Confidence interval estimation in heteroscedastic linear models with replicates
A.K. Mohamand and F.A. Graybill (1980). Confidence interval estimation in heteroscedastic linear models with replicates . Journal of Statistical Research, Vol. 14, No. 1-2, pp. ??-??.
Abstract
The least squares[Ls] method is applied to obtain approximate confidence intervals for the simple linear model with replicates and unequal, unknown error variances. Approximate expressions are derived for the degrees of freedom to be used for finding confidence intervals. Performance of the method is reported on the basis of a simulation study which indicates the appropriateness of the method in some cases and underlines the need for further Investigations In others.