Comparing results from 2 different tests using a common "control"I have 2 separate market tests. In test #1, each panelist graded BOTH prototype A and prototype B. In test #2, a different group of panelists was used and each graded BOTH prototype A and prototype C. I would like to compare prototype B to prototype C. The data is stacked (see attached). How do I normalize the data based on prototype A (the prototype used in both tests)?
Re: Comparing results from 2 different tests using a common "control"You might use Analyze > Fit Model. Put Data in the Y role. Add Prototype and Panelist to the effects. Select Panelist in the effects list, click the red triangle next to Attributes, and select Random Effect. Then click Run.
You can now click the red triangle at the top and select Estimates > Custom Test or Multiple Comparisons. The analysis is based on the Least Squares Means, which accomplish the same thing that you intended by normalizing.
