Which statistical analysis should I use for this design?

Question

I'm writing a research proposal for a study involving two independent variables (3 levels to each) and three dependent variables. The IVs are Caffeine dosage (0, 300mg, 600mg) and Sleep (0 hrs, 4 hrs, and 8 hrs). The dependent variables will be performance on three types of cognitive tasks: Memory, Reasoning, and Attention.

Basically, subjects will be exposed to a certain number of hours of sleep, a certain dose of caffeine, and then be tested on the three cognitive tasks. What would be the best statistical analysis for this? I was thinking a MANOVA, but I'm not quite sure how I would set it up.

Answer 1

You can either have two categorical variables with three categories each, in which case you'll create dummy variables during the analysis phase, or (what I'd do) you can treat them as numerical variables and not have to mess around with creating the synthetics.  That would also end up providing an estimate as to how hours of sleep correspond to MG of caffeine.