The main contribution of this work lies in challenging the common practice of inferential statistics in the realm of simple linear regression for attaining a higher degree of accuracy when multiple observations are available, at least, at one level of the regressor variable. We derive sufficient conditions under which one can improve the accuracy of the interval estimations at quite affordable ...
