Data collection is sometimes performed without considering the effect poorly collected data has in the strength of statistical conclusions. When researchers properly design an experiment, they are able to control important factors associated with the outcome of interest, therefore collecting useful data that may provide strong evidence for causality.
In this short course, we will discuss the three basic principles of design of experiments, namely: 1) randomization (to ensure results are not biased by time, order, fatigue, etc.), 2) replication (to improve our estimation of the effects), and 3) local control of error (to reduce variability from known sources). We will also use two example data sets to illustrate the three principles and their application to two straightforward design scenarios: The Complete Randomized Design with the Plasma Etching Experiment and the Randomized Complete Block Design with the Vascular Craft Experiment, both found the Douglas Montgomery Design and Analysis of Experiments book. We will enter and, and analyze the data in JMP, software that can be acquired in Torgesen Bridge . This course is aimed for students with no design of experiments background and assumes that the audience is familiar with concepts such as hypothesis testing and p-value.