# Should I use ANOVA or MANOVA for repeated measures experiment with two groups and several DVs?

I have data on two groups of subjects. Each group was measured three times, and there were four dependent variables all measuring the level of fitness (score of a person's heart rate in five stages of the test, minutes they achieved, distance they achieved and rated perceived exertion).

My hypothesis is that there will be a significant difference in group A's fitness level compared to group B. group A is in advanced training programme and group B is in a 'normal' training programme and their fitness is measured before training, at 4 weeks and at 6 months follow-up.

I'm leaning toward one-way repeated measures ANOVA, but I'm also starting to wonder if I should be using one-way repeated measures MANOVA because I have 4 dependent variables. Or does this not matter? Could I still use repeated measures ANOVA?

• What model/hypothesis do you want to fit/test? May 17, 2014 at 10:01
• Hi nick, thank you for your comment. My hypothesis is that there will be a significant difference in group A's fitness level compared to group B. group A is in advanced training programme and group B is in a 'normal' training programme and their fitness is measured before training, at 4 weeks and at 6 months follow-up
– Emma
May 17, 2014 at 10:23
• Okay, but how do you want to handle "fitness level"? Do you have a measurement model in mind that combines your four dependent variables? They are not four levels of one variable BTW; might want to check your understanding of the difference between levels and variables, unless I'm misunderstanding what you've written. If you don't have anything like a latent factor model in mind for combining the four dependent variables, you'll probably need to test your hypotheses for each separately, or need to develop a latent factor model, which could be tricky. May 17, 2014 at 10:31
• @HorstGrünbusch hi there, thanks for your comment. I think my own understanding of what a dependent variable with 4 levels means was wrong. I should say 4 dependent variables as appose to one dependent variable with 4 levels
– Emma
May 17, 2014 at 11:04
• So you measured the fitness level 4 times? Your first comment suggested 3 measurements: Baseline, 4 weeks, 6 weeks. So I got confused. You are sure that fitness level is metric, e.g. data from a bicycle ergometer? And: Have a look at this older question, please. May 17, 2014 at 11:20