The researchers stated that the distributions of failure loads were not Normally distributed, but there are no extreme outliers. The test p-value of 0.001 is Question 10 options: unreliable no matter how the data are distributed. inaccurate because the populations must be exactly Normally distributed to perform ANOVA. inaccurate because the populations must be strongly skewed with many extreme outliers to perform ANOVA. reasonably accurate anyway because the sample sizes are large (26 each).