Hey people! I'm planing to do a randomized clinical trial in which the objective is to know how does caffeine influences the age that Huntington's disease begins. It is known that Huntington's disease, it is influenced by physical activity and mental exercises like chess and sudoku. Knowing that the initial sample was constituted by 100 people, in which 50 of them were exposed were exposed to caffeine and the other 50 will be exposed to a placebo.
Do you think it is a good idea to divide each group that include 10 people that make mental exercise frequently ( 5- pratice physical activity; 5- do not pratice regurlaly) and 40 people that does not pratice any mental exercise (20-pratice some sport) and other 20 does not) making a total of eight groups.
It is always beneficial, in any study, to be as definitive as possible for greater statistical inference, but for adequate statistical robustness/power, it depends on the number of dependent variables you will be collecting. Optimally, for multivariate analyses, it is suggested that you have 10 subjects per dependent variable (5 subjects per DV would be considered minimal)--the more subjects/DV, the better. Hopefully, that should provide a starting point to how many groups you can make out of 100 subjects in order to have any confidence in your statistical outcomes. Or, you can increase your overall subject size to fit the number of DVs you really need. Finally, ensure that all cell sizes/groups have an equal number of subjects.
You'll never find enough people with Huntington's disease to make this study work and have any useful statitiscal power.