Abstract

This thesis is a study of Estimators, particularly in Linear Models. The newest technology of Bootstrap Methodology is employed in the estimation procedure. We present a survey of the Bootstrap Methodology in the beginning and move on to some serious problems in Linear Model estimation procedure. We have worked out the conditions under which the estimators of nonstandard linear models will be best linear unbiased estimators. Furthermore, we have shown that the estimators of other linear models bear a linear relationship with least squares estimators. Finally, we have worked out the finite sample properties of two-stage least-squares using the Bootstrap Methodology.