To ensure that your results are valid, consider the following guidelines when you perform the analysis and interpret your results.
- Your worksheet must contain either a 2-level factorial design or a mixtures design
- If you do not enter a response variable, a cube plot displays only the design points that represent the design.
- If you want to display fitted means, use Analyze Factorial Design first
- A cube plot obtains the fitted means from the most recent factorial design model that you fit for a response. If you did not analyze a factorial design, you do not have a model to use. Without a stored model, cube plot displays the data means. For more information, go to Stored model overview and Comparison of data means and fitted means.
For mixture designs, you do not need to fit the model first because the cube plot displays only data means.
- Verify that the model meets the assumptions of the analysis
- If the stored model does not meet the assumptions of Analyze Factorial Design, the cube plot might be inaccurate. For more information, go to the Data considerations and Interpret the key results for Analyze Factorial Design.