consider a case in which the number of unknown parameters is s maller than 5. We show that the usual constant coverage probability estimat or is admissible in the usual sense in this case. Note that this estimator is inadmissible in the usual sense in the other case where the number of un known parameters is greater than 4. (C) 2001 Academic Press.