A note on detecting outliers in short autocorrelated data using joint estimation and exponentially weighted moving average

Comprehensive results for the joint estimation outlier detection method and the exponentially weighted moving average method with regard to their performance as statistical process control methods or outlier detection methods for short-run autocorrelated data are reported. Both methods are shown to be effective. Extensive tables are presented which may be used by practitioners to determine the best time-series lengths and smoothing constants or critical values for use with these methods. In addition, several of the tables report results at the lowest level for five factors.