As long as the experiment is repeated a sufficient number of times, it does not matter how many variables are used. Yes, an experiment should test only one variable at a time. This ensures that the experimental outcome is clearly due to one identifiable factor.