Degrees of freedom are the
number of data observations beyond the minimum necessary to calculate a
regression coefficient or a statistic.
For example, to calculate the intercept term at least one observation is
needed. Therefore, in a problem with 55
total observations the corrected total degrees of freedom are 54. Similarly, to calculate a regression
coefficient and an intercept two observations are required for the model
degrees of freedom. Finally, the
difference between the corrected total and model degrees of freedom is the
error degrees of freedom.