It's okay to measure one thing with something else that's variable. For example let's say I want to determine aluminum's coefficient of thermal expansion. I have a block of aluminum which I am measuring with a steel ruler. Both objects will change size if I vary the temperature, but by measuring both at several temperatures I can determine the ratio of their coefficients of thermal expansion. Funnily enough, if I'm using a mercury thermometer I'm really measuring everything relative to mercury's coefficient of thermal expansion.