ABSTRACT
A dependent variable that represents measurements taken from an economic system is unlikely to be related only to a single explanatory variable, and it is necessary to consider a more general version of the linear model which allows the list of explanatory variables to be extended. The knowledge that parameter estimates can be obtained by solution of the normal equations is sufficient to define the least squares method. The simple correlation coefficient was used as a measure of the extent to which the behaviour of the dependent variable is explained by the behaviour of a single genuine explanatory variable. The predicted values represent a component of the observations on the dependent variable that can be explained as an exact linear function of the explanatory variable observations. The type of logarithms most frequently used for performing calculations are to base 10, but in economic modelling one would usually employ natural logarithms, which use the mathematical constant e as the base.
