I have the following experimental data and desperately lack expert advice:
Of a total of 60 Subjects, 20 are randomly assigned to one of three conditions (1X3) during treatment
Each subject performs six tasks in random order. This is because I did not have access to a sufficient number of subjects to have one subject perform one task per treatment.
There are three dependent variables. The dependent variables are based on task outcomes and are to be related to the treatment condition.
I have to control for the fact that task outcomes are not independent within subjects (referring to the 3 * 6 outcomes by each subject)
Subject specific variables should also be controlled as they supposedly influence task behavior (e.g. gender, product involvement)
My thesis advisor first suggested a MANCOVA model for analysis, but after reading into the topic, I don’t find it appropriate at all since it doesn’t account for the within subject error.
However, I want to do things correctly. I believe the correct model would be a multi-level model but I hardly find information how to render my design in a multi-level model. I have even trouble to formulate my design in a multi-level model, if I only have one dependent variable. I would greatly appreciate some help: What would my model look like, where can I read about my specific problem. Thanks.