Cross-Over Design Experiment (Using SAS)

Abstract

Cross-over experiments are a special class of repeated measures experiments. Up to now the repeated measures experiments that we have been investigating, are where the experimental unit is measured repeatedly after treatment. Only one treatment is applied to each of the experimental units. Inter observation time do not given other treatments, so it just sort of pure observation from one moment to the next. In this experimental design, the general assumption used is that the experimental units are relatively homogeneous. In experiments involving biological creatures, like animals, human studies, psychology, etc., we will find it difficult to obtain experimental unit as much as we need, which has a homogeneous condition. Because in this study, there are things that physically we can’t find a striking difference but basically every individual has the distinction, and this can make the measurements that we do less valid. In addition, the number of treatments with repeats that we demand that we have enough experimental units, but often problems to get it. Cross-Over design experiment, implementing the experiment in which one unit of the experiment received all treatments with a specific sequence and the observations were made at each meal treatment. This design is often also referred to as a Change-Over Design. Using a cross-over experiment allows for an increase in precision when less variability is expected within subjects than between subjects. One feature of cross-over designs with more than two periods, is the ability to measure any possible carryover effects. Carryover effects are when the results in subsequent treatments are influenced by treatments given in previous periods or may also be caused by a “learning effect”.