Mixed-Level Factorial Design in Small Blocks

Experiments of block size two has been found useful in microarray experiments. Many researcher has worked on this topic since Draper and Guttman(1997). Two different approaches has been considered in different aspects. One aspect is to emphasize the precision of estimators, and the other is to save costs when all available effects are estimated. One approach uses the traditional designing method and the other does the assigning method using orthogonal arrays. As well known, it needs many runs when estimating all the available effects in experiment with small blocks. Kerr (2006) suggested good methods for design experiments in blocks of size two with the applications in microarray experiments. However, her designs require many experimental runs. Most of researches deal with the cases with two-level factors in blocks of size two. In this project, we intend to find designs with four-level factors involved in small sizes.

Project ID:PA9907-0334
External Project ID:NSC99-2118-M182-001