Department of Statistics Seminar
North Carolina State University

presents

Dr. Raymond H. Myers

Virginia Tech

"Standard Experimental Designs In Nonstandard Conditions"

ABSTRACT

Standard Response Surface designs for model fitting in industry are historically constructed to give good performance under ideal conditions, i.e., normal iid errors with homogeneous error variance. Industrial applications involving regression or response surface model building are not a great deal unlike applications in areas of biostatistics and biopharmaceutics where nonnormal responses such as counts or binary outcomes often are encountered. Much attention has been given to the use of generalized linear models in industry. In this talk, we discuss the efficiency of standard experimental design when the problem involves binary, Poisson and possibly other clearly nonnormal responses.

Friday, April 24, 2000

3:35 - 4:35 pm

206 Cox Hall

Refreshments will be served on the second floor of Dabney Hall (left of Room 222) at 3:00 pm.