The independent variable is the only one that can change. Outcome variable depends on the test variable, so it isn't purposely changed. The control group isn't changed, because then it wouldn't be a control. And if you changed the hypothesis, you would basically be cheating and saying you knew what would happen before the experiment.