by Robert J. Seely, Amgen, Inc. Louis Munyakazi, John Haury, Heather Simmerman, W. Heath Rushing, and Thomas F. Curry Determining whether a data point is an "outlier" ? a result that doesn't fit, that is too high or too low, that is extreme or discordant ? is difficult when using small data sets (such as the data from three, four, or five conformance runs). The authors show that the Weisberg t-test is a powerful tool for detecting deviations in small data sets.
by Robert J. Seely, Louis Munyakazi, and John Haury Regulatory approval and successful manufacturing depend on establishing meaningful and reasonable acceptance criteria for process validations and ongoing monitoring. Three methods are presented here to correct for the likely underestimation of process limits due to small samples.