Further Details of an Analysis for Unreplicated Fractional Factorials.
Abstract
Recent work by Box and Meyer has outlined a new approach to analyzing unreplicated fractional factorials, utilizing the factor sparsity principle. It has been shown to be a formal analog to normal probability plotting. Further details of the analysis are explored here, in particular, sensitivity to choice of prior parameters, standard errors of effects, and extention of ideas to blocked and replicated designs. Keywords: Bayesian analysis; Experimental design; Factorial design.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1986
- Accession Number
- ADA180770
Entities
People
- R. D. Meyer
Organizations
- University of Wisconsin–Madison