USING HISTORICAL DATA TO INTRODUCE THE STATISTICAL REGRESSION CONCEPT

Abstract

To initiate the student in the concept of Linear Regression, and in that of Associated Correlation, in an introductory course of Descriptive Statistics the use of historical problems related to the subject is proposed, in particular, of the problems addressed by Galton at the end of the century XIX, associated with genetic inheritance. In addition to transferring the student to the historical context in which they emerged, giving them names of pioneer scientists, we proceed with the resolution of these problems intuitively, visualizing them graphically, testing with different measures of descriptive statistics already known to them, with trial and error, with the support of calculator and spreadsheet. For the end we leave the formalization of the theory. The experience proposes, then, a reverse course to what is usually habitual, from practice to theory. First the student faces a set of problems in which the base is the relationship between variables, unknown by them until the moment of facing them, and using the statistical instruments related to the management of a variable, such as the mean and the standard deviation. Our role as a teacher is to introduce clues that help you overcome obstacles on a practical level. Then those tracks, formalized, will become the theoretical basis of the chapter devoted to regression analysis. We proceed to the evaluation of the experience. The results of the same, valued in three ways, resolution of practical exercises, survey of students and exam grades, show that the process has been positive for the ultimate goal of learning.