Abstract
A model of non observable component measurement. A formal presentation of the axiomatic formula tion of the classical test theory. The aim of this paper is to develop in a strict and abstract way the basis of the linear-true score model in the classical test theory. Its purpose is to overcome both the inaccuracy that is frequently seen in the psychometric literature with regard to the aleator y or fixed nature of the model components and its confusing writing down for its reading and understanding. The axioms of the model, their basic properties and their main results about reliability and validity are introduced pointing out the importance of defining them taking into account the non observable component of the measurement. The importance of an abstract presentation for the model lies in the interest of being able to apply it in other contexts -not necessarily of psychological nature- w here the events also involve non observable variables as the true scores in the classical test theory.
