- Three-Level designs
- 6x3 via Theorem 8
- 9x7 via Theorem 4
- 9x16 and 9x12 via Theorems 6 and 7
- 18x12 via Theorem 8
- 18x12 via Theorem 9
- 27x25 via Theorem 4
- 27x26 via Theorem 5
- 27x169 and 27x156 via Theorems 6 and 7
- 54x39 via Theorem 8
- 54x39 via Theorem 9

- Four-level designs
- 8x4 and 12x4 via Theorem 8
- 16x9 via Theorem 4
- 16x15 via Theorem 6
- 32x20 and 48x20 via Theorem 8
- 48x20 via Theorem 9
- 64x41 via Theorem 4
- 64x231 via Theorem 6
- Five-level designs
- 10x5, 15x5 and 20x5 via Theorem 8
- 25x11 via Theorem 4
- 25x36 and 25x30 via Theorems 6 and 7
- 50x30, 75x30 and 100x30 via Theorem 8
- 50x30 , 75x30 and 100x30 via Theorem 9

- Published [pdf] in
*Annals of Statistics*(2005) 33, 2811-2836. - Xu and Wu (2003) UCLA Statistics Electronic Publications preprint 356
- Slides for talk presented at the "INTERNATIONAL
CONFERENCE ON DESIGN OF

EXPERIMENTS: THEORY AND APPLICATIONS", Memphis, TN.

Send comments and questions to

Hongquan Xu

May 2005.