For this example, one variable at a time is eliminated the one whose dlimination will cause the smallest increase in the residual sum of squares. The process will stop when the partial sums of squares for all variables remaining in the model are significant (α = .05)