What effect does increasing the sample size have on the power of a hypothesis test, assuming all other factors remain constant?