If lots of things were changed (independent variable) then you wouldn’t know what caused the trend(s) in your results so you have to control all but one variable so you know that you are just changing the one thing so that is the only thing that could cause the trend. I hope that made sense
