On the Construction of Trend-Resistant Designs for Comparing a Set of Test Treatments with a Set of Controls

Abstract

In this article we consider the problem of comparing a set of standard treatments to a set of test treatments where the treatments are to be applied to experimental units sequentially over time or space and where there may be an unknown time or spatial trend. Methods are given here for constructing run orders of treatments for those situations that are “optimal” and robust against trends that can be expressed as a polynomial function of the order in which the observations are taken.

Key Words:

• A-optimal
• MV-optimal
• Trend free