New method identifies outliers in linear regression unaffected by masking effects.
The article discusses a new method for finding and testing outliers in linear regression. The researchers use a scale ratio test to identify outliers and test for their presence. They propose a forward sequential procedure that is more effective than traditional methods, as it is not affected by certain issues that can distort results. The new method was compared to classical procedures using real data, showing its superiority in outlier detection.