If case weighting is in effect, [C]NLR uses case weights when calculating the residual
sum of squares and derivatives. However, the degrees of freedom in
the ANOVA table are always based
on unweighted cases.
When the model program is first invoked for each
case, the weight variable's value is set equal to its value in the
active dataset. The model program may recalculate that value. For
example, to effect a robust estimation, the model program may recalculate
the weight variable's value as an inverse function of the residual
magnitude. [C]NLR uses the weight
variable's value after the model program is executed.