Improved estimation of the slope parameter in a linear ultrastructural model when measurement errors are not necessarily normal

An ultrastructural model framework of linear regression relationship between the study and explanatory variables is considered which allows a comprehensive treatment of the classical linear regression model which is free from contamination and the two popular forms, viz., the functional and the structural, of the measurement error model under one roof. Assuming knowledge of the variance of the measurement errors associated with explanatory variable, a consistent class of the slope parameter has been considered and large-sample asymptotic properties have been studied when distributions of measurement errors are not necessarily normal.