Also known as the standard deviation, the sigma measures sample variability: i.e., how closely various data points are clustered around the mean in a given data set. In a normally distributed sample, approximately 68% of the cases lie within plus or minus one standard deviation of the mean, 95.44% lie within plus or minus 2 standard deviations, and 99.74% of the cases lie within 3 standard deviations of the mean.