Tukey’s Outlier Filter
Data fields (or variables, such as royalty rates) contain outliers, and the larger the sample size the higher the probability of getting at least one outlier (called “one-wild”).  When outliers are present, estimates of the average and the standard deviation are distorted.  Therefore, these statistics should not be used to produce confidence intervals about the average of the data distribution.  Two strategies can be pursued when outliers are detected.  First, we can use quartiles, because quartiles are less sensitive to outliers.  Second, we can remove the identified outliers, and then produce estimates of the average, the standard deviation, and confidence intervals using the data without outliers.  John Tukey provided a method to set aside outliers based on the following quartile rule:  Set aside observation Y from the computation of the average and the standard deviation if:

      Y < (Q1 − 1.5 IQR) or Y > (Q3 + 1.5 IQR),

where Q1 denotes the lower quartile, Q3 denotes the upper quartile, and IQR = (Q3 − Q1) denotes the interquartile range.  

See David Hoaglin, Frederick Mosteller, and John Tukey (editors), Understanding Robust and Exploratory Data Analysis, New York, John Wiley & Sons, 1983, pp. 39, 54, 62, 223.