Screening is just the beginning

Evaluation of the screening experiment will identify the significant variables. Some variables have probably turned out to have a negligible or only a minor influence. 

The insignificant variables and variables with only minor influence can be fixed at favourable or convement settings when the reaction is further investigated. 

Check also the consequences for the interpretation of the result if the insignificant variables are removed from the design by projecting the screening design down to the space spanned by the significant variables. 

Try to fit a response model using the significant variables only. Make diagnostic plots of the residuals to analyze the fit. If an independent estimate of the experimental error is available, use analysis of variance of the reduced model to find out if there is a significant lack of fit. If the reduced model withstands the diagnostic tests, it is a good indication that the important variables have been found. 

The best validation ofa model is, of course, that predictions from the model can be verified by running the corresponding experiment 

---