Why are people disagreeing with this? If you run an experiment looking at a variable change and it doesn't change much in some places but changes a lot in others, then wouldn't this prove that there are other variables at play - that it couldn't possibly be that single variable itself?