Experimental Designs

Study designs will be study-specific. The feasibility of different study designs will be strongly influenced by characteristics of the different designs and by the available opportunities for applying the treatment (i.e., available treatment structures). Other, more practical considerations include characteristics of study subjects, study sites, the time available for the study, the time period of interest, the existence of confounding variables, budget, and the level of interest in the outcome of the study by others. The collection of concomitant variables will almost always improve the study design. Regardless of the environment within which studies are conducted, all protocols should follow good scientific methods. Even with the best of intentions, though, study of results will seldom lead to clear-cut statistical inferences. There is no single combination of design and treatment structures appropriate for all situations. Our advice is to seek assistance from a statistician and let common sense be your guide.