McMASTER UNIVERSITY STATISTICS SEMINAR

Week of February 28 - March 3, 2000

SPEAKER:	Peter Kupchak Department of Statistics, University of Toronto
TITLE:	"Optimal design theory for nonlinear dose-response models"
DAY:	Wednesday, March 1, 2000
TIME:	3:30 p.m. [Coffee & cookies in BSB-202 at 3:00 p.m.]
PLACE:	BSB-108

SUMMARY

When designing for the precise estimation of the parameters in a linear model, one can make use of the well-known fact that the information matrix associated with one's design is independent of the parameter values. However, the efficiency of a given design relative to a specified optimality criterion depends on the values of the unknown parameters when the underlying model is nonlinear. In this talk, I present a general overview of optimal design theory for linear models, and show how the theory can be applied to the specific problem of testing for interaction between two drugs, given that the probability of response to an applied dose combination can be modelled using a bivariate logistic function.

ABOUT THE SPEAKER

• undergraduate degree in Materials Science and Engineering obtained from McMaster Unversity;
• undergraduate degree in Education obtained from the University of Toronto;
• M.Sc. in Statistics obtained from McMaster University, 1996;
• currently a Ph.D. candidate in the Department of Statistics, University of Toronto.

REFERENCES

The references below have been suggested by Peter Kupchak as useful background for his talk. They have been placed on reserve at Thode Library (STATS 770: Statistics Seminar).

[1] Atkinson, A.C. and Donev, A.N. (1992). Optimum Experimental Designs. Oxford: Clarendon Press. [A clearly-written monograph, especially suited for students and researchers who are new to the topic.]

[2] Fedorov, V.V. and Hackl, P. (1997). Model-Oriented Design of Experiments. New York: Springer-Verlag.

[3] Silvey, S.D. (1980). Optimal Design. London: Chapman and Hall.