The Annals of Statistics

Optimal designs for mixed models in experiments based on ordered units

Dibyen Majumdar and John Stufken

Full-text: Open access

Abstract

We consider experiments for comparing treatments using units that are ordered linearly over time or space within blocks. In addition to the block effect, we assume that a trend effect influences the response. The latter is modeled as a smooth component plus a random term that captures departures from the smooth trend. The model is flexible enough to cover a variety of situations; for instance, most of the effects may be either random or fixed. The information matrix for a design will be a function of several variance parameters. While data will shed light on the values of these parameters, at the design stage, they are unlikely to be known, so we suggest a maximin approach, in which a minimal information matrix is maximized. We derive maximin universally optimal designs and study their robustness. These designs are based on semibalanced arrays. Special cases correspond to results available in the literature.

Article information

Source
Ann. Statist., Volume 36, Number 3 (2008), 1090-1107.

Dates
First available in Project Euclid: 26 May 2008

Permanent link to this document
https://projecteuclid.org/euclid.aos/1211819557

Digital Object Identifier
doi:10.1214/07-AOS518

Mathematical Reviews number (MathSciNet)
MR2418650

Zentralblatt MATH identifier
1216.62114

Subjects
Primary: 62K05: Optimal designs
Secondary: 62K10: Block designs

Keywords
Trend effects variance components block designs orthogonal array of type II semibalanced arrays universal optimality

Citation

Majumdar, Dibyen; Stufken, John. Optimal designs for mixed models in experiments based on ordered units. Ann. Statist. 36 (2008), no. 3, 1090--1107. doi:10.1214/07-AOS518. https://projecteuclid.org/euclid.aos/1211819557


Export citation

References

  • [1] Bradley, R. A. and Yeh, C.-M. (1980). Trend-free block designs: Theory. Ann. Statist. 8 883–893.
  • [2] Chai, F.-S. (1995). Construction and optimality of nearly trend-free designs. J. Statist. Plann. Inference 48 113–129.
  • [3] Chai, F.-S. and Majumdar, D. (1993). On the Yeh–Bradley conjecture on linear trend-free block designs. Ann. Statist. 21 2087–2097.
  • [4] Cheng, C. S. (1988). A note on the optimality of semibalanced arrays. In Optimal Design and Analysis of Experiments (Y. Dodge, V. Federov and H. Wynn, eds.) 115–122. North-Holland, Amsterdam.
  • [5] Cox, D. R. (1951). Some systematic experimental designs. Biometrika 38 312–323.
  • [6] Hedayat, A. S., Sloane, N. A. J. and Stufken, J. (1999). Orthogonal Arrays: Theory and Applications. Springer, New York.
  • [7] Jacroux, M., Majumdar, D. and Shah, K. (1995). Efficient block designs in the presence of trends. Statist. Sinica 5 605–615.
  • [8] Jacroux, M., Majumdar, D. and Shah, K. (1997). On the determination and construction of optimal block designs in the presence of trends. J. Amer. Statist. Assoc. 92 375–382.
  • [9] Kiefer, J. (1961). Optimum experimental designs. V. With applications to systematic and rotatable designs. Proc. Fourth Berkeley Sympos. Math. Statist. Probab. 1 381–405. Univ. California Press, Berkeley.
  • [10] Kiefer, J. and Wynn, H. P. (1984). Optimum and minimax exact treatment designs for one-dimensional autoregressive error processes. Ann. Statist. 12 431–450.
  • [11] Lin, M. and Dean, A. M. (1991). Trend-free block designs for varietal and factorial experiments. Ann. Statist. 19 1582–1596.
  • [12] Lin, W.-C. and Stufken, J. (1998). Varietal trials in the presence of trends. J. Combin. Inform. System Sci. 23 295–316.
  • [13] Lin, W.-C. and Stufken, J. (1999). On finding trend-free block designs. J. Statist. Plann. Inference 78 57–70.
  • [14] Majumdar, D. and Martin, R. J. (2002). Finding optimal designs in the presence of trends. J. Statist. Plann. Inference 106 177–190.
  • [15] Majumdar, D. and Martin, R. J. (2004). Efficient designs based on orthogonal arrays of type I and type II for experiments using ordered units over time or space. Statist. Methodol. 1 19–35.
  • [16] Mukerjee, R. and Sengupta, S. (1994). A-optimal run orders in the presence of a trend. Austral. J. Statist. 36 115–122.
  • [17] Phillips, J. P. N. (1968). Methods of constructing one-way and factorial designs balanced for trend. Appl. Statist. 17 162–170.
  • [18] Rao, C. R. (1961). Combinatorial arrangements analogous to OAs. Sankhyā Ser. A 23 283–286.
  • [19] Rao, C. R. (1973). Some combinatorial problems of arrays and applications to experimental designs. In A Survey of Combinatorial Theory (J. N. Srivastava, ed.) 349–359. North-Holland, Amsterdam.
  • [20] Stufken, J. (1988). On the existence of linear trend-free block designs. Comm. Statist. Theory Methods 17 3857–3863.
  • [21] Yeh, C.-M. and Bradley, R. (1983). Trend-free block designs: Existence and construction results. Comm. Statist. Theory Methods 12 1–24.
  • [22] Yeh, C.-M., Bradley, R. and Notz, W. I. (1985). Nearly trend-free block designs. J. Amer. Statist. Assoc. 80 985–992.