Identification of Active Factors in Unreplicated Fractional Factorial Experiments.
Technical summary rept.,
WISCONSIN UNIV-MADISON MATHEMATICS RESEARCH CENTER
Pagination or Media Count:
Box and Meyer proposed a method for analyzing unreplicated factorial experiments in which contrasts too large to be attributed to noise are identified. It then remains to associate supposed active contrasts with experimental factors, either as main effects or interactions. This article explores a modification of the Box and Meyer method which allows it to be used for this purpose. The analysis is shown to take the confounding structure of the design into account when assessing the activity of each factor. In addition, the sensitivity of the analysis to various assumptions is explored.
- Statistics and Probability