To ensure that your results are valid, consider the following guidelines when you collect data, perform the analysis, and interpret your results.
A designed response surface experiment in Minitab must have at least 2 continuous factors for a central composite design and at least 3 continuous factors for a Box-Behnken design.
A binary response has two outcomes, such as pass or fail. Response data should be organized into two worksheet columns using the event/trial format. For more information, see Enter your data for Analyze Binary Response for Response Surface Design.
If your response is continuous, then you should use Analyze Response Surface Design.
If the variability in your measurement system is too great, your experiment may lack the power to find important effects.
Randomization reduces the chance that uncontrolled conditions will bias the results. Randomization also lets you estimate the inherent variation in materials and conditions so that you can make valid statistical inferences based on the data from your experiment.
If the model does not fit the data, the results can be misleading. In the output, use the residual plots, the diagnostic statistics for unusual observations, and the model summary statistics to determine how well the model fits the data.