Title: Statistical tests for range-based comparisons: Increase flexibility for producers without increasing risk for consumers?
Authors: Gerhard Goessler - University of Graz (Austria)
Vera Hofer - University of Graz (Austria) [presenting]
Walter Goessler - University of Graz (Austria)
Abstract: Based on the concept of k-equivalence some statistical tests for range-based comparisons of quality attributes are discussed. Their properties are compared to statistical approaches proposed by agencies like the FDA or by researchers in academia and in industry. It is demonstrated that most approaches are easy to apply but they are not suited for a range-based comparison of quality attributes. Moreover, advantageous properties of certain statistical approaches are obtained only for a relatively large sample size. The industry can often not utilized them since typically being restricted to a small sample size. Using low numbers of observations may even render statistical comparisons unacceptable owing to the level of patient or producer risk they entail.