it is error or variation not attributable to the effect we’re investigating. The simplest way, therefore, to test whether the model fits the data or whether our hypothesis is a good explanation of the data we have observed, is to compare the systematic variation against the unsystematic variation.
