Misconception CVM021:

When testing for the effect of a particular variable on an outcome, it is important for that variable to vary, but only one other relevant variable has to be controlled and the rest can vary.