Rodent body weight data from National Toxicology Program studies is an important end point used to determine if a toxicant causes adverse effects. Statistical tests for differences between body weights of control and treated groups often assume that the data are normally distributed (i.e., are bell-shaped curves). This study evaluated the importance of the normality assumption in statistical testing of rodent body weight measurements. It was found that the normality test used in this project, Shapiro-Wilk’s test for normality, has 6-56% power to detect skew normal distributions with samples sizes of 50 animals or less. However, statistical tests that compare body weights in a control group to a treated group were able to detect 10% differences in body weight with at least 80% power for sample sizes of 10, 20, and 50 rodents for both normally distributed data and data with a skew normal distribution while keeping False Positive Rates at an acceptable level of approximately 5%.