Determining Statistically Valid Sample Size