Is this a 2-way or a 1-way ANOVA?

I did an experiment where I took one baseline measure and then randomized people's legs to receive independent treatments. Leg 1 did exercise, while the opposite leg did not (Leg 2). We measured the response after 6wk of exercise and then subjected BOTH legs to further exercise and took another measurement. So I have:

Baseline measure (from a leg at random)
PHASE 1: Leg 1 - exercises, Leg 2 does nothing
- Measurement again of BOTH legs at 6wk
PHASE 2: Legs 1 AND 2 BOTH do a different exercise (for 6wk)
- Measurement again of BOTH legs at 12wk (6wk for phase 1 + 6wk for phase 2)

I can easily run this as a 1-way RM ANOVA and determine differences between means. BUT, my advisor says this is a 2-way RM ANOVA, but if that's so then how do I account for the single baseline measure that is really the 'common baseline' for BOTH legs? IF I had made the measurement at baseline in both legs, then I get it... it's a complete 2-way design with time and leg as factors. However, I only made a single baseline measure... can't go back and redo this and I am stumped!

ANY help would be appreciated!


Active Member
Then one possibility is to subtract the baseline data from all the measurements and ignore the baseline. It is now no longer repeated measures.