In a study designed to measure whether the number of hours of sleep significantly predicts higher test grades, researchers vary the number of hours of sleep that subjects get per night. This method of manipulating one variable to measure change in a second variable is called: