Design of Experiments

Abstract

MAJOR FACTORS TO BE CONSIDERED IN THE DESIGN OF RESEARCH EXPERIMENTS ARE NOTED BRIEFLY. FOUR ARE CONSIDERED ESSENTIAL TO VALIDATE EXPERIMENTAL RESULTS: CAREFUL AND EXPLICIT PROBLEM DEFINITION, COST-EFFECTIVE REPLICATION, ANALYSIS OF "INFERENCE SPACE" (LIMITATIONS OF DATA, POPULATION, OR EXPERIMENTAL DESIGN THAT CONSTRAIN APPLICATION OF THE RESULTS WITHIN CERTAIN BOUNDS), AND UNDERSTANDING OF RESTRICTIONS ON RANDOMIZATION. METHODS OF HANDLING EXTRANEOUS VARIABLE AND THE USE OF HISTORICAL DATA ARE DISCUSSED, AND AN EXAMPLE IS GIVEN OF THE USE OF RANDOMIZATION RESTRICTION IN THE DESIGN OF A HEART VALVE EXPERIMENT.