Statistical Programs |
College of Agriculture | University of Idaho |
Seminar Announcement |
"Applied Statistics in Agriculture" |
Comparing k Test Treatments to Both
Positive and Negative Controls
Presented By |
Dr. Nairanjana Dasgupta |
Program in Statistics Washington State University |
Tuesday, Feb. 11 3:30 P. M. Room 62 College of Agriculture |
In the past, most comparison to control problems have dealt with testing the k test
treatments to either a positive control or a negative control. However, in certain
situations it is imperative to include both types of control in the study. These
problems came up as specific consulting problems and these are discussed as data
examples. I discuss a method to compare several test treatments to both a negative as
well as a positive control simultaneously. Specifically, to see if the test treatments
are worse than the negative control, or are the better than the positive control, while
controlling Family Wise Error. The least favorable configuration under our composite
null is found analytically and one-sided confidence intervals are developed. The
possible alternatives to the proposed method are using Dunnett's (1955) method or the
Bonferroni correction. We show via Monte Carlo simulation that the proposed method is
superior in terms of both Type I error and marginal power. Key words: boundary condition of the null; least favorable configuration; Family-wise Error; one-sided confidence intervals; error control; marginal power; multiple comparison. Joint work with: Eleanne Solozarno, University of New Hampshire and Tony Hayter, Georgia Tech University. |
Return to Seminar Listing. |