The error term refers to the difference
(a residual) between what is observed (ACTUAL) and what is measured
or estimated (FITTED) using a regression model. This error term
also referred to as the unsystematic component of the regression
ACTUAL - FITTED = ERROR
The inclusion of an error term in the regression
model is justified on the following grounds:
- As recognition of an imperfect model.
That is, some variables that help to "explain" the
behavior of the dependent variable might not be included, either
because the analyst cannot measure them or simply because of
- The included variables may have been measured
- There is always some component in the
variation of the dependent variable that is completely random.