Prediction profiler for several responses fit by different models

[#1]

When I have done a multi-factor experiment I fit models to each of my responses. Many times the different responses don't have the same significant terms and interactions as each other. After i reduce my model for each response I want to investigate the effects of changing factors on each response simultaneously. How do I do this when the models I fit are different for each response?

[#2 · Accepted solution]

For each response save the prediction formula to the data table (red triangle hot spot-Save Columns>Prediction Formula). Then use the Graph>Profiler platform to specify the formulas that you saved to the columns in the data table as your Y (Prediction Formula) responses. This will bring all of your responses with their specific fit data into a combined profiler.

[#3]

Great. That was just what I needed. Thank you a bunch.

[#4]

You can save the Predicted standard error too if you need that in your profiler as well.