Caution
: This option creates a large SAS data set and opens it in JMP for plotting. It can greatly slow execution time for large trials.
Residuals are computed as observed d
dependent variable
values minus the predicted values from the
ANOVA
model. Studying the residuals can help you decide on the validity of the assumptions underlying the
ANOVA
model, namely, that the errors are approximately normally distributed and independent. This option creates
quantile
-quantile (
Q-Q
) plots (see
Q-Q Plot
) and
Scatterplot
s
of the standardized residuals from the
ANOVA
model fits. These plots are useful for assessing the quality of each fit.
