Identification of Significant Variables

Computerized routines for plotting make it possible to obtain graphical illustrations by projections of the response surface models and such plots are often sufficient for identifying influential variables. An example from the reduction of enamines is shown below. 

However, when the number of experiments is of the same order as the number of estimated model parameters, which is often the case when fractional designs have been used, no useful estimate of the experimental error variance is available. 

---