| A checklist
of how to compare DOE results to the experimental goals |
Assume that you have a final
model that has passed all the relevant tests (visual and quantitative)
and you are ready to make conclusions and decisions. These should be responses
to the questions or outputs dictated by the original experimental goals.
Checklist relating DOE conclusions or outputs to experimental goals
or experimental purpose:
Are the factor levels significantly different? (comparative experiment
goal)
Which are the significant effects, or terms in final model? (screening
experiment goal)
What is the model for estimating responses?
Full factorial case (main effects plus significant interactions)
Fractional factorial case (main effects plus significant interactions that
are not confounded with other possibly real effects)
RSM case (allowing for quadratic or possibly cubic models, if needed)
What responses are predicted and how can responses be optimized? (RSM goal)
Contour plots
JMP prediction profiler (or other software
aids)
Settings for confirmation runs and prediction intervals for results
|
