The variables of your survey or experiment are meaningless by themselves: What gives them meaning is your hypothesis. In this article, I show how ignoring this principle can lead to very different results when analyzing data from the same experiment.
