Problems involved in the specification and interpretation of quantitative structure-activity relationships. Part 1: A...

It has been demonstrated that ordinary regression coefficients of a quantitative structure-activity equation depend strongly on the scaling both of the biological parameter and the explanatory variables (regressors). Even in the case of only two regressors this fact may lead to challengable interpretations regarding the relative performances of the explanatory variables, say physicochemical parameters. The use of standardized variables and, consequently, of standardized regression coefficients is proposed in order to get a better insight into the mechanistic features of the mode of action of the congeneric series under investigation. The procedure has been illustrated with examples adapted from the literature.