Because of small sample sizes, differences are not significant as the probability of making a Type I error exceeds 5%